Skip to Content

Rule

Medicare Program; Prospective Payment System for Long-Term Care Hospitals RY 2008: Annual Payment Rate Updates, and Policy Changes; and Hospital Direct and Indirect Graduate Medical Education Policy Changes

Document Details

Information about this document as published in the Federal Register.

Published Document

This document has been published in the Federal Register. Use the PDF linked in the document sidebar for the official electronic format.

Start Preamble Start Printed Page 26870

AGENCY:

Centers for Medicare & Medicaid Services (CMS), HHS.

ACTION:

Final Rule.

SUMMARY:

This final rule updates the annual payment rates for the Medicare prospective payment system (PPS) for inpatient hospital services provided by long-term care hospitals (LTCHs). The final payment amounts and factors used to determine the updated Federal rates that are described in this final rule were determined based on the LTCH PPS rate year July 1, 2007 through June 30, 2008. The annual update of the long-term care diagnosis-related group (LTC-DRG) classifications and relative weights remains linked to the annual adjustments of the acute care hospital inpatient diagnosis-related group system, and continue to be effective each October 1. The final outlier threshold for July 1, 2007, through June 30, 2008, is derived from the LTCH PPS rate year calculations. We are also finalizing policy changes which include revisions to the GME and IME policies. In addition, we are adding a technical amendment correcting the regulations text at § 412.22.

EFFECTIVE DATE:

These regulations are effective on July 1, 2007.

Start Further Info

FOR FURTHER INFORMATION CONTACT:

Tzvi Hefter, (410) 786-4487 (General information).

Judy Richter, (410) 786-2590 (General information, payment adjustments for special cases, and onsite discharges and readmissions, interrupted stays, co-located providers, and short-stay outliers).

Michele Hudson, (410) 786-5490 (Calculation of the payment rates, LTC-DRGs, relative weights and case-mix index, market basket, wage index, budget neutrality, and other payment adjustments).

Ann Fagan, (410) 786-5662 (Patient classification system).

Miechal Lefkowitz, (410) 786-5316 (Graduate Medical Education payments).

Linda McKenna, (410) 786-4537 (Payment adjustments, interrupted stay, and transition period).

Renate Rockwell, (410) 786-4645 (Graduate Medical Education payments).

Elizabeth Truong, (410) 786-6005 (Federal rate update, budget neutrality, other adjustments, and calculation of the payment rates).

Michael Treitel, (410) 786-4552 (High cost outliers and cost-to-charge ratios).

Table of Contents

I. Background

A. Legislative and Regulatory Authority

B. Criteria for Classification as a LTCH

1. Classification as a LTCH

2. Hospitals Excluded from the LTCH PPS

C. Transition Period for Implementation of the LTCH PPS

D. Limitation on Charges to Beneficiaries

E. Administrative Simplification Compliance Act (ASCA) and Health Insurance Portability and Accountability Act (HIPAA) Compliance

II. Summary of the Provisions of the Final Rule

A. Summary of Major Contents of this Final Rule

B. Responses to Comments

III. Long-Term Care Diagnosis-Related Group (LTC-DRG) Classifications and Relative Weights

A. Background

B. Patient Classifications into DRGs

C. Organization of DRGs

D. Update of LTC-DRGs

1. Background

2. Method for Updating the LTC-DRG Relative Weights

3. Budget Neutrality (BN) Requirement for the Annual LTC-DRG Update

E. ICD-9-CM Coding System

1. Uniform Hospital Discharge Data Set (UHDDS) Definitions

2. Maintenance of the ICD-9-CM Coding System

3. Coding Rules and Use of ICD-9-CM Codes in LTCHs

IV. Changes to the LTCH PPS Payment Rates for the 2008 LTCH PPS Rate Year

A. Overview of the Development of the Payment Rates

B. LTCH PPS Market Basket

1. Overview of the RPL Market Basket

2. Market Basket Estimate for the 2008 LTCH PPS Rate Year

C. Standard Federal Rate for the 2008 LTCH PPS Rate Year

1. Background

2. Update to the Standard Federal Rate for the 2008 LTCH PPS Rate Year

3. Standard Federal Rate for the 2008 LTCH PPS Rate Year

D. Calculation of LTCH Prospective Payments for the 2008 LTCH PPS Rate Year

1. Adjustment for Area Wage Levels

a. Background

b. Geographic Classifications/Labor Market Area Definitions

c. Labor-Related Share

d. Wage Index Data

2. Adjustment for Cost-of-Living in Alaska and Hawaii

3. Adjustment for High-Cost Outliers (HCOs)

a. Background

b. Cost-to-charge ratios (CCRs)

c. Establishment of the Fixed-Loss Amount

d. Reconciliation of Outlier Payments Upon Cost Report Settlement

e. Application of Outlier Policy to Short-Stay Outlier (SSO) Cases

4. Other Payment Adjustments

5. Budget Neutrality (BN) Offset to Account for the Transition Methodology

6. One-time Prospective Adjustment to the Standard Federal Rate

V. Other Policy Changes for the 2008 LTCH PPS Rate Year

A. Short-Stay Outlier (SSO) Cases

1. Background

2. Additional Discussion of the SSO Payment Formula (Includes Technical Correction)

3. Determination of Cost-to-Charge Ratios (CCRs)

4. Reconciliation of SSO Cases

B. Expansion of Special Payment Provisions for LTCH Hospitals within Hospitals (HwHs) and LTCH Satellites: Expansion of the 25 Percent Rule to Certain Situations Not Currently Covered Under Existing § 412.534

VI. Computing the Adjusted Federal Prospective Payments for the 2008 LTCH PPS Rate Year

VII. Transition Period

VIII. Payments to New LTCHs

IX. Method of Payment

X. Monitoring

XI. MedPAC Recommendations: The RTI Contract

XII. Graduate Medical Education (GME)

A. GME Background

B. Resident Training in Nonhospital Settings

1. Background

2. Moratorium on Disallowances of Allopathic or Osteopathic Family Practice Residents Training Time in Nonhospital Settings, and Questions and Answers (Qs&As) on CMS Web site (Section 713 of the MMA and § 413.78)

3. Requirements for Written Agreements for Residency Training in Nonhospital Settings (§ 413.78(e))

4. Modification of the Definition of “All or Substantially All of the Costs for the Training Program in the Nonhospital Setting”

5. Implementation of a 90 Percent Cost Threshold

C. Other Issues to be Considered

D. Summary of Final Provisions

XIII. Technical Amendment

XIV. Collection of Information Requirements

XV. Regulatory Impact Analysis

A. Introduction

1. Executive Order 12866

2. Regulatory Flexibility Act (RFA)

3. Impact on Rural Hospitals

4. Unfunded Mandates

5. Federalism Start Printed Page 26871

6. Alternatives Considered

B. Anticipated Effects of Payment Rate Changes

1. Budgetary Impact

2. Impact on Providers

3. Calculation of Prospective Payments

4. Results

5. Effects on the Medicare Program

C. Impact of Other Policy Changes

1. Effects of Policy Expansion of the Special Payment Provisions for LTCH HwHs and LTCH Satellites to Certain Situations Not Presently Covered by Existing § 412.534 for Subclause (I) LTCHs

2. Effects of Policy Change Relating to Payment for Direct Graduate Medical Education (GME)

D. Accounting Statement

Addendum: Tables

Acronyms

Because of the many terms to which we refer by acronym in this final rule, we are listing the acronyms used and their corresponding terms in alphabetical order below:

AAMC Association of American Medical Colleges

AFMAA Academic Family Medicine Advocacy Alliance

AHA American Hospital Association

AHIMA American Health Information Management Association

ALOS Average length of stay

ALTHA Acute Long Term Hospital Association

AMGA American Medical Group Association

AMPRA American Medical Peer Review Association

AOA American Osteopathic Association

APR All patient refined

ASCA Administrative Simplification Compliance Act of 2002 (Pub. L. 107-105)

BBA Balanced Budget Act of 1997 (Pub. L. 105-33)

BBRA Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act of 1999 (Pub. L. 106-113)

BIPA Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Benefits Improvement and Protection Act of 2000 (Pub. L. 106-554)

BN Budget neutrality

CBSA Core-based statistical area

CCR Cost-to-charge ratio

C&M Coordination and maintenance

CMI Case-mix index

CMS Centers for Medicare & Medicaid Services

COLA Cost of living adjustment

CS Consolidated severity-adjusted

CY Calendar year

DSH Disproportionate share of low-income patients

DRGs Diagnosis-related groups

FI Fiscal intermediary

FMC Family Medicine Center

FTE Full-time equivalent

FY Federal fiscal year

GME Graduate medical education

HCO High-cost outlier

HCRIS Hospital cost report information system

HHA Home health agency

HHS (Department of) Health and Human Services

HIPAA Health Insurance Portability and Accountability Act (Pub. L. 104-191)

HIPC Health Information Policy Council

HwHs Hospitals within hospitals

ICD-9-CM International Classification of Diseases, Ninth Revision, Clinical Modification (codes)

IME Indirect medical education

I-O Input-Output

IPF Inpatient psychiatric facility

IPPS [Acute Care Hospital] Inpatient Prospective Payment System

IRF Inpatient rehabilitation facility

LOS Length of stay

LTC-DRG Long-term care diagnosis-related group

LTCH Long-term care hospital

MCE Medicare code editor

MDC Major diagnostic categories

MedPAC Medicare Payment Advisory Commission

MedPAR Medicare provider analysis and review

MMA Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (Pub. L. 108-173)

MSA Metropolitan statistical area

NAICS North American Industrial Classification System

NALTH National Association of Long Term Hospitals

NCHS National Center for Health Statistics

OACT [CMS'] Office of the Actuary

OBRA 86 Omnibus Budget Reconciliation Act of 1986 (Pub. L. 99-509)

OMB Office of Management and Budget

OPM U.S. Office of Personnel Management

O.R. Operating room

OSCAR Online Survey Certification and Reporting (System)

OTN One-Time Notification

PIP Periodic interim payment

PLI Professional liability insurance

PMSA Primary metropolitan statistical area

PPI Producer Price Indexes

PPS Prospective payment system

PRA Per resident amount

PSF Provider specific file

QIO Quality Improvement Organization (formerly Peer Review organization (PRO))

RIA Regulatory impact analysis

RPL Rehabilitation psychiatric long-term care (hospital)

RTI Research Triangle Institute, International

RY Rate year (begins July 1 and ends June 30)

SIC Standard industrial code

SNF Skilled nursing facility

SSO Short-stay outlier

TEFRA Tax Equity and Fiscal Responsibility Act of 1982 (Pub. L. 97-248)

TEP Technical expert panel

UHDDS Uniform hospital discharge data set

I. Background

A. Legislative and Regulatory Authority

Section 123 of the Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act of 1999 (BBRA) (Pub. L. 106-113) as amended by section 307(b) of the Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000 (BIPA) (Pub. L. 106-554) provides for payment for both the operating and capital-related costs of hospital inpatient stays in long-term care hospitals (LTCHs) under Medicare Part A based on prospectively set rates. The Medicare prospective payment system (PPS) for LTCHs applies to hospitals described in section 1886(d)(1)(B)(iv) of the Social Security Act (the Act), effective for cost reporting periods beginning on or after October 1, 2002.

Section 1886(d)(1)(B)(iv)(I) of the Act defines a LTCH as “a hospital which has an average inpatient length of stay (as determined by the Secretary) of greater than 25 days.” Section 1886(d)(1)(B)(iv)(II) of the Act also provides an alternative definition of LTCHs: Specifically, a hospital that first received payment under section 1886(d) of the Act in 1986 and has an average inpatient length of stay (LOS) (as determined by the Secretary of Health and Human Services (the Secretary)) of greater than 20 days and has 80 percent or more of its annual Medicare inpatient discharges with a principal diagnosis that reflects a finding of neoplastic disease in the 12-month cost reporting period ending in fiscal year (FY) 1997.

Section 123 of the BBRA requires the PPS for LTCHs to be a “per discharge” system with a diagnosis-related group (DRG) based patient classification system that reflects the differences in patient resources and costs in LTCHs. It also requires that the “per discharge” system maintain budget neutrality (BN). We believe the statutory mandate for BN applies only to the first year of the Start Printed Page 26872implementation of the LTCH PPS such that estimated payments in the first year of the PPS were projected to equal payments that would have been paid for operating and capital-related costs of LTCHs had this new payment system not been enacted.

Section 307(b)(1) of the BIPA, among other things, mandates that the Secretary shall examine, and may provide for, adjustments to payments under the LTCH PPS, including adjustments to DRG weights, area wage adjustments, geographic reclassification, outliers, updates, and a disproportionate share adjustment.

In the August 30, 2002 Federal Register, we issued a final rule that implemented the LTCH PPS authorized under BBRA and BIPA (67 FR 55954). This system uses information from LTCH patient records to classify patients into distinct long-term care diagnosis-related groups (LTC-DRGs) based on clinical characteristics and expected resource needs. Payments are calculated for each LTC-DRG and provisions are made for appropriate payment adjustments. Payment rates under the LTCH PPS are updated annually and published in the Federal Register.

The LTCH PPS replaced the reasonable cost-based payment system under the Tax Equity and Fiscal Responsibility Act of 1982 (TEFRA) (Pub. L. 97-248) for payments for inpatient services provided by a LTCH with a cost reporting period beginning on or after October 1, 2002. (The regulations implementing the TEFRA reasonable cost-based payment provisions are located at 42 CFR part 413.) With the implementation of the PPS for acute care hospitals authorized by the Social Security Amendments of 1983 (Pub. L. 98-21), which added section 1886(d) to the Act, certain hospitals, including LTCHs, were excluded from the PPS for acute care hospitals and were paid their reasonable costs for inpatient services subject to a per discharge limitation or target amount under the TEFRA system. For each cost reporting period, a hospital-specific ceiling on payments was determined by multiplying the hospital's updated target amount by the number of total current year Medicare discharges. (Generally, in this document when we refer to discharges, the intent is to describe Medicare discharges.) The August 30, 2002 final rule further details the payment policy under the TEFRA system (67 FR 55954).

In the August 30, 2002 final rule, we also presented an in-depth discussion of the LTCH PPS, including the patient classification system, relative weights, payment rates, additional payments, and the BN requirements mandated by section 123 of the BBRA. The same final rule that established regulations for the LTCH PPS under 42 CFR part 412, subpart O, also contained LTCH provisions related to covered inpatient services, limitation on charges to beneficiaries, medical review requirements, furnishing of inpatient hospital services directly or under arrangement, and reporting and recordkeeping requirements. We refer readers to the August 30, 2002 final rule for a comprehensive discussion of the research and data that supported the establishment of the LTCH PPS (67 FR 55954).

In the June 6, 2003 Federal Register, we published a final rule that set forth the FY 2004 annual update of the payment rates for the Medicare PPS for inpatient hospital services furnished by LTCHs (68 FR 34122). It also changed the annual period for which the payment rates are effective. The annual updated rates are now effective from July 1 through June 30 instead of from October 1 through September 30. We refer to the July through June time period as a “long-term care hospital rate year” (LTCH PPS rate year). In addition, we changed the publication schedule for the annual update to allow for an effective date of July 1. The payment amounts and factors used to determine the annual update of the LTCH PPS Federal rate is based on a LTCH PPS rate year. While the LTCH payment rate update is effective July 1, the annual update of the LTC-DRG classifications and relative weights are linked to the annual adjustments of the acute care hospital inpatient DRGs and are effective each October 1.

In the Prospective Payment System for Long-Term Care Hospitals RY 2007: Annual Payment Rate Updates, Policy Changes, and Clarifications final rule (71 FR 27798) (hereinafter referred to as the RY 2007 LTCH PPS final rule), we set forth the 2007 LTCH PPS rate year annual update of the payment rates for the Medicare PPS for inpatient hospital services provided by LTCHs. We also adopted the “Rehabilitation, Psychiatric, Long-Term Care (RPL)” market basket under the LTCH PPS in place of the excluded hospital with capital market basket. In addition, we implemented a zero percent update to the LTCH PPS Federal rate for RY 2007. We also revised the existing payment adjustment for short stay outlier (SSO) cases by reducing part of the current payment formula and adding a fourth component to that payment formula. In addition, we sunsetted the surgical DRG exception to the payment policy established under the 3-day or less interruption of stay policy. Finally, we clarified the policy at § 412.534(c) for adjusting the LTCH PPS payment so that the LTCH PPS payment is equivalent to what would otherwise be payable under § 412.1(a).

B. Criteria for Classification as a LTCH

1. Classification as a LTCH

Under the existing regulations at § 412.23(e)(1) and (e)(2)(i), which implement section 1886(d)(1)(B)(iv)(I) of the Act, to qualify to be paid under the LTCH PPS, a hospital must have a provider agreement with Medicare and must have an average Medicare inpatient LOS of greater than 25 days. Alternatively, § 412.23(e)(2)(ii) states that for cost reporting periods beginning on or after August 5, 1997, a hospital that was first excluded from the PPS in 1986 and can demonstrate that at least 80 percent of its annual Medicare inpatient discharges in the 12-month cost reporting period ending in FY 1997 have a principal diagnosis that reflects a finding of neoplastic disease must have an average inpatient LOS for all patients, including both Medicare and non-Medicare inpatients, of greater than 20 days.

Section 412.23(e)(3) provides that, subject to the provisions of paragraphs (e)(3)(ii) through (e)(3)(iv) of this section, the average Medicare inpatient LOS, specified under § 412.23(e)(2)(i) is calculated by dividing the total number of covered and noncovered days of stay for Medicare inpatients (less leave or pass days) by the number of total Medicare discharges for the hospital's most recent complete cost reporting period. Section 412.23 also provides that subject to the provisions of paragraphs (e)(3)(ii) through (e)(3)(iv) of this section, the average inpatient LOS specified under § 412.23(e)(2)(ii) is calculated by dividing the total number of days for all patients, including both Medicare and non-Medicare inpatients (less leave or pass days) by the number of total discharges for the hospital's most recent complete cost reporting period.

In the RY 2005 LTCH PPS final rule (69 FR 25674), we specified the procedure for calculating a hospital's inpatient average length of stay (ALOS) for purposes of classification as a LTCH. That is, if a patient's stay includes days of care furnished during two or more separate consecutive cost reporting periods, the total days of a patient's stay would be reported in the cost reporting period during which the patient is discharged (69 FR 25705). Therefore, we revised § 412.23(e)(3)(ii) to specify that, Start Printed Page 26873effective for cost reporting periods beginning on or after July 1, 2004, in calculating a hospital's ALOS, if the days of an inpatient stay involve days of care furnished during two or more separate consecutive cost reporting periods, the total number of days of the stay are considered to have occurred in the cost reporting period during which the inpatient was discharged.

Fiscal intermediaries (FIs) verify that LTCHs meet the ALOS requirements. We note that the inpatient days of a patient who is admitted to a LTCH without any remaining Medicare days of coverage, regardless of the fact that the patient is a Medicare beneficiary, will not be included in the above calculation. Because Medicare would not be paying for any of the patient's treatment, data on the patient's stay would not be included in the Medicare claims processing systems. As described in § 409.61, in order for both covered and noncovered days of a LTCH hospitalization to be included, a patient admitted to the LTCH must have at least one remaining benefit day (68 FR 34123).

The FI's determination of whether or not a hospital qualifies as an LTCH is based on the hospital's discharge data from the hospital's most recent complete cost reporting period as specified in § 412.23(e)(3) and is effective at the start of the hospital's next cost reporting period as specified in § 412.22(d). However, if the hospital does not meet the ALOS requirement as specified in § 412.23(e)(2)(i) and (ii), the hospital may provide the FI with data indicating a change in the ALOS by the same method for the period of at least 5 months of the immediately preceding 6-month period (69 FR 25676). Our interpretation of § 412.23(e)(3) was to allow hospitals to submit data using a period of at least 5 months of the most recent data from the immediately preceding 6-month period.

As we stated in the FY 2004 Inpatient Prospective Payment System (IPPS) final rule, published in the August 1, 2003 Federal Register, prior to the implementation of the LTCH PPS, we did rely on data from the most recently submitted cost report for purposes of calculating the ALOS (68 FR 45464). The calculation to determine whether an acute care hospital qualifies for LTCH status was based on total days and discharges for LTCH inpatients. However, with the implementation of the LTCH PPS, for the ALOS specified under § 412.23(e)(2)(i), we revised § 412.23(e)(3)(i) to only count total days and discharges for Medicare inpatients (67 FR 55970 through 55974). In addition, the ALOS specified under § 412.23(e)(2)(ii) is calculated by dividing the total number of days for all patients, including both Medicare and non-Medicare inpatients (less leave or pass days) by the number of total discharges for the hospital's most recent complete cost reporting period. As we discussed in the FY 2004 IPPS final rule, we are unable to capture the necessary data from our present cost reporting forms (68 FR 45464). Therefore, we have notified FIs and LTCHs that until the cost reporting forms are revised, for purposes of calculating the ALOS, we will be relying upon census data extracted from Medicare Provider Analysis and Review (MedPAR) files that reflect each LTCH's cost reporting period (68 FR 45464). Requirements for hospitals seeking classification as LTCHs that have undergone a change in ownership, as described in § 489.18, are set forth in § 412.23(e)(3)(iv).

2. Hospitals Excluded From the LTCH PPS

The following hospitals are paid under special payment provisions, as described in § 412.22(c) and, therefore, are not subject to the LTCH PPS rules:

  • Veterans Administration hospitals.
  • Hospitals that are reimbursed under State cost control systems approved under 42 CFR part 403.
  • Hospitals that are reimbursed in accordance with demonstration projects authorized under section 402(a) of the Social Security Amendments of 1967 (Pub. L. 90-248) (42 U.S.C. 1395b-1) or section 222(a) of the Social Security Amendments of 1972 (Pub. L. 92-603) (42 U.S.C. 1395b-1 (note)) (Statewide all-payer systems, subject to the rate-of-increase test at section 1814(b) of the Act).
  • Nonparticipating hospitals furnishing emergency services to Medicare beneficiaries.

C. Transition Period for Implementation of the LTCH PPS

In the August 30, 2002 final rule (67 FR 55954), we provided for a 5-year transition period. During this 5-year transition period, a LTCH's total payment under the PPS was based on an increasing percentage of the Federal rate with a corresponding decrease in the percentage of the LTCH PPS payment that is based on reasonable cost concepts. However, effective for cost reporting periods beginning on or after October 1, 2006, total LTCH PPS payments are based on 100 percent of the Federal rate.

D. Limitation on Charges to Beneficiaries

In the August 30, 2002 final rule, we presented an in-depth discussion of beneficiary liability under the LTCH PPS (67 FR 55974 through 55975). In the RY 2005 LTCH PPS final rule (69 FR 25676), we clarified that the discussion of beneficiary liability in the August 30, 2002 final rule was not meant to establish rates or payments for, or define Medicare-eligible expenses. Under § 412.507, if the Medicare payment to the LTCH is the full LTC-DRG payment amount, as consistent with other established hospital prospective payment systems, a LTCH may not bill a Medicare beneficiary for more than the deductible and coinsurance amounts as specified under § 409.82, § 409.83, and § 409.87 and for items and services as specified under § 489.30(a). However, under the LTCH PPS, Medicare will only pay for days for which the beneficiary has coverage until the SSO threshold is exceeded. (See section V.A.1.a. of this preamble.) Therefore, if the Medicare payment was for a SSO case (§ 412.529) that was less than the full LTC-DRG payment amount because the beneficiary had insufficient remaining Medicare days, the LTCH could also charge the beneficiary for services delivered on those uncovered days (§ 412.507).

E. Administrative Simplification Compliance Act (ASCA) and Health Insurance Portability and Accountability Act (HIPAA) Compliance

Claims submitted to Medicare must comply with both the Administrative Simplification Compliance Act (ASCA) (Pub. L. 107-105), and Health Insurance Portability and Accountability Act (HIPAA) (Pub. L. 104-191). Section 3 of the ASCA requires that the Medicare Program deny payment under Part A or Part B for any expenses incurred for items or services “for which a claim is submitted other than in an electronic form specified by the Secretary.” Section 1862(h) of the Act (as added by section 3(a) of the ASCA) provides that the Secretary shall waive such denial in two specific types of cases and may also waive such denial “in such unusual cases as the Secretary finds appropriate” (68 FR 48805). Section 3 of the ASCA operates in the context of the ASCA provisions of HIPAA, which include, among other provisions, the transactions and code sets standards requirements codified as 45 CFR parts 160 and 162, subparts A and I through R (generally known as the Transactions Rule). The Transactions Rule requires covered entities, including covered health care Start Printed Page 26874providers, to conduct the covered electronic transactions according to the applicable transactions and code sets standards.

II. Summary of the Provisions of the Final Rule

A. Major Contents of This Final Rule

In this final rule, we are setting forth the annual update to the payment rates for the Medicare LTCH PPS, as well as, other policy changes. The following is a summary of the major areas that we have addressed in this final rule.

In section III. of this preamble, we discuss the LTCH PPS patient classification and the relative weights which remain linked to the annual adjustments of the acute care hospital inpatient DRG system, and are based on the annual revisions to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes effective each October 1.

Also, in section III. of this preamble, we have established a BN requirement for the annual update of the LTC-DRG classifications and relative weights to reflect changes in relative LTCH resource use. This requirement ensures that estimated aggregate LTCH PPS payments will not decrease or increase as a result of the annual update to the LTC-DRG classifications and relative weights based on the most recent available data. In this section, we also summarize the proposed severity adjusted MS-LTC-DRGs and the development of the proposed relative weights for FY 2008 presented in the FY 2008 IPPS proposed rule.

As discussed in section IV.C. of this preamble, we are implementing a 0.71 percent update to the LTCH PPS Federal rate for the 2008 LTCH PPS rate year based on an adjustment to account for changes in coding practices. Also in section IV. of this preamble, we discuss the prospective payment rate for RY 2008, and in section VI., we discuss the applicable adjustments to the payment rates, including the revisions to the wage index, the labor-related share, the cost-of-living adjustment (COLA) factors, and the outlier threshold, for the 2008 LTCH PPS rate year.

In section V.A. of this preamble, we discuss our change to the current payment formula for certain SSO cases. That is, those cases with a LOS that is less than or equal to one standard deviation of the ALOS of an IPPS discharge that was grouped into the same DRG. However, in situations where the SSO cases would exceed the IPPS discharge that was grouped in the same DRG, payment would continue to be paid under the existing formula.

In section V.B. of this preamble, we discuss the expansion of the present 25 percent admission policy at § 412.534(c) to those certain situations not already affected by the existing policy. Previously, this policy only applied to co-located LTCHs and LTCH satellites whose percentage of discharges exceeded the 25 percent threshold (or the applicable percentage). This is extended to include an adjusted payment to LTCH discharges that were admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH where those discharges exceed the 25 percent (or applicable percentage) threshold. The final policy also applies to grandfathered LTCHs and satellite facilities of LTCHs that have Medicare discharges that were admitted from a hospital co-located with the LTCH or satellite facility of the grandfathered LTCH.

In section X. of this preamble, we will discuss our on-going monitoring protocols under the LTCH PPS.

In section XI. of this preamble, we discuss the recommendations made by the Research Triangle Institute, International's (RTI) evaluation of the feasibility of adopting recommendations made in the June 2004 Medicare Payment Advisory Commission (MedPAC) Report.

In section XII. of this preamble, we discuss our revisions to redefine the statutory term “all or substantially all of the costs for the training program in the nonhospital setting.” The statute requires that hospitals must pay “all or substantially all” of the costs for a training program in a nonhospital setting in order to count FTE residents training in the nonhospital setting for Medicare graduate medical education (GME) payment purposes. We are revising § 413.75(b) to introduce a new definition of “all or substantially all of the costs for the training program in the nonhospital setting” to mean, at least 90 percent of the total of the costs of the residents' salaries and fringe benefits (including travel and lodging where applicable) and the portion of the cost of teaching physicians' salaries attributable to nonpatient care direct GME activities. In addition, we are revising § 412.105(f)(1)(ii)(C) for IME and § 413.78 to reflect this new definition of “all or substantially all” of the GME costs in a nonhospital setting, effective for cost reporting periods beginning on or after July 1, 2007.

In section XV. of this preamble, we analyze the impact of the changes presented in this final rule on Medicare expenditures, Medicare-participating LTCHs, and Medicare beneficiaries.

B. Responses to Comments

We received 270 comments on the RY 2007 LTCH PPS proposed rule. Comments and responses follow the appropriate policy section in this rule. The following is a comment we received regarding the schedule of the LTCH PPS update.

Comment: One commenter urged CMS to consolidate the July 1 update of the LTCH PPS rates and the October 1 development of the LTC-DRG weights into one publication cycle, a step which the commenter states would be very beneficial for the LTCH industry.

Response: We appreciate the commenter's suggestion and we will evaluate whether such a consolidation is a workable alternative to our present schedule.

III. Long-Term Care Diagnosis-Related Group (LTC-DRG) Classifications and Relative Weights

A. Background

Section 123 of the BBRA requires that the Secretary implement a PPS for LTCHs (that is, a per discharge system with a DRG-based patient classification system reflecting the differences in patient resource use and costs). Section 307(b)(1) of the BIPA modified the requirements of section 123 of the BBRA by requiring that the Secretary examine “the feasibility and the impact of basing payment under such a system [the LTCH PPS] on the use of existing (or refined) hospital DRGs that have been modified to account for different resource use of LTCH patients, as well as the use of the most recently available hospital discharge data.”

In accordance with section 123 of the BBRA as amended by section 307(b)(1) of the BIPA and § 412.515, we use information derived from LTCH PPS patient records to classify these cases into distinct LTC-DRGs based on clinical characteristics and estimated resource needs. The LTC-DRGs used as the patient classification component of the LTCH PPS correspond to the hospital inpatient DRGs in the IPPS. (As discussed in greater detail below in this section, in the FY 2008 IPPS proposed rule, we have proposed to adopt the severity-weighted patient classification system, the proposed MS-LTC-DRGs, for the LTCH PPS beginning in FY 2008, which is the same patient classification system proposed for use under the IPPS for FY 2008.) We assign an appropriate weight to the LTC-DRGs to account for the difference in resource use by patients exhibiting the case complexity and multiple medical problems characteristic of LTCHs. Start Printed Page 26875

In a departure from the IPPS, we use low volume LTC-DRGs (less than 25 LTCH cases) in determining the LTC-DRG weights, since LTCHs do not typically treat the full range of diagnoses as do acute care hospitals. To manage the large number of low volume DRGs (all DRGs with fewer than 25 cases), we group low volume DRGs into 5 quintiles based on average charge per discharge. (A listing of the current composition of low volume quintiles used in determining the FY 2007 LTC-DRG relative weights appears in the FY 2007 IPPS final rule (71 FR 47974 through 47978). A listing of the proposed composition of low volume quintiles used in determining the proposed FY 2008 MS-LTC-DRG relative weights appears in the FY 2008 IPPS proposed rule.) We also account for adjustments to payments for cases in which the stay at the LTCH is less than or equal to five-sixths of the geometric ALOS and classify these cases as SSO cases. (A detailed discussion of the application of the Lewin Group model that was used to develop the LTC-DRGs appears in the August 30, 2002 LTCH PPS final rule (67 FR 55978).)

B. Patient Classifications Into DRGs

Generally, under the LTCH PPS, a Medicare payment is made at a predetermined specific rate for each discharge; that payment varies by the LTC-DRG to which a beneficiary's stay is assigned. Consistent with our historical practice of having LTC-DRGs correspond to the DRGs applicable under the IPPS, we will continue to model the LTCH-DRGs after their predecessor CMS DRGs. In addition, we are proposing to use the FY 2008 GROUPER Version 25.0 to be effective for discharges occurring on or after October 1, 2007 through September 30, 2008.

Cases are classified into LTC-DRGs for payment based on the following six data elements:

(1) Principal diagnosis.

(2) Up to eight additional diagnoses.

(3) Up to six procedures performed.

(4) Age.

(5) Sex.

(6) Discharge status of the patient.

As indicated in the August 30, 2002 LTCH PPS final rule, upon the discharge of the patient from a LTCH, the LTCH must assign appropriate diagnosis and procedure codes from the most current version of the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). HIPAA Transactions and Code Sets Standards regulations at 45 CFR parts 160 and 162 require that no later than October 16, 2003, all covered entities must comply with the applicable requirements of subparts A and I through R of part 162. Among other requirements, those provisions direct covered entities to use the ASC X12N 837 Health Care Claim: Institutional, Volumes 1 and 2, version 4010, and the applicable standard medical data code sets for the institutional health care claim or equivalent encounter information transaction (see 45 CFR 162.1002 and 45 CFR 162.1102).

Medicare FIs/MACs enter the clinical and demographic information into their claims processing systems and subject this information to a series of automated screening processes called the Medicare Code Editor (MCE). These screens are designed to identify cases that require further review before assignment into a DRG can be made. During this process, the following types of cases, among others, are selected for further development:

  • Cases that are improperly coded. (For example, diagnoses are shown that are inappropriate, given the sex of the patient. Code 68.6, Radical abdominal hysterectomy, would be an inappropriate code for a male.)
  • Cases including surgical procedures not covered under Medicare. (For example, organ transplant in a non-approved transplant center.)
  • Cases requiring more information. (For example, ICD-9-CM codes are required to be entered at their highest level of specificity. There are valid 3-digit, 4-digit, and 5-digit codes. That is, code 262, Other severe protein-calorie malnutrition, contains all appropriate digits, but if it is reported with either fewer or more than 3 digits, the claim will be rejected by the MCE as invalid.)

After screening through the MCE, each claim will be classified into the appropriate LTC-DRG by the Medicare LTCH GROUPER software. As indicated in the August 30, 2002 LTCH PPS final rule, the Medicare GROUPER software, which is used under the LTCH PPS, is specialized computer software, and is the same GROUPER software program used under the IPPS. The GROUPER software was developed as a means of classifying each case into a DRG on the basis of diagnosis and procedure codes and other demographic information (age, sex, and discharge status). Following the LTC-DRG assignment, the Medicare FI/MAC determines the prospective payment by using the Medicare PRICER program, which accounts for hospital-specific adjustments. Under the LTCH PPS, we provide an opportunity for the LTCH to review the LTC-DRG assignments made by the FI and to submit additional information within a specified timeframe as specified in § 412.513(c).

The GROUPER software is used both to classify past cases to measure relative hospital resource consumption to establish the DRG weights and to classify current cases for purposes of determining payment. The records for all Medicare hospital inpatient discharges are maintained in the MedPAR file. The data in this file are used to evaluate possible DRG classification changes and to recalibrate the DRG weights during our annual update under both the IPPS (§ 412.60(e)) and the LTCH PPS (§ 412.517). As discussed in greater detail in sections III.D. and E. of this preamble, with the implementation of section 503(a) of the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 (MMA) (Pub. L. 108-173), there is the possibility that one feature of the GROUPER software program may be updated twice during a Federal FY (October 1 and April 1) as required by the statute for the IPPS (69 FR 48954 through 48957). Specifically, as we discussed in the FY 2007 IPPS final rule, diagnosis and procedure codes for new medical technology may be created and added to existing CMS DRGs in the middle of the Federal FY on April 1 (71 FR 47959 and 47971). However, this policy change will have no effect on the LTC-DRG relative weights during the FY, which will continue to be updated only once a year on October 1, nor will there be any impact on Medicare payments under the LTCH PPS during the FY as a result of this policy. The use of the ICD-9-CM code set is also compliant with the current requirements of the Transactions and Code Sets Standards regulations at 45 CFR parts 160 and 162, published in accordance with HIPAA.

In the IPPS proposed rule, we proposed to create and implement MS-DRGs for FY 2008; that is, the proposed MS-DRGs would be effective beginning with discharges on or after October 1, 2007 through September 30, 2008. The proposed MS-DRGs are a severity-based system of DRGs in which all existing CMS DRGs were refined to better recognize severity of illness among patients. The details of this proposal can be reviewed online at http://www.cms.hhs.gov/​AcuteInpatientPPS/​downloads/​CMS-1533-P.pdf.

Under the broad authority of section 123(a) of the BBRA as modified by section 307(b) of the BIPA, we intend to model the proposed MS-LTC-DRGs on the corresponding CMS DRGs as described in the FY 2008 IPPS proposed rule if this DRG system is implemented for the IPPS in FY 2008. In addition, as Start Printed Page 26876stated above in this section, we intend to use the FY 2008 GROUPER Version 25.0, effective for discharges occurring on or after October 1, 2007 through September 30, 2008 for the LTCH PPS if the IPPS system is implemented for FY 2008.

To elaborate, if the proposed MS-DRGs are adopted for use by the IPPS, the LTC-DRGs will use the same structure as the proposed MS-DRGs, and will be referred to as the MS-LTC-DRGs. Cases will continue to be classified into MS-LTC-DRGs using the six data elements listed above, and will be subject to review by the MCE as they have in the past. After screening through the MCE, claims will be classified into the appropriate MS-LTC-DRG by the LTCH PPS GROUPER software. Following the MS-LTC-DRG assignment, the Medicare FI/MAC determines the appropriate payment using the Medicare PRICER program.

C. Organization of DRGs

The DRGs are organized into 25 major diagnostic categories (MDCs), most of which are based on a particular organ system of the body; the remainder involve multiple organ systems (such as MDC 22, Burns). Accordingly, the principal diagnosis determines MDC assignment. Within most MDCs, cases are then divided into surgical DRGs and medical DRGs. Surgical DRGs are assigned based on a surgical hierarchy that orders operating room (O.R.) procedures or groups of O.R. procedures by resource intensity. The GROUPER software program does not recognize all ICD-9-CM procedure codes as procedures that affect DRG assignment, that is, procedures which are not surgical (for example, EKG), or minor surgical procedures (for example, 86.11, Biopsy of skin and subcutaneous tissue).

The medical DRGs are generally differentiated on the basis of diagnosis. Both medical and surgical DRGs may be further differentiated based on age, sex, discharge status, and presence or absence of complications or comorbidities (CC). The proposed MS-DRGs, as defined in the FY 2008 IPPS proposed rule, and the MS-LTC-DRGs contain base DRGs that have been subdivided into one, two, or three severity levels. The most severe level has at least one code that is a major CC, referred to as “with MCC”. The next lower severity level contains cases with at least one CC, referred to as “with CC”. Those DRGs without an MCC or a CC are referred to as “without CC/MCC”. When data did not support the creation of three severity levels, the base DRG was divided into either two levels or the base was not subdivided. The proposed two-level subdivisions consist of one of the following subdivisions:

  • With CC/MCC.
  • Without CC/MCC.

In this type of subdivision, cases with at least one code that is on the CC or MCC list are assigned to the “with CC/MCC” DRG. Cases without a CC or an MCC are assigned to the “without CC/MCC” DRG.

The other type of proposed two-level subdivision is as follows:

  • With MCC.
  • Without MCC.

In this type of subdivision, cases with at least one code that is on the MCC list are assigned to the “with MCC” DRG. Cases that do not have an MCC are assigned to the “without MCC” DRG. This type of subdivision could include cases with a CC code, but no MCC.

We note that CCs are defined by certain secondary diagnoses not related to, or not inherently a part of, the disease process identified by the principal diagnosis. (For example, the GROUPER software would not recognize a code from the 800.0x series, Skull fracture, as a CC when combined with principal diagnosis 850.4, Concussion with prolonged loss of consciousness, without return to preexisting conscious level.) In addition, we note that the presence of additional diagnoses does not automatically generate a CC, as not all MS-DRGs or MS-LTC-DRGs recognize comorbid or complicating conditions in their definition. (For example, proposed MS-DRG 069, Transient Ischemia (formerly CMS DRG 524, Transient Ischemia), is based solely on the principal diagnosis, without consideration of additional diagnoses for DRG determination.)

As discussed in greater detail in the FY 2007 IPPS final rule (71 FR 47898 through 47912 and 47973), in its March 2005 Report to Congress, “Physician-Owned Specialty Hospitals,” MedPAC recommended that the Secretary improve payment accuracy in the hospital IPPS by, among other things, “refining the current DRGs to more fully capture differences in severity of illness among patients.” (Recommendation 1, p. 93.) As we discussed in that same final rule (71 FR 47973), we did not adopt a new severity-adjusted patient classification system under the IPPS, for FY 2007, but we did refine the CMS DRG patient classification system for Version 24.0 of the GROUPER software to improve the CMS DRG system's recognition of severity of illness for FY 2007. The updates to the CMS DRG patient classification system used under the IPPS for FY 2007 (GROUPER Version 24.0), were also applied to the LTC-DRGs used under the LTCH PPS for FY 2007.

In the FY 2008 IPPS proposed rule, we presented the changes to the proposed MS-DRG patient classification system for FY 2008. In that rule, we proposed the IPPS GROUPER Version 25.0 for FY 2008 to process LTCH PPS claims for LTCH discharges occurring from October 1, 2007 through September 30, 2008. As noted above in this section and as we also discussed in the FY 2007 IPPS final rule, in its March 1, 2005 Report to Congress on Medicare Payment Policy (page 64) and in Recommendation 1 of the 2005 Report to Congress on Physician-Owned Specialty Hospitals, MedPAC recommended that CMS, among other things, refine the current DRGs under the IPPS to more fully capture differences in severity of illness among patients.

D. Update of LTC-DRGs

1. Background

We propose to modify the existing LTC-DRGs so that they reflect the changes made to the CMS DRGs under the proposed IPPS notice. As discussed in greater detail in the FY 2008 IPPS proposed rule, under the LTCH PPS, relative weights for each proposed MS-LTC-DRG are a primary element used to account for the variations in cost per discharge and resource utilization among the payment groups (that is, proposed MS-LTC-DRGs). To ensure that Medicare patients classified to each proposed MS-LTC-DRG have access to an appropriate level of services and to encourage efficiency, each year based on the best available data, we calculate a relative weight for each proposed MS-LTC-DRG that represents the resources needed by an average inpatient LTCH case in that proposed MS-LTC-DRG. For example, cases in a proposed MS-LTC-DRG with a relative weight of 2 will, on average, cost twice as much as cases in a proposed MS-LTC-DRG with a relative weight of 1. Under § 412.517, the proposed MS-LTC-DRG classifications and weighting factors (that is, relative weights) are adjusted annually to reflect changes in factors affecting the relative use of LTCH resources, including treatment patterns, technology and number of discharges.

For FY 2008, the proposed MS-LTC-DRG classifications and relative weights were updated based on LTCH data from the FY 2005 MedPAR file, which contained hospital bills data from the December 2006 update. The proposed MS-LTC-DRG patient classification system is based upon 745 MS-DRGs that formed the structure of the FY 2008 Start Printed Page 26877LTCH PPS GROUPER program. The FY 2008 proposed MS-LTC-DRGs continues to include two “error DRGs.” As in the IPPS, we included two error DRGs in which cases that cannot be assigned to valid DRGs will be grouped. These two proposed error MS-LTC-DRGs are MS-LTC-DRG 999 (Principal Diagnosis Invalid as a Discharge Diagnosis) and MS-LTC-DRG 998 (Ungroupable). The other 743 proposed MS-LTC-DRGs are the same MS-DRGs used in the IPPS GROUPER program for FY 2008 (Version 25.0).

For FY 2008, as discussed in greater detail in the FY 2008 IPPS proposed rule, we proposed to adopt the MS-LTC-DRGs for the LTCH PPS for RY 2008. (Additional information on the proposed MS-LTC-DRG classifications and proposed MS-LTC-DRG relative weights can be found in the FY 2008 IPPS proposed rule.)

In the past, the annual update to the CMS DRGs was based on the annual revisions to the ICD-9-CM codes and was effective each October 1. The ICD-9-CM coding update process was revised as discussed in greater detail in the FY 2005 IPPS final rule (69 FR 48953 through 48957). Specifically, section 503(a) of the MMA includes a requirement for updating diagnosis and procedure codes twice a year instead of the current process of annual updates on October 1 of each year. This requirement is included as part of the amendments to the Act relating to recognition of new medical technology under the IPPS. (For additional information on this provision, including its implementation and its impact on the LTCH PPS, refer to the FY 2005 IPPS final rule (69 FR 48953 through 48957), the RY 2006 LTCH PPS final rule (70 FR 24172 through 24177), and the RY 2008 LTCH PPS proposed rule (72 FR 4783 through 4784).)

As discussed in the RY 2008 proposed rule (72 FR 4784), in implementing section 503(a) of the MMA, there will only be an April 1 update if diagnosis and procedure codes are requested and approved. We note that any new codes created for April 1 implementation will be limited to those diagnosis and procedure code revisions primarily needed to describe new technologies and medical services. However, we reiterate that the process of discussing updates to the ICD-9-CM has been an open process through the ICD-9-CM Coordination and Maintenance (C&M) Committee since 1995. Requestors will be given the opportunity to present the merits for a new code and make a clear and convincing case for the need to update ICD-9-CM codes through an April 1 update.

At the September 2006 ICD-9-CM C&M Committee meeting, there were no requests for an April 1, 2007 implementation of ICD-9-CM codes, and therefore, the next update to the ICD-9-CM coding system will not occur until October 1, 2007 (FY 2008). Presently, as there were no coding changes suggested for an April 1, 2007 update, the ICD-9-CM coding set implemented on October 1, 2006, will continue through September 30, 2007 (FY 2007). As discussed above in this section, the next update to the proposed MS-LTC-DRGs and relative weights for proposed FY 2008 will be presented in the FY 2008 IPPS proposed rule. Furthermore, we will notify LTCHs of any revisions to the GROUPER software used under the IPPS and LTCH PPS that would be implemented April 1, 2008. As noted previously in this section, in the FY 2007 IPPS final rule (71 FR 47973), we established the use of Version 24.0 of the CMS GROUPER, which is used under the IPPS for FY 2007, to classify cases for LTCH PPS discharges that would occur on or after October 1, 2006 and on or before September 30, 2007.

2. Method for Updating the LTC-DRG Relative Weights

As discussed in the August 30, 2002 LTCH PPS final rule that implemented the LTCH PPS, under the LTCH PPS, each LTCH will receive a payment that represents an appropriate amount for the efficient delivery of care to Medicare patients (67 FR 55984). The system must be able to account adequately for each LTCH's case-mix to ensure both a fair distribution of Medicare payments and access to care for those Medicare patients whose care is more costly. Therefore, in § 412.523(c), we adjust the standard Federal PPS rate by the LTC-DRG relative weights in determining payment to LTCHs for each case. As we have noted above, we are proposing to adopt the MS-LTC-DRGs for the LTCH PPS for FY 2008. However, as discussed in the FY 2008 IPPS proposed rule, this proposed change in the patient classification system does not affect the basic principles of the development of relative weights under a DRG-based PPS. For purposes of clarity, in the general discussion below in which we describe the basic methodology of the patient classification system in use since the start of the LTCH PPS, we use the acronym “MS-LTC-DRG” to specify the proposed DRG patient classification system to be used by the LTCH PPS in FY 2008. Although the proposed adoption of the MS-LTC-DRGs would result in some modifications of existing procedures for assigning weights (for example, in cases of zero volume and/or nonmonotonicity, as discussed below), the basic methodology for developing the proposed FY 2008 MS-LTC-DRG relative weights presented in the FY 2008 IPPS proposed rule continued to be determined in accordance with the general methodology established in the August 30, 2002 LTCH PPS final rule (67 FR 55989 through 55991), which is discussed below. Therefore, in the discussion below, the term “LTC-DRGs” will be used in descriptions of the basic methodology established at the beginning of the LTCH PPS that will remain unchanged if we adopt the proposed MS-LTC-DRGs. The use of the term “MS-LTC-DRGs” in the following discussion will indicate a discussion of specifics aspects of our proposed adoption of the severity-weighted patient classification system for FY 2008 as presented in the FY 2008 IPPS proposed rule.)

Under the LTCH PPS, relative weights for each LTC-DRG are a primary element used to account for the variations in cost per discharge and resource utilization among the payment groups as described in § 412.515. To ensure that Medicare patients who are classified to each LTC-DRG have access to services and to encourage efficiency, we calculate a relative weight for each LTC-DRG that represents the resources needed by an average inpatient LTCH case in that LTC-DRG. For example, cases in a LTC-DRG with a relative weight of 2 will, on average, cost twice as much as cases in a LTC-DRG with a weight of 1.

As we discussed in the FY 2007 IPPS final rule, the LTC-DRG relative weights effective under the LTCH PPS for FY 2007 were calculated using the March 2006 update of FY 2005 MedPAR data and Version 24.0 of the GROUPER software (71 FR 47973). We use total days and total charges in the calculation of the LTC-DRG relative weights.

LTCHs often specialize in certain areas, such as ventilator-dependent patients and rehabilitation or wound care. Some case types (DRGs) may be treated, to a large extent, in hospitals that have (from a perspective of charges) relatively high (or low) charges. Distribution of cases with relatively high (or low) charges in specific LTC-DRGs has the potential to inappropriately distort the measure of average charges. To account for the fact that cases may not be randomly distributed across LTCHs, we use a hospital-specific relative value method to calculate relative weights. We believe this method removes this hospital-specific source of bias in measuring Start Printed Page 26878average charges. Specifically, we reduce the impact of the variation in charges across providers on any particular LTC-DRG relative weight by converting each LTCH's charge for a case to a relative value based on that LTCH's average charge. (See the FY 2007 IPPS final rule for further information on the application of the hospital-specific relative value methodology under the LTCH PPS (71 FR 47974 through 47975).)

To account for LTC-DRGs with low volume (that is, with fewer than 25 LTCH cases), we grouped those low volume LTC-DRGs into 1 of 5 categories (quintiles) based on average charges, for the purposes of determining relative weights. For FY 2007 based on the FY 2005 MedPAR data, we identified 180 LTC-DRGs that contained between 1 and 24 cases. This list of low volume LTC-DRGs was then divided into 1 of the 5 low volume quintiles, each containing 36 LTC-DRGs (180/5 = 36). Each of the low volume LTC-DRGs grouped to a specific quintile received the same relative weight and ALOS using the formula applied to the regular LTC-DRGs (25 or more cases). (See the FY 2007 IPPS final rule for further explanation of the development and composition of each of the 5 low volume quintiles for FY 2007 and their composition (71 FR 47975 through 47978).)

After grouping the cases in the appropriate LTC-DRG, we calculated the relative weights by first removing statistical outliers and cases with a LOS of 7 days or less. Next, we adjusted the number of cases remaining in each LTC-DRG for the effect of SSO cases under § 412.529. The short-stay adjusted discharges and corresponding charges were used to calculate “relative adjusted weights” in each LTC-DRG using the hospital-specific relative value method. We also adjusted the LTC-DRG relative weights to account for nonmonotonically increasing relative weights. That is, we made an adjustment if cases classified to the LTC-DRG “with CCs” of a “with CC”/“without CC” pair had a lower average charge than the corresponding LTC-DRG “without CCs” by assigning the same weight to both LTC-DRGs in the “with CC”/“without CC” pair. (See the FY 2007 IPPS final rule for further details on the steps for calculating the LTC-DRG relative weights (71 FR 47978 through 47984).)

In addition, of the 538 LTC-DRGs in the LTCH PPS for FY 2007, based on LTCH cases in the FY 2005 MedPAR files, we identified 183 LTC-DRGs for which there were no LTCH cases in the database. That is, no patients who would have been classified to those DRGs were treated in LTCHs during FY 2005, and therefore, no charge data were reported for those DRGs. Thus, in the process of determining the relative weights of LTC-DRGs, we were unable to determine weights for these 183 LTC-DRGs using the method described in this section of the preamble. However, since patients with a number of the diagnoses under these LTC-DRGs may be treated at LTCHs beginning in FY 2007, we assigned relative weights to each of the 183 “no volume” LTC-DRGs based on clinical similarity and relative costliness to one of the remaining 355 (538-183 = 355) LTC-DRGs for which we were able to determine relative weights, based on the FY 2005 claims data. (A list of the current no-volume LTC-DRGs and further explanation of their FY 2007 relative weight assignment can be found in the FY 2007 IPPS final rule (71 FR 47980 through 47984).)

Furthermore, for FY 2007, we established LTC-DRG relative weights of 0.0000 for heart, kidney, liver/intestinal, lung, simultaneous pancreas/kidney, and pancreas transplants (LTC-DRGs 103, 302, 480, 495, 512 and 513, respectively) because presently no LTCH meets the applicable requirements to perform Medicare covered transplant procedures. However, if in the future, a LTCH seeks to meet such requirements as a Medicare-approved transplant center to perform Medicare-covered transplant procedures, we believe that the application and approval procedure would allow sufficient time for us to propose appropriate weights for the LTC-DRGs affected. At the present time, we included these 6 transplant LTC-DRGs in the GROUPER software program for administrative purposes. As the LTCH PPS uses the same GROUPER software program for LTCHs as is used under the IPPS, removing these DRGs would be administratively burdensome.

As we noted previously in this section, there were no new ICD-9-CM code requests for an April 1, 2007 update. Therefore, Version 24.0 of the DRG GROUPER software established in the FY 2007 IPPS final rule will continue to be effective until October 1, 2007. Moreover, the LTC-DRGs and relative weights for FY 2007 established in Table 11 of that same IPPS final rule (71 FR 48321 through 48331) will continue to be effective until October 1, 2007, (just as they would have been even if there had been any new ICD-9-CM code requests for an April 1, 2007 update). Accordingly, Table 3 in the Addendum to this final rule lists the LTC-DRGs and their respective relative weights, geometric ALOS, and five-sixths of the geometric ALOS that we will continue to use for the period of July 1, 2007 through September 30, 2007. (This table is the same as Table 11 of the Addendum to the FY 2007 IPPS final rule.) The next update to the ICD-9-CM coding system will be presented in the FY 2008 IPPS proposed rule (since there will be no April 1, 2007 updates to the ICD-9-CM coding system).

In addition, the proposed DRGs and GROUPER for FY 2008 that would be effective October 1, 2007, will be presented in the IPPS FY 2008 proposed rule. Below we provide a summary of the development of the proposed LTC-DRG relative weights for FY 2008 presented in that same proposed rule. To calculate the proposed MS-LTC-DRG relative weights for FY 2008 in the FY 2008 IPPS proposed rule, we obtained total Medicare allowable charges from FY 2006 Medicare LTCH bill data from the December 2006 update of the MedPAR file, which are the best available data at this time, and we used the proposed Version 25.0 of the CMS GROUPER used under the IPPS (as discussed in section II.B. of the preamble of that proposed rule) to classify cases. To calculate the final MS-LTC-DRG relative weights for FY 2008, we proposed that, if more recent data are available (for example, data from the March 2007 update of the MedPAR file), we would use those data and the finalized Version 25.0 of the CMS GROUPER used under the IPPS. We continued to use total days and total charges in the calculation of the proposed MS-LTC-DRG relative weights. We also continued to use the hospital-specific relative value methodology, described above, for determining the proposed MS-LTC-DRG relative weights for FY 2008.

As noted above in this section, although the proposed adoption of the MS-LTC-DRGs would result in some modifications of existing procedures discussed above for assigning relative weights under the current system (as discussed in detail below), the basic methodology for developing the proposed FY 2008 MS-LTC-DRG relative weights in the FY 2008 IPPS proposed rule continue to be determined in accordance with the general methodology established in the August 30, 2002 LTCH PPS final rule (67 FR 55989 through 55991) summarized above. With the implementation of the LTCH PPS for FY 2003, we established a procedure to address setting relative weights for LTC-DRG “pairs” that were differentiated on the presence or Start Printed Page 26879absence of CCs (71 FR 47979). As discussed in the FY 2008 IPPS proposed rule, our proposal to adopt a severity-based patient classification system for the LTCH PPS, the MS-LTC-DRGs described above, required us to adapt our existing approach for setting relative weights for the severity levels within a specific base DRG. We are also proposed to modify our existing methodology for maintaining monotonicity when setting relative weights for the proposed MS-LTC-DRGs.

As under the existing procedure, under the proposed MS-LTC-DRGs, for purposes of the annual setting of the relative weights, there continue to be three different categories of DRGs based on volume of cases within specific LTC-DRGs. LTC-DRGs with at least 25 cases are each assigned a relative weight; low-volume proposed MS-LTC-DRGs (that is, proposed MS-LTC-DRGs that contain between 1 and 24 cases annually) are grouped into quintiles (described below) and assigned the weight of the quintile. Cases with no-volume proposed MS-LTC-DRGs (that is, no cases in the database were assigned to those proposed MS-LTC-DRGs) are cross-walked to other proposed MS-LTC-DRGs based on the clinical similarities and assigned the weight of the quintile that is closest to the relative weight of the cross-walked proposed MS-LTC-DRG. (For in-depth discussions of our proposals regarding proposed relative weight setting for low-volume MS-LTC-DRGs and for no-volume MS-LTC-DRGs, see the FY 2008 IPPS proposed rule.)

As noted above, for FY 2008, we are proposing to adopt the MS-DRGs for use in both the LTCH PPS and the IPPS. While the LTCH PPS and the IPPS use the same patient classification system, the methodology that is used to set the DRG weights for use in each payment system differs because the overall volume of cases in the LTCH PPS is much less than in the IPPS. As a general rule, as described in the FY 2008 IPPS proposed rule, we are proposing to set the weights for the proposed MS-LTC-DRGs using the following steps: (1) If an MS-LTC-DRG has at least 25 cases, it is assigned its own relative weight; (2) if an MS-LTC-DRGs has between 1 and 24 cases, it is assigned to a quintile to which we will assign a relative weight; and (3) if an MS-LTC-DRG has no cases, it is cross-walked to another DRG based upon clinical similarities and assigned the appropriate relative weight. Theoretically, as with the existing LTC-DRG system, cases under the proposed MS-LTC-DRG system that are more severe require greater expenditure of medical care resources and will result in higher average charges. Therefore, in the three severity levels of the base MS-LTC-DRG, relative weights should increase with severity, from lowest to highest. If the relative weights do not increase (that is, if based on the relative weight calculation using the most recent LTCH claims data, a proposed MS-LTC-DRG with MCC would have a lower relative weight than one with CC, or the DRG without CC/MCC would have a higher relative weight than either of the others), there is a problem with monotonicity.

As discussed above in this section, to account for LTC-DRGs with low volume (that is, with fewer than 25 LTCH cases), we group those “low-volume LTC-DRGs” (that is, DRGs that contained between 1 and 24 cases annually) into one of five categories (quintiles) based on average charges, for the purposes of determining relative weights. As discussed in the FY 2008 IPPS proposed rule, we proposed to continue to employ this treatment of low-volume proposed MS-LTC-DRGs with a modification to combine proposed MS-LTC-DRGs for the purpose of computing a relative weight in cases where necessary to maintain monotonicity in determining the proposed FY 2008 MS-LTC-DRG relative weights using the best available LTCH data. In that proposed rule, using LTCH cases from the December 2006 update of the FY 2006 MedPAR file, we identified 307 proposed MS-LTC-DRGs that contained between 1 and 24 cases. This list of proposed MS-LTC-DRGs was then divided into one of the 5 low-volume quintiles, each containing a minimum of 61 proposed MS-LTC-DRGs (307/5 = 61, with a remainder of 2 proposed MS-LTC-DRGs). Consistent with our current methodology, we are proposing to make an assignment to a specific low-volume quintile by sorting the low-volume proposed MS-LTC-DRGs in ascending order by average charge. (See the FY 2008 IPPS proposed rule for further explanation of the development and composition of each of the 5 low volume quintiles for FY 2007 and their proposed composition.)

As we noted previously, although the proposed adoption of the MS-LTC-DRGs would result in some modifications of existing procedures for assigning relative weights, the proposed FY 2008 MS-LTC-DRG relative weights presented in Table 11 of the FY 2008 IPPS proposed rule are based on the methodology established in the August 30, 2002 LTCH PPS final rule (67 FR 55989 through 55991). In summary, as described in greater detail in that same proposed rule, LTCH cases would be grouped to the appropriate proposed MS-LTC-DRG, while taking into account the low-volume proposed MS-LTC-DRGs as described above, before the proposed FY 2008 MS-LTC-DRG relative weights can be determined. After grouping the cases to the appropriate proposed MS-LTC-DRG, we proposed to calculate the proposed relative weights for FY 2008 by first removing statistical outliers and cases with a LOS of 7 days or less and to adjust the number of cases in each proposed MS-LTC-DRG for the effect of SSO cases under § 412.529. The short-stay adjusted discharges and corresponding charges are used to calculate “relative adjusted weights” in each proposed MS-LTC-DRG using the HSRV method described above.

Next we proposed to determine relative weights for the no-volume proposed MS-LTC-DRGs. As discussed in the FY 2008 IPPS proposed rule, of the 745 proposed MS-LTC-DRGs for FY 2008, we identified 124 proposed MS-LTC-DRGs for which there were no LTCH cases in the database. That is, no patients who would have been classified to those proposed MS-LTC-DRGs were treated in LTCHs during FY 2006, and therefore, no charge data were reported for those proposed MS-LTC-DRGs. Thus, in the process of determining the proposed MS-LTC-DRG relative weights, we are unable to determine weights for these 124 proposed MS-LTC-DRGs using the methodology described above. However, because patients with a number of the diagnoses under these proposed MS-LTC-DRGs may be treated at LTCHs beginning in FY 2008, we are proposing to assign relative weights to each of the 124 no-volume proposed MS-LTC-DRGs based on clinical similarity and relative costliness to one of the remaining 621 (745-124 = 621) proposed MS-LTC-DRGs for which we are able to determine proposed relative weights, based on FY 2006 LTCH claims data. In general, we determined proposed relative weights for the 124 proposed MS-LTC-DRGs with no LTCH cases in the FY 2006 MedPAR file used in this proposed rule by cross-walking these proposed MS-LTC-DRGs to other proposed MS-LTC-DRGs and then grouping them to the appropriate proposed low-volume quintile. (A list of the proposed no-volume MS-LTC-DRGs and further explanation of their proposed FY 2008 relative weight assignment can be found in the FY 2008 IPPS proposed rule.) We also adjusted the proposed MS-LTC-DRG relative weights to account for nonmonotonically increasing relative weights, including any no volume Start Printed Page 26880proposed MS-LTC-DRGs, where applicable, as described above.

Furthermore, for FY 2008 we proposed to establish proposed MS-LTC-DRG relative weights of 0.0000 for the following transplant proposed MS-LTC-DRGs: Heart transplant or implant of heart assist system w MCC (proposed MS-LTC-DRG 1); Heart transplant or implant of heart assist system w/o MCC (proposed MS-LTC-DRG 2); Liver transplant w MCC or intestinal transplant (proposed MS-LTC-DRG 5); Liver transplant w/o MCC (proposed MS-LTC-DRG 6); Lung transplant (proposed MS-LTC-DRG 7); Simultaneous pancreas/kidney transplant (proposed MS-LTC-DRG 8); and Pancreas transplant (proposed MS-LTC-DRG 10). As explained in the FY 2008 IPPS proposed rule, this is because Medicare will only cover these procedures if they are performed at a hospital that has been certified for the specific procedures by Medicare and presently no LTCH has been so certified. If in the future a LTCH applies for certification as a Medicare-approved transplant center, we believe that the application and approval procedure would allow sufficient time for us to determine appropriate weights for the proposed MS-LTC-DRGs affected. At the present time, we would only include these seven proposed transplant MS-LTC-DRGs in the GROUPER program for administrative purposes only. Because we use the same GROUPER program for LTCHs as is used under the IPPS, removing these proposed MS-LTC-DRGs would be administratively burdensome. (See the FY 2008 IPPS proposed rule for further details on the steps for calculating the proposed MS-LTC-DRG relative weights for FY 2008.)

3. Budget Neutrality (BN) Requirement for the Annual LTC-DRG Update

As noted above in this section, currently under § 412.517, the LTC-DRG classifications and relative weights are adjusted annually to reflect changes in factors affecting the relative use of LTCH resources, such as treatment patterns, technology and number of discharges. Currently, there are no statutory or regulatory requirements that the annual update to the LTC-DRG classifications and relative weights be done in a budget neutral manner. Historically, since the initial implementation of the LTCH PPS in FY 2003, we have updated the LTC-DRG relative weights each year without a BN adjustment based on the most recent available LTCH claims data, which reflect current LTCH patient mix and coding practices, and appropriately reflected more or less resource use than the previous year's LTC-DRG relative weights (71 FR 47991). When we proposed changes to the LTC-DRGs for FY 2007 in the FY 2007 IPPS proposed rule, we estimated that those proposed changes to the LTC-DRG classifications and relative weights would result in about an estimated 1.4 percent decrease in estimated aggregate LTCH PPS payments (71 FR 24413). As we discussed in the FY 2007 IPPS final rule (71 FR 47991), several commenters, including MedPAC, urged us to establish a BN requirement for the annual reclassification and recalibration of the LTC-DRGs so that, in future years, the LTCH PPS could avoid an estimated decrease in estimated aggregate payments, such as the estimated 1.4 percent decrease that resulted from the proposed update to the LTC-DRGs and relative weights for FY 2007. In response to previous proposed annual updates to the LTC-DRG relative weights, we also received comments recommending that a BN adjustment be applied in determining the LTC-DRG relative weights to mitigate LTCH PPS payment fluctuations. (See the FY 2005 IPPS final rule (69 FR 48999 through 49000), and the FY 2006 IPPS final rule (70 FR 47333 through 47334).)

In response to those comments, we explained that we understood the commenters' concern with the estimated decrease in payments under LTCH PPS based upon the changes in the LTC-DRGs and relative weights proposed for FY 2007. However, as we discussed in the FY 2007 IPPS final rule, we did not postpone the proposed FY 2007 reclassification and recalibration of the LTC-DRGs, nor did we implement those changes in a budget neutral manner. We noted several reasons for the annual fluctuations in LTC-DRG relative weights that have resulted in both estimated increases and decreases in estimated aggregate LTCH PPS payments in the 4 years since the implementation of the LTCH PPS in FY 2003. Specifically, we reiterated our belief that several factors have affected the changes to the LTC-DRG relative weights over the past 4 years, including actual improvements in coding so that cases are appropriately assigned to LTC-DRGs. We also explained that historically we recalibrated the LTC-DRG relative weights each year based on the most recent available LTCH claims data, which reflect current LTCH patient mix and coding practices, and appropriately reflects more or less resource use than the previous year's LTC-DRG relative weights. The intended purpose of the annual recalibration of the LTC-DRG relative weights is to reflect any variation in coding practices and charges from the previous year and to help ensure that the LTC-DRG relative weights in the upcoming fiscal year will result in appropriate and accurate payments to LTCHs for the resources they expend to treat their Medicare patients. (71 FR 47984 through 47989)

We also reminded the commenters that under the IPPS, there is a statutory requirement that the annual DRG reclassification and recalibration changes be made in a manner that assures that the estimated aggregate payments are neither greater than nor less than the estimated aggregate payments that would have been made without the changes, but there is no corresponding statutory requirement under the LTCH PPS. However, we noted that, given the considerable discretion granted to the Secretary under section 123 of the BBRA and section 307(b) of the BIPA of 2000 to develop the LTCH PPS, it is possible that, at some point, the Secretary would consider using this broad authority to establish a BN policy for the annual update of the LTC-DRG classifications and relative weights. We further stated that if we find that it would be appropriate to propose making the updates to the LTC-DRGs and relative weights in a budget neutral manner, the public would have the opportunity to submit comments on any proposed change during the rulemaking process.

As we discussed in the RY 2007 LTCH PPS proposed rule (72 FR 4784 through 4786), a LTCH's case-mix index (CMI) is defined as its case weighted average LTC-DRG relative weight for all its discharges in a given period. Changes in CMI consist of two components: “real” CMI changes and “apparent” CMI changes. Real CMI increase is defined as the increase in the average LTC-DRG relative weights resulting from the hospital's treatment of more resource intensive patients. Apparent CMI increase is defined as the increase in CMI due to changes in coding practices. The computed (or observed) CMI increase is defined as real CMI increase (due to an increase in patient severity) plus the increase due to changes in coding practices (including better documentation of the medical record by physicians and more complete coding of the medical record by coders). If LTCH patients have more costly impairments, lower functional status, or increased comorbidities, and thus require more resources in the LTCH, we consider this a real change in case-mix. Conversely, if LTCH patients have the same impairments, functional status, and Start Printed Page 26881comorbidities but are coded differently resulting in higher payment, we consider this an apparent change in case-mix. We believe that changes in payment rates, including the LTC-DRG relative weights, should accurately reflect changes in LTCHs' true cost of treating patients (real CMI increase), and should not be influenced by changes in coding practices (apparent CMI increase).

As stated above in this section, apparent CMI increase results from cases being grouped to a LTC-DRG with a higher weight than it would be without such changes in coding practices. As we discussed in the FY 2007 IPPS final rule (71 FR 48343 through 48344), in discussing the impact of the changes to the LTC-DRG classifications and relative weights established for FY 2007 that were estimated to result in an aggregate decrease in LTCH PPS payments of approximately 1.3 percent, we explained that changes in coding practices (rather than patient severity) primarily resulted in fluctuations in the LTC-DRG relative weights in the past. Specifically, based on an analysis of FY 2005 LTCH claims data, we continued to observe that the average LTC-DRG relative weight decreases due to an increase of relatively lower charge cases being assigned to LTC-DRGs with higher relative weights in the prior year. Contributing to this increase in these relatively lower charge cases being assigned to LTC-DRGs with higher relative weights in the prior year are improvements in coding practices, which are typical when moving from a reasonable cost-based payment system to a PPS. The impact of including cases with relatively lower charges into LTC-DRGs that had a relatively higher relative weight in the previous version of the GROUPER software is a decrease in the average relative weight for those LTC-DRGs in the updated version of the GROUPER software.

We noted in the RY 2008 LTCH PPS proposed rule (72 FR 4785) that this same phenomenon of relatively lower charge cases being assigned to LTC-DRGs with higher relative weights in the prior year was also observed when we analyzed the LTCH claims data from FY 2003 and FY 2004 to update the LTC-DRG relative weights for FY 2005 and FY 2006, respectively (see the FY 2005 IPPS final rule (69 FR 48999) and the FY 2006 IPPS final rule (70 FR 47701 through 47702).) However, this phenomenon was more notable based on the FY 2004 LTCH claims data that were used to update the LTC-DRG relative weights for FY 2006, where the changes to the LTC-DRG weights established were estimated to result in a decrease in aggregate LTCH PPS payments of 4.2 percent (as compared to the estimated 1.3 percent decrease in aggregate LTCH PPS payments based on the FY 2005 LTCH claims data used to determine the FY 2007 LTC-DRG relative weights). Because the estimated decrease in aggregate LTCH PPS payments due to the update to the LTC-DRG relative weights based on more recent (FY 2005) LTCH claims data was significantly lower (1.3 percent estimated based on the LTC-DRG changes for FY 2007) than it was based on FY 2004 LTCH claims data (4.2 percent estimated based on the LTC-DRG changes for FY 2006), we believe that, as LTCHs have become more familiar with the ICD-9-CM coding principles and guidelines used under a DRG-based system, annual changes in LTCH CMI are approaching the point where the observed CMI increase is primarily due to changes in real CMI (that is, increased patient severity) rather than apparent CMI (that is, changes in coding practices). In other words, because we have observed that, over time as LTCHs have gained more experience with ICD-9-CM coding, estimated changes in LTCH PPS payments due to recalibration of the LTC-DRG relative weights based on more recent claims data (for example, the FY 2007 LTC-DRG relative weights calculated from FY 2005 LTCH claims data as compared to the FY 2006 LTC-DRG relative weights calculated from FY 2004 LTCH claims data) have diminished over time. That is, we have estimated smaller fluctuations in aggregate LTCH PPS payments as a result of the annual recalibration of the LTC-DRG relative weights based on more recent LTCH claims data generated after the implementation of the LTCH PPS (for example, the 1.3 percent estimated decrease in aggregate LTCH PPS payments for FY 2007 based on FY 2004 LTCH claims data as compared to the 4.2 percent estimated decrease in aggregate LTCH PPS payments for FY 2007 based on FY 2005 LTCH claims data).

For these reasons, as discussed in the RY 2008 LTCH PPS proposed rule (72 FR 4785), we believe that LTCH coding practices have stabilized such that the most recent available LTCH claims data now primarily reflect changes in the resources used by the average LTCH patient in a particular LTC-DRG (and not changes in coding practices). Thus, we believe that the most recent available data (as described below in this section) mainly reflect the true costs of treating LTCH patients, and we believe changes in payment rates, including the LTC-DRGs, should reflect such costs. Furthermore, in that same proposed rule, we explained that a LTCH CMI analysis based on the most recent available LTCH claims data, which is discussed in section IV.C. of this preamble, also supports our belief that observed CMI increase is primarily due to changes in real CMI (that is, increased patient severity) rather than apparent CMI (that is, changes in coding practices). Specifically, this CMI analysis indicates that changes in LTCH coding practices, which resulted in fluctuations in the LTC-DRG relative weights in the past, appear to be stabilizing as LTCHs have become more familiar with a DRG-based system.

Specifically, this LTCH CMI analysis shows that the overall observed change in LTCH CMI from FY 2003 compared to FY 2004 was an increase of approximately 6.75 percent while the overall observed change in LTCH CMI from FY 2004 compared to FY 2005 was an increase of approximately 3.49 percent, which is only about half of the LTCH CMI growth measured from the prior period (that is, the 6.75 percent from FY 2003 to FY 2004). Furthermore, preliminary analysis of FY 2006 LTCH claims data, which reflects over 3 full years of experience under the LTCH PPS for most LTCHs, showed an even smaller overall observed CMI increase of about 1.9 percent from FY 2005 compared to FY 2006. Again, the observed CMI increase from FY 2005 to FY 2006 is only about half of the LTCH CMI growth measured from the prior period (that is, the 3.49 percent from FY 2004 to FY 2005). Because this LTCH CMI analysis shows that observed CMI is declining, we believe that LTCH coding practices have stabilized such that changes in LTCH CMI are now primarily due to changes in real CMI (that is, increased patient severity) rather than apparent CMI (that is, changes in coding practices). In other words, because we believe that the observed annual CMI increase is primarily “real” and not “apparent,” it is no longer necessary to update the LTC-DRGs in a non-budget neutral manner (as discussed in greater detail below in this section). As stated above in this section, we believe that changes in payment rates, including the LTC-DRG relative weights, should accurately reflect changes in LTCHs' true cost of treating patients (real CMI increase) and should not be influenced by changes in coding practices (apparent CMI increase).

In light of these facts, in order to mitigate estimated fluctuations in estimated aggregate LTCH PPS Start Printed Page 26882payments, as urged by past commenters, we stated in the RY 2008 proposed rule (72 FR 4785) that we had given further consideration to the issue of establishing a BN requirement for annual LTC-DRG reclassification and recalibration. Therefore, in that proposed rule, under the broad authority conferred upon the Secretary under section 123 of the BBRA as amended by section 307(b) of the BIPA to develop the LTCH PPS, we proposed that, beginning with the LTC-DRG update for FY 2008, the annual update to the LTC-DRG classifications and relative weights would be done in a budget neutral manner such that estimated aggregate LTCH PPS payments would be unaffected, that is, would be neither greater than nor less than the estimated aggregate LTCH PPS payments that would have been made without the LTC-DRG classification and relative weight changes. Accordingly, we proposed to revise § 412.517 to specify that annual changes to the LTC-DRG classifications and the recalibration of the LTC-DRG relative weights would be made in a budget neutral manner such that estimated aggregate LTCH PPS payments are not affected.

Comment: Numerous commenters, including MedPAC, supported our proposal to recalibrate the LTC-DRGs annually in a budget neutral manner. Some commenters also recommended that we should monitor the recalibration so that any reweighting of the LTC-DRGs is conducted in a manner that does not result in a redistribution of payments from high acuity DRGs to lower acuity DRGs, pending implementation of revised certification criteria designed to screen out LTCH inappropriate patients.

Response: We appreciate the commenters' support of our proposed BN requirement for the annual LTC-DRG update. As discussed in the RY 2008 LTCH PPS proposed rule (72 FR 4785 through 4786), we explained that we believe that it would be appropriate to update the LTC-DRG classifications and relative weights in a budget neutral manner at this time for the reasons discussed below. As noted above in this section, the relative weight for each LTC-DRG represents the resources needed by an average inpatient LTCH case in that LTC-DRG, such that LTCH cases in a LTC-DRG with a relative weight of 2 will, on average, cost twice as much as cases in a LTC-DRG with a relative weight of 1.

In the past when we recalibrated the LTC-DRG relative weights each year without a BN adjustment based on the most recent available LTCH claims data, we believe that the resulting LTC-DRG relative weights appropriately reflected more or less resource use than the previous year's LTC-DRG relative weights, and that the estimated aggregate payment changes were appropriate given that the LTCH claims data used to determine those LTC-DRG relative weights reflected changes in coding practices, as well as changes in actual resource use. Historically, we have not updated the LTC-DRGs in a budget neutral manner because we believed that past fluctuations in the LTC-DRG relative weights were primarily due to changes in LTCH coding practices, which included both “real” and “apparent” changes in LTCHs' case-mix (as discussed above in this section). We believe that changes in the LTCH PPS payment rates, including the LTC-DRG relative weights, should accurately reflect changes in LTCHs' true cost of treating patients (real CMI increase), and should not be influenced by changes in coding practices (apparent CMI increase). Therefore, in the past we did not update the LTC-DRGs in a budget neutral manner so that “apparent” CMI changes were not permanently built into the LTCH PPS payment rates.

Because LTCH 2006 claims data does not appear to significantly reflect changes in LTCH coding practices in response to the implementation of the LTCH PPS (as explained above in this section), we believe that it may be appropriate to update the LTC-DRGs so that estimated aggregate LTCH PPS payments would neither increase or decrease since we believe that changes in the LTC-DRG classifications and relative weights should accurately reflect changes in LTCHs' resource use (that is, true cost of treating patients) and should not be influenced by changes in coding practices, and that the most recent such LTCH claims data primarily reflects changes in the resources needed by an average LTCH case in a particular LTC-DRG (and not changes in coding practices).

Thus, we now believe it would be reasonable and appropriate to update the LTC-DRGs in a budget neutral manner, beginning in FY 2008, so that estimated aggregate payments under the LTCH PPS would be unaffected (that is, estimated aggregate LTCH PPS payments would not be greater than or less than they would have been without the proposed LTC-DRG classification and relative weight changes) by any changes resulting from the annual reclassification and recalibration of the LTC-DRGs. Updating the LTC-DRGs in a budget neutral manner would result in an annual update to the individual LTC-DRG classifications and relative weights based on the most recent available data to reflect changes in relative LTCH resource use; however, the LTC-DRG relative weights would be uniformly adjusted to ensure that estimated aggregate payments under the LTCH PPS would not be affected (that is, decreased or increased).

In this final rule, under the broad authority conferred upon the Secretary under section 123 of the BBRA as amended by section 307(b) of the BIPA to develop the LTCH PPS, beginning with the LTC-DRG update for FY 2008 (discussed in greater detail below), the annual update to the LTC-DRG classifications and relative weights will be done in a budget neutral manner such that estimated aggregate LTCH PPS payments will be unaffected, that is, will be neither greater than nor less than the estimated aggregate LTCH PPS payments that would have been made without the LTC-DRG classification and relative weight changes. Accordingly, we are revising § 412.517 to specify that annual changes to the LTC-DRG classifications and the recalibration of the LTC-DRG relative weights are made in a budget neutral manner such that estimated aggregate LTCH PPS payments are not affected.

As discussed above, we believe that the most recent available LTCH claims data reflects the intensity of resource use of the treatment of Medicare patients based on current LTCH coding and treatment practices. Accordingly, we believe that annually updating the LTC-DRG relative weights using the most recent available LTCH claims data reflects more or less resource use than the previous year's LTC-DRG relative weights based on the current LTCH practices. Therefore, we believe that any redistribution in payments as a result of the annual recalibration of the LTC-DRG relative weights based on this updated LTCH claims data appropriately reflects LTCH resource use in the treatment of their Medicare patients. While we will continue to monitor LTCH data, including any redistribution of payments upon the annual update of the LTC DRGs, for the reasons discussed above, we are not adopting the commenters' suggestion to establish a requirement that the annual recalibration of the relative weights be done in a manner that would adjust for redistribution of payments from high acuity LTC-DRGs to lower acuity LTC-DRGs.

As we explained in the RY 2008 LTCH PPS proposed rule (72 FR 4786), we intend to update the LTC-DRG classifications and relative weights for FY 2008 based on the best available data Start Printed Page 26883at the time to allow for changes in factors affecting hospital resource use, including but not limited to, practice patterns and new technology. This will be done in a budget neutral manner, such that estimated aggregate payments under the LTCH PPS would neither decrease or increase as a result of the changes due to the annual reclassification and recalibration of the LTC-DRGs. Because we will continue to use the most recent available LTCH data, the updated LTC-DRG relative weights will continue to reflect changes in LTCH resource use (as is the case under the current (non-budget neutral) LTC-DRG update methodology). Thus, for example, if the most recent LTCH claims data showed that the resource use for hypothetical LTC-DRG “ABC” is double the resource use for hypothetical LTC-DRG “XYZ,” then the value of the relative weight for LTC-DRG “ABC” would be about twice the value of relative weight for LTC-DRG “XYZ.”

In addition to accounting for changes in relative resource use, to include a BN requirement for the annual update to the LTC-DRGs, the updated LTC-DRG relative weights will need to be uniformly adjusted to ensure that estimated aggregate LTCH PPS payments will not be affected. That is, a BN factor will need to be computed to ensure that the LTC-DRG reclassification and recalibration process, by itself, neither increases nor decreases estimated aggregate LTCH PPS payments.

As discussed in the FY 2008 IPPS proposed rule, to accomplish BN when annually updating the LTC-DRG classifications and relative weights under revised § 412.517, we proposed to use a method that is similar to the methodology used under the IPPS. (Information on the IPPS DRG BN adjustment can be found in the FY 2007 IPPS final rule (71 FR 47970).) As noted above, we proposed to adopt the MS-LTC-DRGs for the LTCH PPS for FY 2008. Therefore, in the discussion that follows, we will refer to the development of the proposed budget neutrality factor in terms of the proposed MS-LTC-DRG severity-weighted patient classification system. Specifically, after recalibrating the proposed MS-LTC-DRG relative weights, as we do under our existing methodology (as described in detail in the FY 2007 IPPS final rule (71 FR 47978 through 47981)), as described in greater detail in the FY 2008 IPPS proposed rule, we would calculate and apply a normalization factor (which will be published annually in the IPPS proposed and final rules when we update the LTC-DRGs and relative weights) to the proposed MS-LTC-DRG relative weights to ensure that estimated aggregate LTCH PPS payments are not influenced by changes in the composition of case types or changes made to the classification system. That is, the normalization adjustment is intended to ensure that the recalibration of the proposed MS-LTC-DRG relative weights (that is, the process itself) neither increases nor decreases total estimated payments. To calculate the normalization factor, we proposed to use the most recent available claims data (FY 2006) and apply the proposed GROUPER (Version 25.0) to calculate the proposed MS-LTC-DRG relative weights. (We also proposed to use the most recent available claims data in the analysis for this final rule.) These weights were determined such that the average CMI value is 1.0. Then, we proposed to group the same claims data (FY 2006) using the current GROUPER (Version 24.0) and current LTC-DRG relative weights. The average CMI was calculated for the claims data using the current GROUPER and relative weights. Finally, the ratio of the average CMI of the claims data set under the current GROUPER and the proposed GROUPER was calculated as the proposed normalization factor.

For FY 2008, based on the latest available data, the proposed normalization factor is estimated as 1.020302, which was applied to each proposed MS-LTC-DRG relative weight. (We also stated that if more current data become available prior to publication of the final rule, we will use those data to determine the normalization factor.) That is, each proposed MS-LTC-DRG relative weight was multiplied by 1.020302 in the first step of the BN process.

We are also proposed to ensure that estimated aggregate LTCH PPS payments (based on the most recent available LTCH claims data) after recalibration (the proposed relative weights) would be equal to estimated aggregate LTCH PPS payments (for the same most recent available LTCH claims data) before recalibration (the existing relative weights). Therefore, we proposed to calculate the BN adjustment factor by simulating estimated payments under both sets of GROUPERs and relative weights. We proposed to simulate total estimated payments under the current payment policies (RY 2007) using the most recent available claims data (FY 2006) and using the proposed GROUPER (Version 25.0), and normalized relative weights. Then, we proposed to simulate estimated payments using the most recent available claims data (FY 2006) and apply the proposed GROUPER (Version 25.0). We next calculated payments using the same claims data (FY 2006) with the current GROUPER (Version 24.0). The ratio of the estimated average payment under the current GROUPER and the proposed GROUPER was calculated as the proposed BN factor. Then each of the proposed normalized relative weights was multiplied by the proposed BN factor to determine the proposed budget neutral relative weight for each proposed MS-LTC-DRG. Accordingly, based on the most recent available data, we proposed to apply a BN factor of 1.003924 to the relative weights after normalizing. To calculate the proposed MS-LTC-DRG relative weights for FY 2008, we obtained total Medicare allowable charges from FY 2006 Medicare LTCH bill data from the December 2006 update of the MedPAR file, which are the best available data at that time. We also proposed that if more current data become available prior to publication of the final rule, we will use those data to determine the budget neutrality factor. The proposed FY 2008 MS-LTC-DRG relative weights are presented in Table 11 in the Addendum of the FY 2008 IPPS proposed rule, which reflect the budget neutral adjustment described above.

In the recently issued FY 2008 IPPS proposed rule, we proposed significant refinements to the DRGs used under both the IPPS and LTCH PPS to better recognize severity of illness among patients. The proposed refinements would be effective October 1, 2007. The proposed new MS-DRG and MS-LTC-DRG systems present opportunities to acute care hospitals and LTCHs, respectively, to improve documentation and coding to receive higher payments without a real increase in patient severity of illness. The Office of the Actuary estimates an adjustment of −2.4 percent to the IPPS rates for each of FY 2008 and FY 2009 will be necessary to account for the anticipated improvements in coding and documentation. In the FY 2008 IPPS proposed rule, we proposed to apply this −2.4 percent adjustment for case mix increase in FY 2008 and in FY 2009 in both the IPPS and LTCH PPS systems to address the proposed change to the refined severity DRGs. It should be noted that this adjustment is not related to the finalized budget neutrality adjustment included in this LTCH final rule and discussed above. The budget neutrality adjustment in this rule is an annual requirement that is needed to assure that annual recalibration of the DRG weights based on the most recent available claims data, results in no Start Printed Page 26884changes (increase or decrease) in estimated payments that stem from updating the DRG weights, while the proposed −2.4 percent adjustment for FYs 2008 and 2009 is tied solely to the proposed change to the MS-LTC-DRGs. Accordingly, each of the proposed MS-LTC-DRG relative weights in Table 11 of the Addendum to the FY 2008 IPPS proposed rule reflects this proposed adjustment. That is, each proposed MS-LTC-DRG relative weight was multiplied by a factor of 0.976 to account for changes in coding or classification of discharges resulting from the proposed adoption of the new patient classification system. This proposed adjustment is consistent with the proposed adjustment applied to the proposed IPPS rates for FYs 2008 and 2009 to eliminate the effect of changes in coding or classification of discharges that do not reflect real change in case-mix because we believe that adoption of the proposed MS-LTC-DRGs would create a risk of increased aggregate levels of payment as a result of increased documentation and coding.

E. ICD-9-CM Coding System

1. Uniform Hospital Discharge Data Set (UHDDS) Definitions

Because the assignment of a case to a particular LTC-DRG or the proposed MS-LTC-DRG will help determine the amount that will be paid for the case, it is important that the coding is accurate. Classifications and terminology used in the LTCH PPS are consistent with the ICD-9-CM coding scheme and the UHDDS, as recommended to the Secretary by the National Committee on Vital and Health Statistics (“Uniform Hospital Discharge Data: Minimum Data Set, National Center for Health Statistics (NCHS), April 1980”) and as revised in 1984 by the Health Information Policy Council (HIPC) of the Department of Health and Human Services (HHS).

We note that the ICD-9-CM coding terminology and the definitions of principal and other diagnoses of the UHDDS are consistent with the requirements of the HIPAA Administrative Simplification Act of 1996 (45 CFR part 162). Furthermore, the UHDDS was used as a standard for the development of policies and programs related to hospital discharge statistics by both governmental and nongovernmental sectors for over 30 years. In addition, the following definitions (as described in the 1984 Revision of the UHDDS, approved by the Secretary for use starting January 1986) are requirements of the ICD-9-CM coding system, and have been used as a standard for the development of the CMS-DRGs:

  • Diagnoses are defined to include all diagnoses that affect the current hospital stay.
  • Principal diagnosis is defined as the condition established after study to be chiefly responsible for occasioning the admission of the patient to the hospital for care.
  • Other diagnoses (also called secondary diagnoses or additional diagnoses) are defined as all conditions that coexist at the time of admission, that develop subsequently, or that affect the treatment received or the LOS or both. Diagnoses that relate to an earlier episode of care that have no bearing on the current hospital stay are excluded.
  • All procedures performed will be reported. This includes those that are surgical in nature, carry a procedural risk, carry an anesthetic risk, or require specialized training.

We provide LTCHs with a 60-day window after the date of the notice of the initial LTC-DRG or proposed MS-LTC-DRG assignment to request review of that assignment of the discharge to an LTC-DRG or MS-LTC-DRG. Additional information may be provided by the LTCH to the FI as part of that review.

2. Maintenance of the ICD-9-CM Coding System

The ICD-9-CM C&M Committee is a Federal interdepartmental committee, co-chaired by the National Center for Health Statistics (NCHS) and CMS, which is charged with maintaining and updating the ICD-9-CM system. The C&M Committee is jointly responsible for approving coding changes, and developing errata, addenda, and other modifications to the ICD-9-CM to reflect newly developed procedures and technologies and newly identified diseases. The C&M Committee is also responsible for promoting the use of Federal and non-Federal educational programs and other communication techniques with a view toward standardizing coding applications and upgrading the quality of the classification system.

The NCHS has lead responsibility for the ICD-9-CM diagnosis codes included in the Tabular List and Alphabetic Index for Diseases, while CMS has the lead responsibility for the ICD-9-CM procedure codes included in the Tabular List and Alphabetic Index for Procedures. The C&M Committee encourages participation by health-related organizations in this process and holds public meetings for discussion of educational issues and proposed coding changes twice a year at the CMS Central Office located in Baltimore, Maryland. The agenda and dates of the meetings can be accessed on our Web site at http://www.cms.hhs.gov/​ICD9ProviderDiagnosticCodes.

As discussed previously in this section, for the IPPS, section 503(a) of the MMA includes a requirement for updating diagnosis and procedure codes twice a year instead of annual updates on October 1 of each year. This requirement will improve the recognition of new technologies under the IPPS by accounting for them in the GROUPER software at an earlier date. Because this statutory requirement could have a significant impact on health care providers, coding staff, publishers, system maintainers, and software systems, among others, we solicited comments on our proposed provisions to implement this requirement as part of the FY 2005 IPPS proposed rule (69 FR 28220 through 28221). We responded to comments and published our new policy regarding the updating of diagnosis and procedure codes (currently the ICD-9-CM) in the FY 2005 IPPS final rule (69 FR 48953 through 48957). In addition, we established a policy for the possibility of an April 1 ICD-9-CM diagnosis and procedure code update in the RY 2006 LTCH PPS final rule (70 FR 24176) since LTCH systems would be expected to recognize and report those new codes through the channels described in this section even though no DRG additions or deletions or changes to relative weights will occur prior to the usual October 1 update. (For more detailed information on the affect of the statutory mandates directed at the IPPS as amended by section 503(a) of the MMA, refer to the FY 2005 IPPS final rule (69 FR 48954 through 48957) and the RY 2007 LTCH PPS final rule (71 FR 27806 through 27808)).

Current addendum and code title information is published on the CMS Web site at: http://www.cms.hhs.gov/​ICD9ProviderDiagnosticCodes/​04_​addendum.asp. Summary tables showing new, revised, and deleted code titles are also posted on the CMS Web site at http://www.cms.hhs.gov/​ICD9ProviderDiagnosticCodes/​07_​summarytables.asp. Information on ICD-9-CM diagnosis codes can be found at http://www.cms.hhs.gov/​ICD9ProviderDiagnosticCodes/​. Information on new, revised, and deleted ICD-9-CM codes is also available in the American Hospital Association (AHA) publication, the Coding Clinic for ICD-9-CM. AHA also distributes information to publishers and software vendors. We also send copies of all ICD-9-CM coding changes to our contractors for use in updating Start Printed Page 26885their systems and providing education to providers. In addition, of particular note to LTCHs are the invalid diagnosis codes (Table 6C) and the invalid procedure codes (Table 6D) located in the annual proposed and final rules for the IPPS. Claims with invalid codes are not processed by the Medicare claims processing system.

3. Coding Rules and Use of ICD-9-CM Codes in LTCHs

We continue to urge LTCHs to focus on improved coding practices. Inappropriate coding of cases can adversely affect the uniformity of cases in each LTC-DRG or proposed MS-LTC-DRG and produce inappropriate weighting factors at the annual recalibration. Because of concerns raised by LTCHs concerning correct coding, we have asked the AHA to provide additional clarification and instruction on proper coding in the LTCH setting. The AHA will provide this instruction via their established process of addressing questions through their publication, the Coding Clinic for ICD-9-CM. Written questions or requests for clarification may be addressed to the Central Office on ICD-9-CM, American Hospital Association, One North Franklin, Chicago, IL 60606. A form for question(s) is available for download and can be mailed on AHA's Web site at: www.ahacentraloffice.org. In addition, current coding guidelines are available at the NCHS Web site: http://www.cdc.gov/​nchs/​datawh/​ftpserv/​ftpicd9/​ftpicd9.htm#conv.

In conjunction with the cooperating parties (AHA, the American Health Information Management Association (AHIMA), and NCHS), we reviewed actual medical records and continue to emphasize the importance of the quality of the documentation under the LTCH PPS. Based on the LTCH claims data analysis described above in section III.D.2. of this preamble, we fully believe that with some experience under a PPS, the quality of the documentation and coding of LTCHs has improved, as it did for the IPPS. However, because of the need for proper coding by LTCHs, the cooperating parties will assist their members with continued improvement in documentation and coding issues for the LTCHs through specific questions and coding guidelines. The importance of consistent and complete documentation is emphasized in the revised ICD-9-CM Official Guidelines for Coding and Reporting: “A joint effort between the attending physician and coder is essential to achieve complete and accurate documentation, code assignment, and reporting of diagnoses and procedures. The importance of consistent, complete documentation in the medical record cannot be overemphasized. Without this documentation, the application of all coding guidelines is a difficult, if not impossible task” (Coding Clinic for ICD-9-CM, Fourth Quarter 2002, page 115).

To improve medical record documentation, LTCHs should be aware that if the patient is being admitted for continuation of treatment of an acute or chronic condition, guidelines at Section I.B.10 of the Coding Clinic for ICD-9-CM, Fourth Quarter 2002 (page 129) are applicable for the selection of principal diagnosis. To clarify coding advice issued in the August 30, 2002 LTCH PPS final rule (67 FR 55979), at Guideline I.B.12, Late Effects, we state that a late effect is considered to be the residual effect (condition produced) after the acute phase of an illness or injury has terminated (Coding Clinic for ICD-9-CM, Fourth Quarter 2002, page 129). Regarding whether a LTCH should report the ICD-9-CM code(s) for an unresolved acute condition instead of the code(s) for late effects of rehabilitation, we emphasize that each case must be evaluated on its unique circumstances and coded appropriately. Depending on the documentation in the medical record, either a code reflecting the acute condition or rehabilitation could be appropriate in a LTCH.

Since implementation of the LTCH PPS, our Medicare FIs have conducted training and provided assistance to LTCHs in correct coding. We have also issued manuals containing procedures, as well as coding instructions to LTCHs and FIs. We will continue to conduct training and provide guidance on an “as needed” basis. We also refer readers to the detailed discussion on correct coding practices in the August 30, 2002 LTCH PPS final rule (67 FR 55981 through 55983). Additional coding instructions and examples will be published in the Coding Clinic for ICD-9-CM.

IV. Changes to the LTCH PPS Payment Rates for the 2008 LTCH PPS Rate Year

A. Overview of the Development of the Payment Rates

The LTCH PPS was effective beginning with a LTCH's first cost reporting period beginning on or after October 1, 2002. Effective with that cost reporting period, LTCHs are paid, during a 5-year transition period, a total LTCH prospective payment that is comprised of an increasing proportion of the LTCH PPS Federal rate and a decreasing proportion based on reasonable cost-based principles, unless the hospital makes a one-time election to receive payment based on 100 percent of the Federal rate, as specified in § 412.533. New LTCHs (as defined at § 412.23(e)(4)) are paid based on 100 percent of the Federal rate, with no phase-in transition payments.

The basic methodology for determining LTCH PPS Federal prospective payment rates is set forth at § 412.515 through § 412.532. In this section, we discuss the factors that will be used to update the LTCH PPS standard Federal rate for the 2008 LTCH PPS rate year that will be effective for LTCH discharges occurring on or after July 1, 2007 through June 30, 2008. When we implemented the LTCH PPS in the August 30, 2002 LTCH PPS final rule (67 FR 56029 through 56031), we computed the LTCH PPS standard Federal payment rate for FY 2003 by updating the latest available (FY 1998 or FY 1999) Medicare inpatient operating and capital cost data, using the excluded hospital market basket.

Section 123(a)(1) of the BBRA requires that the PPS developed for LTCHs be budget neutral for the initial year of implementation. Therefore, in calculating the standard Federal rate under § 412.523(d)(2), we set total estimated LTCH PPS payments equal to estimated payments that would have been made under the reasonable cost-based payment methodology had the PPS for LTCHs not been implemented. Section 307(a) of the BIPA specified that the increases to the hospital-specific target amounts and the cap on the target amounts for LTCHs for FY 2002 provided for by section 307(a)(1) of the BIPA shall not be considered in the development and implementation of the LTCH PPS.

Furthermore, as specified at § 412.523(d)(1), the standard Federal rate is reduced by an adjustment factor to account for the estimated proportion of outlier payments under the LTCH PPS to total estimated LTCH PPS payments (8 percent). For further details on the development of the FY 2003 standard Federal rate, see the August 30, 2002 LTCH PPS final rule (67 FR 56027 through 56037), and for subsequent updates to the LTCH PPS Federal rate, refer to the following final rules: RY 2004 LTCH PPS final rule (68 FR 34134 through 34140), RY 2005 LTCH PPS final rule (69 FR 25682 through 25684), RY 2006 LTCH PPS final rule (70 FR 24179 through 24180), and RY 2007 LTCH PPS final rule (71 FR 27819 through 27827). Start Printed Page 26886

B. LTCH PPS Market Basket

1. Overview of the RPL Market Basket

Historically, the Medicare program has used a market basket to account for price increases of the services furnished by providers. The market basket used for the LTCH PPS includes both operating and capital-related costs of LTCHs because the LTCH PPS uses a single payment rate for both operating and capital-related costs. The development of the LTCH PPS standard Federal rate, using the excluded hospital with capital market basket, is discussed in further detail in the August 30, 2002 LTCH PPS final rule (67 FR 56027 through 56033).

In the August 30, 2002 final rule (67 FR 56016 through 56017 and 56030), which implemented the LTCH PPS, we established the use of the excluded hospital with capital market basket as the LTCH PPS market basket. The excluded hospital with capital market basket was also used to update the limits on LTCHs' operating costs for inflation under the TEFRA reasonable cost-based payment system. We explained that we believe the use of the excluded hospital with capital market basket to update LTCHs' costs for inflation was appropriate because the excluded hospital market basket (with a capital component) measures price increases of the services furnished by excluded hospitals, including LTCHs. For further details on the development of the excluded hospital with capital market basket, see the RY 2004 LTCH PPS final rule (68 FR 34134 through 34137).

In the RY 2007 LTCH PPS final rule (71 FR 27810), we noted that based on our research, we did not develop a market basket specific to LTCH services. We are still unable to create a separate market basket specifically for LTCHs due to the small number of facilities and the limited amount of data that is reported (for instance, only approximately 15 percent of LTCHs reported contract labor cost data for 2002). In that same final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of the BIPA, we adopted the “Rehabilitation, Psychiatric and Long-Term Care (RPL) market basket” as the appropriate market basket of goods and services under the LTCH PPS for discharges occurring on or after July 1, 2006. Specifically, beginning with the 2007 LTCH PPS rate year, for the LTCH PPS, we adopted the use of the RPL market basket based on FY 2002 cost report data as it was the best available data. We choose to use the FY 2002 Medicare cost reports because these are the most recent, relatively complete cost data for inpatient rehabilitation facilities (IRFs), inpatient psychiatric facilities (IPF), and LTCHs.

The RPL market basket is determined based on the operating and capital costs of IRFs, IPFs and LTCHs. Since all IRFs are now paid under the IRF PPS Federal payment rate, nearly all LTCHs are paid 100 percent of the Federal rate under the LTCH PPS, and most IPFs are transitioning to payment based on 100 percent of the Federal per diem payment amount under the IPF PPS (payments to IPFs will be based exclusively on 100 percent of the Federal rate for cost reporting periods beginning on or after January 1, 2008), the RPL market basket reflects changes in the operating and capital costs for these hospitals. As we explained in that same final rule, we believe a market basket based on the data of IRFs, IPFs and LTCHs is appropriate to use under the LTCH PPS since it is the best available data that reflects the cost structures of LTCHs.

For further details on the development of the RPL market basket, including the methodology for determining the operating and capital portions of the RPL market basket, see the RY 2007 LTCH PPS final rule (71 FR 27810 through 27817).

2. Market Basket Estimate for the 2008 LTCH PPS Rate Year

Consistent with our historical practice, we estimate market basket increase based on Global Insight's forecast using the most recent available data. The most recent estimate of the RPL market basket for July 1, 2007 through June 30, 2008 (the 2008 LTCH PPS rate year), based on Global Insight's 1st quarter 2007 forecast with history through the 4th quarter of 2006, is 3.2 percent. Global Insight, Inc. is a nationally recognized economic and financial forecasting firm that contracts with CMS to forecast changes in the components of the market baskets. Consistent with our historical practice of using market basket estimates based on the most recent available data, we are finalizing 3.2 percent as the estimate of the RPL market basket for the 2008 LTCH PPS rate year.

As discussed in greater detail in this section, for the 2008 LTCH PPS rate year, we are updating the standard Federal rate by 0.71 percent. The update reflects an adjustment based on the most recent market basket estimate (currently 3.2 percent) and an adjustment to account for the increase in case-mix in the prior period (FY 2005) that resulted from changes in coding practices rather than an increase in patient severity.

C. Standard Federal Rate for the 2008 LTCH PPS Rate Year

1. Background

At § 412.523(c)(3)(ii), for LTCH PPS rate years beginning RY 2004 through RY 2006, we updated the standard Federal rate to adjust for the most recent estimate of the projected increases in prices for LTCH inpatient hospital services. We established the policy of annually updating the standard Federal rate by the increase factor described in the RY 2004 LTCH PPS final rule (68 FR 34138) because at that time we believed that was the most appropriate method for updating the LTCH PPS standard Federal rate annually for years after FY 2003. When we moved the date of the annual update of the LTCH PPS from October 1 to July 1 in the RY 2004 LTCH PPS final rule (68 FR 34138), we revised § 412.523(c)(3) to specify that for LTCH PPS rate years beginning on or after July 1, 2003, the annual update to the standard Federal rate for the LTCH PPS would be equal to the previous rate year's Federal rate updated by the most recent estimate of increases in the appropriate market basket of goods and services included in covered inpatient LTCH services. We believed that was the most appropriate method for updating the LTCH PPS standard Federal rate annually for years after RY 2004. In the RY 2007 LTCH PPS final rule (71 FR 27818), we established at § 412.523(c)(3)(iii) that the update to the standard Federal rate for the 2007 LTCH PPS rate year is zero percent. As discussed in that same final rule, we explained that rather than solely using the most recent estimate of the LTCH PPS market basket as the basis of the update factor for the Federal rate for RY 2007, we believed it was appropriate to adjust the rate to account for the changes in coding practices (rather than patient severity) as indicated by our ongoing monitoring activities.

Accordingly, we established the LTCH PPS standard Federal rate, effective from July 1, 2006 through June 30, 2007 (the 2007 LTCH PPS rate year), at $38,086.04 (71 FR 27818). Additionally, in the RY 2007 LTCH PPS proposed rule (71 FR 4742 through 4747), we provided a description of a preliminary model of an update framework under the LTCH PPS. We received few comments on that update framework preliminary model. As discussed in the RY 2007 LTCH PPS final rule (71 FR 27818 through 27819 and 27902 through 27906), although we did not propose to adopt an analytical Start Printed Page 26887update framework, we continued to solicit comments on the framework based on the preliminary model, using the best available data and concepts, and we may propose to adopt a framework at some time in the future. While we did not receive any comments regarding the update framework during the public comment period for the RY 2008 LTCH PPS proposed rule, we continue to be interested in comments and suggestions on the preliminary model of an update framework under the LTCH PPS that was present in Appendix A of the RY 2007 LTCH PPS final rule (71 FR 27902 through 27906).

In the discussion that follows, we explain how we developed the standard Federal rate for the 2008 LTCH PPS rate year. Specifically, we explain our rationale, which is based on our ongoing monitoring activities, for implementing an annual update to the standard Federal rate for RY 2008 that reflects an adjustment for the most recent market basket estimate and an adjustment to account for the increase in case-mix in a prior period (FY 2005) that resulted from changes in coding practices rather than an increase in patient severity.

2. Update to the Standard Federal Rate for the 2008 LTCH PPS Rate Year

Under § 412.523(c)(3)(ii), for RY 2004 through RY 2006, the annual update to the LTCH PPS standard Federal rate was equal to the most recent estimate of increases in the prices of an appropriate market basket of goods and services included in covered inpatient LTCH services. As noted above in this section, in the RY 2007 LTCH PPS final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA to include appropriate adjustments in the establishment of the LTCH PPS, for discharges occurring on or after July 1, 2006 and on or before June 30, 2007 (RY 2007), we specified at § 412.523(c)(3)(iii) that the standard Federal rate from the previous year would be updated by a factor of zero percent. That is, the standard Federal rate for the 2007 LTCH PPS rate year remained the same as the standard Federal rate in effect during the 2006 LTCH PPS rate year (July 1, 2005 through June 30, 2006) (that is, $38,086.04).

As discussed in greater detail in the RY 2007 LTCH PPS final rule (71 FR 27819 through 27827), the update to the standard Federal rate for RY 2007 was determined based on the estimate of the LTCH PPS market basket and an analysis of LTCH case-mix, in conjunction with a review of LTCHs' margins and our ongoing LTCH monitoring activities. Specifically, from our CMI analysis, we calculated the observed CMI increase between FY 2003 and FY 2004 (6.75 percent) and determined that a significant portion of the 6.75 percent increase in CMI between FY 2003 and FY 2004 is due to changes in coding practices, which we define as “apparent” increase in case-mix, rather than the treatment of more resource intensive patients. We also noted that the large observed increase in LTCH case-mix was not accompanied by a corresponding increase in Medicare costs. Finally, we noted in the RY 2007 LTCH PPS final rule (71 FR 27826 through 27827) that although the most recent update of the market basket discussed in that final rule is 0.2 percent lower than the estimate of the market basket discussed in the RY 2007 LTCH PPS proposed rule, we believed that finalizing a zero percent update to the Federal rate for RY 2007 was appropriate for several reasons.

First, we did not believe that there was a significant difference between the most recent estimates of the market basket for RY 2007 (3.4 percent) and the estimate used in the RY 2007 LTCH PPS proposed rule (3.6 percent). Furthermore, there could be some minimal variation in how much of the observed case-mix increase represents real case-mix changes. Finally, because the proposed update for RY 2007 at § 412.523(c)(3)(iii) explicitly specified that the RY 2007 standard Federal rate would be the previous LTCH PPS rate year updated by an update factor of zero percent, we believe some commenters may not have been aware that the final update for RY 2007 could have been different than (that is, greater than or less than) zero percent. Thus, we believed that the best approach was to adopt an update factor of zero percent in the final rule for RY 2007, which reflected both the market basket estimate and an adjustment to account for the increase in case-mix in a prior period (FY 2004) that resulted from changes in coding practices rather than an increase in patient severity. In that same final rule (71 FR 27821), we stated that the revision to § 412.523(c)(3) only addressed an update to the LTCH PPS Federal rate for the 2007 LTCH PPS rate year (§ 412.523(c)(3)(iii)), and that we would propose future revisions to § 412.523(c)(3) to address future proposed updates to the LTCH PPS Federal rates in future rate years based on an analysis of the most recent available LTCH data.

In determining the update to the standard Federal rate for the 2008 LTCH PPS rate year, we again performed a CMI analysis using the most recent available LTCH claims data and found the observed CMI increase between FY 2004 and FY 2005 to be 3.49 percent. We believe that there is still some component of apparent CMI increase within the observed CMI increase of 3.49 percent that is due to coding practices rather than the treatment of more resource intensive patients (real CMI increase). Therefore, we believe it is appropriate to apply an adjustment to the market basket update for RY 2008 to account for the apparent CMI increase for a subsequent prior period (that is, CMI increase due to changes in coding practices during FY 2005).

Comment: Many commenters urged us to provide the full market basket update rather than finalize the proposed update factor of 0.71 percent. Several commenters maintained that market basket is a measure of the expected increase in price inputs for the upcoming year that raise the cost of resources used in providing care to Medicare patients. Furthermore, some commenters believed that an increase of less than the market basket would not account for the costs of goods and services required to deliver LTCH services and will result in rates below the cost of care.

Response: As we have discussed previously in the RY 2007 final rule (71 FR 27798), as well as throughout this section of the preamble of this final rule, while we continue to believe that an update to the 2008 LTCH PPS rate year should be based on the most recent estimate of the LTCH PPS market basket, we also believe it appropriate that the rate be adjusted by an adjustment to account for changes in coding practices. In essence, we updated the standard Federal rate for the 2008 LTCH PPS rate year by a factor (+3.2 percent) for the full market basket in addition to applying a factor (−2.49 percent) to eliminate the effect of coding or classification changes that do not reflect real changes in LTCHs' case-mix during FY 2005. This adjustment is necessary in order to account for payments that were made based on improved coding (rather than increased patient severity) in a prior year.

We note that MedPAC had recommended a zero percent update for RY 2008 (March 2007 MedPAC Report to Congress, MedPAC Payment Policy, Recommendation 3D, p. 221) and that the proposed update factor of 0.71 percent is higher than what MedPAC had believed appropriate at the time. Therefore, we disagree with the comment that an increase of less than the market basket would not account for the costs of goods and services required Start Printed Page 26888to deliver LTCH services and will result in rates below the cost of care.

Comment: Several commenters noted that in addition to case mix, other elements that would affect the price of inputs include wages, drugs, products, and supplies; therefore, the commenters question our use of “case-mix as determinative of an appropriate market basket increase.” A commenter also noted that “the market basket update is a prospective measure of price inflation, and CMS provides no data suggesting that prices will not increase by 3.2 percent over RY 2008. CMS also does not provide any data showing that prices from 2004 to 2005 and from 2005 to 2006 (years included in the agency's case-mix analysis) increased less than the market basket update amount for those years.” Consequently, the commenter believed that we have not explained adequately how case mix changes are related to the market basket to warrant a reduction in the full market basket.

Response: We believe these commenters misunderstood our approach in applying the findings from our case mix analysis. First, we do not disagree that the estimated market basket is a prediction of the increase in the costs of goods and services in the coming year. Accordingly, we have based the update to the standard Federal rate each year since RY 2004 on the most recent estimate of the market basket. For RY 2004 through RY 2006, the annual update to the LTCH PPS standard Federal rate was equal to the most recent estimate of the market basket. Beginning in RY 2007, our monitoring activities and CMI analysis determined that a significant portion of the observed increase in CMI between FY 2003 and FY 2004 is due to changes in coding practices, rather than the treatment of more resource intensive patients. Accordingly, we updated the standard Federal rate for RY 2007 based both on the full estimate of market basket and an adjustment to account for the excessive payments that were made based on improved coding (rather than increased patient severity) in a prior period (between FY 2003 and FY 2004) which consequently resulted in a zero percent update. This approach was replicated for RY 2008 which resulted in a net update to the rate for RY 2008 of 0.71 percent.

Comment: Some commenters believed there is no regulatory basis for CMS to adjust the market basket update to account for apparent case-mix increase in a previous year. Specifically, a commenter wrote, “Other than the availability of data, CMS provides no logical explanation as to why an estimation of the “apparent” increase in case-mix derived from FY 2004 and FY 2005 claims should be applied to the market basket increase in RY 2008.” Furthermore, some commenters believed the proposed update factor of 0.71 percent is not based on verifiable or relevant data.

Response: Section 123 of the BBRA as amended by section 307(b) of the BIPA conferred upon the Secretary broad discretion to determine the standard rate and make appropriate adjustments to the system. We note that while § 412.523(c)(3) specifies the update to the standard rate for each year since FY 2003, the regulations do not specifically require that the Secretary automatically apply a market basket increase to prospective years. On the contrary, the regulations are to be updated each year to reflect any update to the standard rate as a result of rulemaking. Furthermore, we consistently use the most recent available data to determine the appropriate update factor. Accordingly, for this final rule we used the most recent available data, including the most recent estimate of the RPL market basket for July 1, 2007 through June 30, 2008, based on Global Insight's 1st quarter 2007 forecast with history through the 4th quarter of 2006, and the case-mix data from FY 2004 compared to FY 2005, to establish the 0.71 percent update factor.

As discussed in detail in the RY 2007 LTCH PPS final rule (71 FR 27819 through 27827), in determining the update to the LTCH PPS Federal rate for RY 2007, we used 2.75 percent as the proxy for “real” CMI change during RY 2004. We noted in that same final rule (71 FR 27822) that we were aware of a well-established RAND Corporation (RAND) study [“Has DRG Creep Crept Up? Decomposing the Case-Mix Index Change Between 1987 and 1988” by G. M. Carter, J. P. Newhouse, and D. A. Relles, R-4098-HCFA/ProPAC (1991)]. Based upon such study, we determined that real case-mix change for IPPS hospitals was a fairly steady 1.0 and 1.4 percent per year. We also noted that in updating IPPS rates, we have consistently assumed that real case-mix change was between 1.0 to 1.4 percent per year, which is a more conservative estimate of real case-mix increase than the 2.75 percent used in determining the update to the Federal rate for RY 2007 (71 FR 27822). For further information on the update to the Federal rate for RY 2007, see the RY 2007 final rule (71 FR 27819 through 27827).

For this final rule, the CMI analysis performed in determining the Federal rate update for RY 2008 is based on the observed CMI increase from FY 2004 to FY 2005 (the first and second full years of the LTCH PPS, respectively). We believe that as the LTCH PPS matured and LTCHs have become more familiar with the DRG-based payment system, it is more appropriate to utilize the estimate of real case-mix increase (1.0 percent to 1.4 percent) based on the RAND study that is typically found in acute care hospitals under the IPPS. Furthermore, an analysis of the most recent available LTCH claims data (FY 2005 LTCH claims data from the March 2006 update of the MedPAR files) show a steady decrease in the observed CMI from year to year since FY 2003 (the observed CMI change between FY 2003 and FY 2004 is 6.75 percent, between FY 2004 and FY 2005 is 3.49 percent, and between FY 2005 and FY 2006 is estimated to be 1.9 percent), which suggests that both apparent and real components of CMI are decreasing as the LTCH PPS matures. Given the estimated 1.9 percent observed CMI increase for FY 2006, it appears that it is inappropriate to assume a constant annual real case mix of 2.75 percent.

Therefore, for periods beyond the first full year of the LTCH PPS, we believe it is no longer appropriate to use such a generous estimate of real CMI. (Many LTCHs have cost reporting periods beginning in August and thus were not paid under the LTCH PPS until August 2003. For those hospitals, the first full year of the LTCH PPS was during FY 2004.) While the well-established “real” case-mix parameters based on the RAND study are based on IPPS data, we believe they are appropriate to apply under the LTCH PPS for the reasons explained below in this section. In the RY 2008 LTCH PPS proposed rule, we solicited comments on other data sources that could be used to determine a proxy for real LTCH PPS case-mix change other than the 1.0 to 1.4 percent per year case-mix parameters based on the RAND study. Although we did not receive any comments suggesting alternative data sources that could be used to determine a proxy for real LTCH PPS case-mix change, we did receive comments pertaining to using 1.0 as the proxy for real case mix.

As we have discussed numerous times in previous LTCH PPS proposed and final rules, acute care hospitals paid under the IPPS and LTCHs paid under the LTCH PPS have much in common. Hospitals paid under both systems are required to meet the same certification criteria set forth in section 1861(e) of the Act to participate as a hospital in the Medicare program. LTCHs are certified as acute care hospitals but are classified as LTCHs for payment purposes solely because such hospitals generally have Start Printed Page 26889an inpatient ALOS of greater than 25 days (as set forth in section 1886(d)(1)(B)(iv)(I) of the Act). Furthermore, the LTCH PPS uses the same patient classification system that is used under the IPPS, and several LTCH PPS payment policies, such as the area wage adjustment (§ 412.525(c)), COLA for Alaska and Hawaii (§ 412.525(b)), and high cost outlier (HCO) policy (§ 412.525(a)) are modeled after the similar IPPS policies.

Therefore, we believe it is appropriate to utilize the estimate of real CMI increase based on the RAND study of 1.0 percent as the proxy for the portion of the observed 3.49 percent CMI increase from FY 2004 to FY 2005 that represents real CMI changes for use in determining the proposed RY 2008 Federal rate update. We are using the more conservative 1.0 percent (rather than the 1.4 percent) as a proxy for real CMI increase because it is consistent with what is used under the IPPS and we believe the similarities between LTCHs and acute care hospitals are significant as we explained previously. (For a more detailed discussion on the 1.0 percent for real CMI increase utilized in the IPPS, see the FY 2007 IPPS final rule (71 FR 48156 through 48158), and the FY 1994 IPPS proposed rule (58 FR 30444).) Accordingly, since the observed CMI change for FY 2005 is estimated at 3.49 percent (based on the most recent available LTCH case-mix data from FY 2004 compared to FY 2005), accounting for the real CMI change of 1.0 percent, we believe that 2.49 percent (3.49-1.0 = 2.49) of that increase reflects CMI increase that is due to changes in coding practices (rather than patient severity).

Comment: Some commenters disagreed with our estimate of real case mix increase which is based on a study of acute care hospitals conducted by RAND using claims data from 1987 to 1988. The commenters did not believe the old data from acute care hospitals is relevant to LTCHs.

Response: As we have discussed numerous times in previous LTCH PPS proposed and final rules, as well as in the previous section of this preamble, we continue to believe that acute care hospitals paid under the IPPS and LTCHs paid under the LTCH PPS have much in common. Hospitals paid under both systems are required to meet the same certification criteria set forth in section 1861(e) of the Act to participate as a hospital in the Medicare program. The commenters did not provide any alternative data sources to determine real case mix for LTCHs. Accordingly, we continue to believe that it is appropriate to utilize the same 1.0 percent factor to project real case mix for both, the IPPS and the LTCH PPS.

Comment: Some commenters believed we proposed to use the more conservative estimate of real case-mix increase (1.0 percent) rather than the upper bound based on the RAND study (1.4 percent) without sufficient justification. However, commenters agreed that we requested comments on other data sources that could be used to determine a proxy for real LTCH PPS case-mix changes. While we did not receive any comments providing alternative data sources to determine real case-mix increase, several commenters suggested that the best proxy for real case-mix increase is the observed case-mix increase adjusted to eliminate any provider with atypical case mix changes.

Response: We continue to believe that using the more conservative 1.0 percent (rather than the 1.4 percent) as a proxy for real CMI increase is appropriate because it is consistent with what is used under the IPPS and we believe the similarities between LTCHs and acute care hospitals are significant as we explained previously.

As we discussed in greater detail in the RY 2007 LTCH PPS final rule (71 FR 27819 through 27827), while we continue to believe that an update to the LTCH PPS Federal rate year should be based on the most recent estimate of the LTCH PPS market basket, we believe it appropriate that the rate be offset by an adjustment to account for changes in coding practices that do not reflect increased patient severity. Such an adjustment protects the integrity of the Medicare Trust Funds by ensuring that the LTCH PPS payment rates better reflect the true costs of treating LTCH patients (71 FR 27798 through 27820). Therefore, in determining the RY 2008 update to the LTCH PPS Federal rate, we believe it is appropriate to apply an adjustment to eliminate the effect of coding or classification changes in a prior period (FY 2005) that do not reflect real changes in LTCHs' case-mix. Specifically, the case-mix adjustment in determining the RY 2008 Federal rate is meant to reduce current payments to account for the increase in payments in FY 2005 that resulted from the CMI increase that was attributable to the apparent case-mix increase in that year. As was the case when we determined the RY 2007 update factor, this adjustment would be necessary to account for payments that were made based on improved coding (rather than increased patient severity) in prior years. Therefore, in this final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of the BIPA to include appropriate adjustments, including updates, in the establishment of the LTCH PPS, we are revising § 412.523(c)(3), to specify that, for discharges occurring on or after July 1, 2007 and on or before June 30, 2008, the standard Federal rate from the previous year will be updated by 0.71 percent, which is based on the most recent market basket estimate (3.2 percent) adjusted by the apparent CMI (2.49 percent) due to changes in coding practice rather than an increase in patient severity. As explained above in this section, the update factor for RY 2008 is based on the most recent estimate of the LTCH PPS market basket offset by an adjustment to account for changes in case-mix in prior periods due to changes in coding practices rather than increased patient severity. We note that the update factor of 0.71 percent is higher than the zero percent update recommended by the MedPAC for RY 2008 (MedPAC Public Meeting, January 9, 2007, Meeting Transcript pp. 225-226). In the RY 2008 LTCH PPS proposed rule, we solicited comments on a possible zero percent update to the standard Federal rate for RY 2008. While most commenters recommended a full market basket update, we did receive some comments noting that in light of MedPAC's recommendation of a zero percent update, the commenters were pleased that we did not propose to implement a zero percent update and the commenters supported our proposal of a 0.71 percent update.

Furthermore, since we are using the most recent estimates of the market basket and CMI increase in the prior period (FY 2005) for calculating the update factor to the LTCH PPS Federal rate, we noted in the proposed rule that at the time the analysis must be performed for the final rule, we would consider comments received on this proposed rule and would also use the most recent estimates available at that time, if appropriate, which may be different from the data used in the proposed rule. Therefore, we explained that the proposed update factor applied to the standard Federal rate may change in the final rule.

At this time, the most recent estimate of the LTCH PPS market basket remains at 3.2 percent, and based on FY 2005 LTCH claims data from the March 2006 update of the MedPAR files, the most recent estimate of apparent CMI increase in the prior period (FY 2005), that is, case-mix increase due to changes in coding practices, also remains at 2.49 percent. Additionally, since we did not receive any comments suggesting alternative data sources to use in Start Printed Page 26890determining a proxy for real case mix and for the reasons stated previously, we are continuing to use 1.0 percent as the proxy for the real case mix. Therefore, the RY 2008 update factor to the LTCH PPS Federal rate will be 0.71 percent (3.2−2.49 = 0.71), which reflects the adjustment to the most recent market basket estimate and accounts for the increase in case-mix in the prior period that resulted from changes in coding practices rather than an increase in patient severity. Accordingly, under the same broad authority conferred upon the Secretary under the BBRA and the BIPA referenced above in this section, we are specifying under § 412.523(c)(3)(iv), that, for discharges occurring on or after July 1, 2007 and on or before June 30, 2008, the standard Federal rate from the previous year would be updated by 0.71 percent, determined based on an adjustment to the most recent estimate of the market basket to account for case-mix increase in the prior period (FY 2005) that is due to changes in coding practices rather than patient severity.

Comment: Numerous commenters stated that we have made changes to the LTCH PPS in the last several years that have slowed the growth in the number of new LTCHs and has controlled margins. The commenters believe that the cumulative effect of these payment changes, including the reweighting of the DRGs in October 2005 and October 2006, the adoption of the original 25 percent rule, the adjustments to the SSO policy, and a zero percent update for RY 2007, has been to bring LTCH margins close to zero. With the addition of the proposed payment changes for RY 2008, the commenters believe that payment to LTCHs will be inadequate. Using our impact analysis table from the proposed rule and MedPAC's estimated margins for FY 2007 as a base for comparison, two commenters attempted to estimate LTCHs' margins for RY 2008. The commenters asserted that, according to their analyses, estimated margins for RY 2008 could be as low as −3.7 percent to −5.7 percent. Numerous commenters expressed concern that the combined effect of changes to the LTCH PPS (from the last 2 years, as well as the proposed changes for RY 2008) would reduce reimbursement below the estimates of costs. Furthermore, one commenter wrote, “A fundamental premise of the Medicare program and its payment systems is that Medicare should not knowingly reimburse providers and suppliers below the cost of care.”

Response: We acknowledge that the changes to the payment system implemented in the last several years have affected the LTCH industry. In fact, we have observed that LTCHs adapt to our regulatory changes by modifying their business model to maximize profitability while operating under the new changes. For example, when we implemented the 25 percent (or applicable percentage) threshold payment adjustment in FY 2005 for co-located LTCHs and satellites, we are aware that LTCHs shifted emphasis from developing co-located facilities to developing freestanding LTCHs. With the proposed expansion of the 25 percent (or applicable percentage) threshold payment adjustment to apply to LTCH or satellite patients that were admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH, we anticipate that LTCHs could adapt by increasing the number of admissions of patients that are HCOs from referring hospitals (exempt from the 25 percent rule). In addition, since LTCHs on average get 20 percent of their discharges from sources other than acute care hospitals, it will be possible for LTCHs to adapt by admitting more of those types of patients, thus making it easier for a LTCH to stay within the applicable threshold. We have also been informed by members of the LTCH industry that in places where there are multiple acute care hospitals, the LTCHs will be able to plan their discharges to assure that they do not exceed the threshold.

Consequently, while the commenters have conducted margins analyses based on current LTCH behaviors and assert that our changes may result in negative margins, we do not believe this will prove to be the case. Indeed, commenters made similar allegations in their objection to the changes for RY 2007, and predicted that we would see many LTCHs put out of business due to our drastically-changed policies. In actuality, we did not see a drastic reduction in either the number of LTCHs or the overall number of LTCH cases. Furthermore, reports in trade journals suggest that certain members of the LTCH industry believe they are well situated to expand in the future. Similarly, we believe LTCHs have the ability to screen patients coming to a LTCH to assure that they are truly LTC patients. However, in the case of the revised SSO policy, we believe that a payment, for those patients that have a LOS comparable to an IPPS patient for that DRG (that is, the IPPS comparable threshold) at a level comparable to the IPPS payment, is an appropriate payment.

3. Standard Federal Rate for the 2008 LTCH PPS Rate Year

In the RY 2007 LTCH PPS final rule (71 FR 27827), we established a standard Federal rate of $38,086.04 for the 2007 LTCH PPS rate year that was based on the best available data and policies established in that final rule. In this final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of the BIPA, consistent with the proposed rule, we are applying an annual update to the standard Federal rate for RY 2008 that reflects an adjustment for the most recent market basket estimate and an adjustment to account for the increase in case-mix in a prior period (FY 2005) that resulted from changes in coding practices rather than an increase in patient severity. Therefore, based on the update factor for RY 2008 of 0.71 percent, the standard Federal rate for RY 2008 will be $38,356.45. Since the standard Federal rate for the 2008 LTCH PPS rate year has already been adjusted for differences in case-mix, wages, COLAs, and HCO payments, we are not making any additional adjustments in the standard Federal rate for these factors.

D. Calculation of LTCH Prospective Payments for the 2008 LTCH PPS Rate Year

The basic methodology for determining prospective payment rates for LTCH inpatient operating and capital-related costs is set forth in § 412.515 through § 412.532. In accordance with § 412.515, we assign appropriate weighting factors to each LTC-DRG to reflect the estimated relative cost of hospital resources used for discharges within that group as compared to discharges classified within other groups. The amount of the prospective payment is based on the standard Federal rate, established under § 412.523, and adjusted for the LTC-DRG relative weights, differences in area wage levels, COLA in Alaska and Hawaii, HCOs, and other special payment provisions (SSOs under § 412.529 and interrupted stays under § 412.531).

In accordance with § 412.533, during the 5-year transition period, which is currently in its final year for LTCH cost reporting periods beginning on or after October 1, 2006 (FY 2007), a total LTCH PPS payment was based on the applicable transition blend percentage of the adjusted Federal rate and a percentage based on reasonable cost principles, unless the LTCH made a one-time election to receive payment based on 100 percent of the Federal rate. Start Printed Page 26891In the final year of the 5-year transition period, which began with LTCH cost reporting periods beginning on or after October 1, 2006, as specified at § 412.533, a total LTCH PPS payment is based on 100 percent of the Federal rate. An LTCH defined as “new” under § 412.23(e)(4) is paid based on 100 percent of the Federal rate with no blended transition payments as specified in § 412.533(d). As discussed in the August 30, 2002 LTCH PPS final rule (67 FR 56038), the applicable transition blends are set forth in § 412.533(a).

Accordingly, for cost reporting periods that began during FY 2006 (that is, on or after October 1, 2005 and on or before September 30, 2006), blended payments under the transition methodology were based on 20 percent of the LTCH's rate based on reasonable cost principles and 80 percent of the adjusted LTCH PPS Federal rate. For cost reporting periods beginning on or after October 1, 2006 (FY 2007), Medicare payment to LTCHs are determined entirely (100 percent) under the LTCH PPS Federal rate.

1. Adjustment for Area Wage Levels

a. Background

Under the authority of section 123 of the BBRA as amended by section 307(b) of the BIPA, we established an adjustment to the LTCH PPS Federal rate to account for differences in LTCH area wage levels at § 412.525(c). The labor-related share of the LTCH PPS Federal rate, currently estimated by the FY 2002-based RPL market basket (as discussed in greater detail in section IV.D.1.c. of this preamble), is adjusted to account for geographic differences in area wage levels by applying the applicable LTCH PPS wage index. The applicable LTCH PPS wage index is computed using wage data from inpatient acute care hospitals without regard to reclassification under sections 1886(d)(8) or 1886(d)(10) of the Act. Furthermore, as we discussed in the August 30, 2002 LTCH PPS final rule (67 FR 56015), we established a 5-year transition to the full wage adjustment. The applicable wage index phase-in percentages are based on the start of an LTCH's cost reporting period as shown in Table 1.

Table 1

Cost reporting periods beginning on or afterPhase-in percentage of the full wage index
October 1, 20021/5th (20 percent).
October 1, 20032/5ths (40 percent).
October 1, 20043/5ths (60 percent).
October 1, 20054/5ths (80 percent).
October 1, 20065/5ths (100 percent).

For example, for cost reporting periods beginning on or after October 1, 2005 and on or before September 30, 2006 (FY 2006), the applicable LTCH wage index value is four-fifths of the applicable full LTCH PPS wage index value. The wage index adjustment will be completely phased-in beginning with cost reporting periods beginning in FY 2007, that is, for cost reporting periods beginning on or after October 1, 2006, the applicable LTCH wage index value will be the full (five-fifths) LTCH PPS wage index value. Therefore, the majority of LTCHs are currently receiving either the four-fifths or full (five-fifths) LTCH PPS wage index value. As we established in the August 30, 2002 LTCH PPS final rule (67 FR 56018), the applicable full LTCH PPS wage index value is calculated from acute-care hospital inpatient wage index data without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act.

b. Geographic Classifications/Labor Market Area Definitions

As discussed in the August 30, 2002 LTCH PPS final rule, which implemented the LTCH PPS (67 FR 56015 through 56019), in establishing an adjustment for area wage levels under § 412.525(c), the labor-related portion of a LTCH's Federal prospective payment is adjusted by using an appropriate wage index based on the labor market area in which the LTCH is located. In the 2006 LTCH PPS rate year final rule (70 FR 24184 through 24185), in § 412.525(c), we revised the labor market area definitions used under the LTCH PPS effective for discharges occurring on or after July 1, 2005 based on the Office of Management and Budget's (OMB's) Core Based Statistical Area (CBSA) designations based on 2000 Census data because we believe that those new labor market area definitions will ensure that the LTCH PPS wage index adjustment most appropriately accounts for and reflects the relative hospital wage levels in the geographic area of the hospital as compared to the national average hospital wage level. As set forth in § 412.525(c)(2), a LTCH's wage index is determined based on the location of the LTCH in an urban or rural area as defined in § 412.64(b)(1)(ii)(A) through (C). An urban area under the LTCH PPS is defined at § 412.64(b)(1)(ii)(A) and (B). In general, an urban area is defined as a Metropolitan Statistical Area (MSA) as defined by the OMB. (In addition, a few counties located outside of MSAs are considered urban as specified at § 412.64(b)(1)(ii)(B).) Under § 412.64(b)(1)(ii)(C), a rural area is defined as any area outside of an urban area.

We note that these are the same CBSA-based designations implemented for acute care inpatient hospitals under the IPPS at § 412.64(b) effective October 1, 2004 (69 FR 49026 through 49034). For further discussion of the labor market area (geographic classification) definitions used under the LTCH PPS, see the 2006 LTCH PPS rate year final rule (70 FR 24182 through 24191).

c. Labor-Related Share

In the August 30, 2002 LTCH PPS final rule (67 FR 56016), we established a labor-related share of 72.885 percent based on the relative importance of the labor-related share of operating costs (wages and salaries, employee benefits, professional fees, postal services, and all other labor-intensive services) and capital costs of the excluded hospital with capital market basket based on FY 1992 data.

As we discussed in LTCH PPS final rules subsequent to the FY 2003 LTCH PPS final rule in which we established the original LTCH PPS labor-related share (68 FR 34142, 69 FR 25685 through 25686, and 70 FR 24182), once our research into the labor-related share methodology was complete, we would update the IPPS and excluded hospital labor-related shares based on that research and the best available data if necessary. Accordingly, we conducted analysis of our labor share methodology, which was completed prior to the development of the RY 2007 LTCH PPS proposed and final rules. In the RY 2007 LTCH PPS final rule (71 FR 27829), we updated the LTCH PPS labor-related share based on the FY 2002-based RPL market basket (discussed in section IV.B. of this preamble) because we believe that this market basket was developed based on the best available data that reflect the cost structures of LTCHs.

Consistent with our historical practice, the labor-related share currently used under the LTCH PPS is determined by identifying the national average proportion of operating costs and capital costs that are related to, influenced by, or vary with the local labor market. Specifically, in the RY 2007 LTCH PPS final rule (71 FR 27829 through 27832), we revised the LTCH PPS labor-related share from 72.885 Start Printed Page 26892percent (as established in the August 30, 2002 final rule (67 FR 56016) based on the FY 1997-based excluded hospital with capital market basket) to 75.665 percent based on the relative importance of the labor-related share of operating costs (wages and salaries, employee benefits, professional fees, and all other labor-intensive services) and capital costs of the proposed RPL market basket based on FY 2002 data from the first quarter of 2006.

In the RY 2008 LTCH PPS proposed rule (72 FR 4794), under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of the BIPA, consistent with our historical practice of determining the labor-related share by identifying the national average proportion of operating costs and capital costs that are related to, influenced by, or varies with the local labor market, and consistent with our historical practice of using the best data available, we proposed to update the LTCH PPS labor-related share from 75.665 percent to 75.511 percent based on the relative importance of the labor-related share of operating costs (wages and salaries, employee benefits, professional fees, and all other labor-intensive services) and capital costs of the FY 2002-based RPL market basket from the 3rd quarter of 2006. The labor-related share is the sum of the relative importance of wages and salaries, fringe benefits, professional fees, labor-intensive services, and a portion of the capital share from an appropriate market basket. We received no comments on our proposal to update the LTCH PPS labor-related share.

Consistent with our historical practice of using the best data available, we also proposed that if more recent data were available to determine the labor-related share of the RPL market basket (used under the LTCH PPS), we would use such data for determining the labor-related share for the 2008 LTCH PPS rate year in the final rule. As discussed above in section IV.B.2. of this preamble, we now have data from the 1st quarter of 2007 (with history through the 4th quarter of 2006). Therefore, in this final rule, for RY 2008, we are using the FY 2002-based RPL market basket costs based on data from the 1st quarter of 2007 to determine the labor-related share for the LTCH PPS effective for discharges occurring on or after July 1, 2007, as this is the most recent available data. The labor-related share for the 2008 LTCH PPS rate year will continue to be the sum of the relative importance of each labor-related cost category, and will reflect the different rates of price change for these cost categories between the base year (FY 2002) and the 2008 LTCH PPS rate year. Accordingly, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of the BIPA, consistent with our historical practice of determining the labor-related share by identifying the national average proportion of operating costs and capital costs that are related to, influenced by, or varies with the local labor market, we are revising the LTCH PPS labor-related share from 75.665 percent to 75.788 percent based on the relative importance of the labor-related share of operating costs (wages and salaries, employee benefits, professional fees, and all other labor-intensive services) and capital costs of the FY 2002-based RPL market basket from the 1st quarter of 2007, as discussed below and shown below in Table 2.

Based on the most recent available data, the sum of the relative importance for 2008 LTCH PPS rate year for operating costs (wages and salaries, employee benefits, professional fees, and labor-intensive services) is 71.767, as shown in Table 2. The portion of capital that is influenced by the local labor market is still estimated to be 46 percent, which is the same percentage used when we established the current labor-related share in the RY 2007 LTCH PPS final rule. Since, based on the most recent available data, the relative importance for capital is 8.742 percent of the FY 2002-based RPL market basket for the 2008 LTCH PPS rate year, we are multiplying the estimated portion of capital influenced by the local labor market (46 percent) by the relative importance for capital (8.742 percent) to determine the labor-related share of capital for the 2008 LTCH PPS rate year. The result is 4.021 percent (0.46 × 8.742 percent), which we add to the 71.767 percent for the operating cost amount to determine the total labor-related share for the 2008 LTCH PPS rate year. Thus, based on the latest available data, we are establishing a labor-related share of 75.788 percent (71.767 percent + 4.021 percent) under the LTCH PPS for the 2008 LTCH PPS rate year. As noted above in this section, this labor-related share is determined using the same methodology as employed in calculating the current LTCH labor-related share (71 FR 27830) and the labor-related shares used under the IRF PPS and IPF PPS, which also use the RPL market basket.

Table 2 shows the 2007 LTCH PPS rate year relative importance labor-related share of the FY 2002-based RPL market basket (established in the RY 2007 LTCH PPS final rule) and the 2008 LTCH PPS rate year relative importance labor-related share of the FY 2002-based RPL market basket.

Table 2.—RY 2007 Labor-Related Share Relative Importance and RY 2008 Labor-Related Share Relative Importance of the FY 2002-Based RPL Market Basket

Cost categoryRY 2007 relative importance*RY 2008 relative importance
Wages and Salaries52.50652.588
Employee Benefits14.04214.127
Professional fees2.8862.907
All other labor intensive services2.1522.145
Subtotal71.58671.767
Labor share of capital costs4.0794.021
Total Labor-related share75.66575.788
* As established in the RY 2007 LTCH PPS final rule (71 FR 27830).
** Other labor intensive services includes landscaping services, services to buildings, detective and protective services, repair services, laundry services, advertising, auto parking and repairs, physical fitness facilities, and other government enterprises.
Start Printed Page 26893

d. Wage Index Data

In the RY 2007 LTCH PPS final rule (71 FR 27830 through 27831), we established LTCH PPS wage index values for the 2007 LTCH PPS rate year calculated from the same data (generated in cost reporting periods beginning during FY 2002) used to compute the FY 2006 acute care hospital inpatient wage index data without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act because that was the best available data at that time. The LTCH wage index values applicable for discharges occurring on or after July 1, 2006 through June 30, 2007 are shown in Table 1 (for urban areas) and Table 2 (for rural areas) in the Addendum to the RY 2007 LTCH PPS final rule (71 FR 27906 through 27930). Acute care hospital inpatient wage index data are also used to establish the wage index adjustment used in the IRF PPS, HHA PPS, and SNF PPS. As we discussed in the August 30, 2002 LTCH PPS final rule (67 FR 56019), since hospitals that are excluded from the IPPS are not required to provide wage-related information on the Medicare cost report and because we would need to establish instructions for the collection of this LTCH data to establish a geographic reclassification adjustment under the LTCH PPS, the wage adjustment established under the LTCH PPS is based on a LTCH's actual location without regard to the urban or rural designation of any related or affiliated provider.

In the RY 2008 proposed rule (72 FR 4795-4796), under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA to determine appropriate adjustments under the LTCH PPS, for the 2008 LTCH PPS rate year, we proposed to use the same data (generated in cost reporting periods beginning during FY 2003) used to compute the FY 2007 acute care hospital inpatient wage index data without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act to determine the applicable wage index values under the LTCH PPS because these data (FY 2003) are the most recent complete data. We proposed to continue to use IPPS wage data as a proxy to determine the LTCH wage index values for the 2008 LTCH PPS rate year because both LTCHs and acute-care hospitals are required to meet the same certification criteria set forth in section 1861(e) of the Act to participate as a hospital in the Medicare program and they both compete in the same labor markets, and, therefore, experience similar wage-related costs. These data are the same FY 2003 acute care hospital inpatient wage data that were used to compute the FY 2007 wage indices currently used under the IPPS, skilled nursing facility (SNF) PPS and home health agency (HHA) PPS. The LTCH wage index values that would be applicable for discharges occurring on or after July 1, 2007 through June 30, 2008, are shown in Table 1 (for urban areas) and Table 2 (for rural areas) in Addendum A to the RY 2008 proposed rule (72 FR 4849 through 4872).

We received no comments on the proposed LTCH wage index values that would be applicable for discharges occurring on or after July 1, 2007 through June 30, 2008. Therefore, in this final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA to determine appropriate adjustments under the LTCH PPS, for the 2008 LTCH PPS rate year, we are using the same data (generated in cost reporting periods beginning during FY 2003) used to compute the FY 2007 acute care hospital inpatient wage index data without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act to determine the applicable wage index values under the LTCH PPS because these data (FY 2003) are the most recent complete data. We are continuing to use IPPS wage data as a proxy to determine the LTCH wage index values for the 2008 LTCH PPS rate year for the reasons stated in the RY 2008 proposed rule (as noted above). The LTCH wage index values that will be applicable for discharges occurring on or after July 1, 2007 through June 30, 2008, are shown in Table 1 (for urban areas) and Table 2 (for rural areas) in the Addendum to this final rule.

As discussed in section IV.D.1.a. of this preamble, the applicable wage index phase-in percentages are based on the start of a LTCH's cost reporting period beginning on or after October 1st of each year during the 5-year transition period. Thus, cost reporting periods beginning on or after October 1, 2005 and before October 1, 2006 (FY 2006), the labor-related portion of the standard Federal rate is adjusted by four-fifths of the applicable LTCH wage index value. The wage index adjustment will be completely phased-in beginning with cost reporting periods beginning in FY 2007. That is, for cost reporting periods beginning on or after October 1, 2006, the labor-related portion of the standard Federal rate is adjusted by the full (five-fifths) applicable LTCH wage index value.

Because the phase-in of the wage index does not coincide with the LTCH PPS rate year (July 1st through June 30th), most LTCHs will experience a change in the wage index phase-in percentages during the LTCH PPS rate year. For example, during the 2008 LTCH PPS rate year, for a LTCH with a September 1st fiscal year, the four-fifths wage index will be applicable for the first 2 months of the 2007 LTCH PPS rate year (July 1, 2007 through August 31, 2007) and the full (five-fifths) wage index will be applicable for the next 10 months of the 2008 LTCH PPS rate year (September 1, 2007 through June 30, 2008). For the remainder of such a LTCH's FY 2006 cost reporting periods, which coincides with the first 2 months of RY 2008, the applicable wage index value would be four-fifths of the full FY 2007 acute-care hospital inpatient wage index data, without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act (as shown in Tables 1 and 2 in the Addendum to this final rule). Beginning with this LTCH's FY 2007 cost reporting period that will begin during RY 2008, the applicable wage index value would be the full (five-fifths) FY 2007 acute care hospital inpatient wage index data, without taking into account geographic reclassification under sections 1886(d)(8) and (d)(10) of the Act (as shown in Tables 1 and 2 in the Addendum to this final rule). We note that since there are no longer any LTCHs in their cost reporting periods that began during FY 2003 through FY 2005 (the first three years of the 5-year wage index phase-in), we are no longer showing the 1/5th, 2/5ths and 3/5ths wage index values in Tables 1 and 2 in the Addendum to this final rule.

2. Adjustment for Cost-of-Living in Alaska and Hawaii

In the August 30, 2002 final rule (67 FR 56022), we established, under § 412.525(b), a COLA for LTCHs located in Alaska and Hawaii to account for the higher costs incurred in those States. In the RY 2007 LTCH PPS final rule (71 FR 27832), for the 2007 LTCH PPS rate year, we established a COLA to payments for LTCHs located in Alaska and Hawaii by multiplying the standard Federal payment rate by the appropriate factor listed in Table 8 of that same final rule.

Similarly, in the RY 2008 proposed rule (72 FR 4796), under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA to determine appropriate adjustments under the Start Printed Page 26894LTCH PPS, for the 2008 LTCH PPS rate year we proposed to apply a COLA to payments to LTCHs located in Alaska and Hawaii by multiplying the proposed standard Federal payment rate by the factors listed in Table 3 of that proposed rule because those were the most recent available data at that time. Those factors were obtained from the U.S. Office of Personnel Management (OPM) and are currently used under the IPPS. In addition, we proposed that if OPM released revised COLA factors before March 1, 2007, we would use them for the development of the payments for the 2008 LTCH rate year and publish them in the LTCH PPS final rule.

We received no comments on our proposed COLA factors for LTCHs located in Alaska and Hawaii for RY 2008. However, we note that OPM released revised COLA factors for certain areas in Alaska prior to March 1, 2007. Specifically, OPM released revised COLA factors for the city of Anchorage and 80-kilometer (50-mile) radius by road, the city of Fairbanks and 80-kilometer (50-mile) radius by road, and the city of Juneau and 80-kilometer (50-mile) radius by road. The COLA factors for all other areas of Alaska were not revised from their current values. (We note that currently there are no LTCHs located in Alaska.)

Therefore, in this final rule were are adopting the revised COLA factors for those areas in Alaska, along with the proposed COLA factors for the other areas of Alaska and Hawaii, for use under the LTCH PPS in RY 2008. We note that the revised COLA factors for certain areas of Alaska have been proposed for use under the IPPS for FY 2008, as discussed in the FY 2008 IPPS proposed rule.

In this final rule, under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA to determine appropriate adjustments under the LTCH PPS, for the 2008 LTCH PPS rate year we are applying a COLA to payments to LTCHs located in Alaska and Hawaii by multiplying the standard Federal payment rate by the factors listed below in Table 3 because these are currently the most recent available data from OPM (as noted above).

Table 3.—Cost-of-Living Adjustment Factors for Alaska and Hawaii Hospitals for the 2008 LTCH PPS Rate Year

Alaska:
City of Anchorage and 80-kilometer (50-mile) radius by road1.24
City of Fairbanks and 80-kilometer (50-mile) radius by road1.24
City of Juneau and 80-kilometer (50-mile) radius by road1.24
All other areas of Alaska1.25
Hawaii:
Honolulu County1.25
Hawaii County1.165
Kauai County1.2325
Maui County1.2375
Kalawao County1.2375

3. Adjustment for High-Cost Outliers (HCOs)

a. Background

Under the broad authority conferred upon the Secretary by section 123 of the BBRA as amended by section 307(b) of BIPA, in the regulations at § 412.525(a), we established an adjustment for additional payments for outlier cases that have extraordinarily high costs relative to the costs of most discharges. Providing additional payments for outliers strongly improves the accuracy of the LTCH PPS in determining resource costs at the patient and hospital level. These additional payments reduce the financial losses that would otherwise be incurred when treating patients who require more costly care and, therefore, reduce the incentives to underserve these patients. We set the outlier threshold before the beginning of the applicable rate year so that total estimated outlier payments are projected to equal 8 percent of total estimated payments under the LTCH PPS. Outlier payments under the LTCH PPS are determined consistent with the IPPS outlier policy.

Under § 412.525(a), we make outlier payments for any discharges if the estimated cost of a case exceeds the adjusted LTCH PPS payment for the LTC-DRG plus a fixed-loss amount. The fixed-loss amount is the amount used to limit the loss that a hospital will incur under the outlier policy for a case with unusually high costs. This results in Medicare and the LTCH sharing financial risk in the treatment of extraordinarily costly cases. Under the LTCH PPS HCO policy, the LTCH's loss is limited to the fixed-loss amount and a fixed percentage of costs above the outlier threshold (LTCH DRG payment plus the fixed-loss amount) determined by the marginal cost factor. We calculate the estimated cost of a case by multiplying the overall hospital cost-to-charge ratio (CCR) by the Medicare allowable covered charge. In accordance with § 412.525(a)(3), we pay outlier cases 80 percent of the difference between the estimated cost of the patient case and the outlier threshold (the sum of the adjusted Federal prospective payment for the LTC-DRG and the fixed-loss amount).

Under the LTCH PPS, we determine a fixed-loss amount, that is, the maximum loss that a LTCH can incur under the LTCH PPS for a case with unusually high costs before the LTCH will receive any additional payments. We calculate the fixed-loss amount by estimating aggregate payments with and without an outlier policy. The fixed-loss amount will result in estimated total outlier payments being projected to be equal to 8 percent of projected total LTCH PPS payments. Currently, MedPAR claims data and CCRs based on data from the most recent provider specific file (PSF) (or to the applicable Statewide average CCR if a LTCH's CCR data are faulty or unavailable) are used to establish a fixed-loss threshold amount under the LTCH PPS.

b. Cost-to-Charge Ratios (CCRs)

In determining outlier payments, we calculate the estimated cost of the case by multiplying the LTCH's overall CCR by the Medicare allowable charges for the case. As we discussed in greater detail in the June 9, 2003 IPPS HCO final rule (68 FR 34506 through 34516), because the LTCH PPS HCO policy at § 412.525 is modeled after the IPPS outlier policy, we believed that it and the SSO policy at § 412.529 are susceptible to the same payment vulnerabilities that became evident under the IPPS and, therefore, merited revision. Thus, we revised the HCO policy at § 412.525(a) and the SSO policy at § 412.529 in that same final rule for the determination of LTCHs' CCRs and the reconciliation of outlier payments.

Under the LTCH PPS, a single prospective payment per discharge is made for both inpatient operating and capital-related costs, and, therefore, we compute a single “overall” or “total” CCR for LTCHs based on the sum of their operating and capital costs (as described in Chapter 3, section 150.24, of the Medicare Claims Processing Manual (CMS Pub. 100-4)) as compared to total charges. Specifically, a LTCH's CCR is calculated by dividing a LTCH's total Medicare costs (that is, the sum of its operating and capital inpatient routine and ancillary costs) by its total Medicare charges (that is, the sum of its operating and capital inpatient routine and ancillary charges). (Instructions regarding the changes established in the Start Printed Page 26895June 9, 2003 IPPS HCO final rule for both LTCHs and IPPS hospitals can be found in Transmittal A-03-058 (Change Request 2785; July 3, 2003).)

As a result of the changes established in the June 9, 2003 IPPS HCO final rule, as we discussed in the RY 2007 LTCH PPS final rule (71 FR 27832 through 27833) and the FY 2007 IPPS final rule (71 FR 48119 through 48121), a LTCH is assigned the applicable Statewide average CCR if, among other things, a LTCH's CCR is found to be in excess of the applicable maximum CCR threshold (that is, the LTCH CCR ceiling). As we explained in the FY 2007 IPPS final rule (71 FR 48117), CCRs above this threshold are most likely due to faulty data reporting or entry, and, therefore, these CCRs should not be used to identify and make payments for outlier cases. Such data are clearly errors and should not be relied upon. Thus, under our established policy, if a LTCH's CCR is above the applicable ceiling, the applicable LTCH PPS Statewide average CCR is assigned to the LTCH instead of the CCR computed from its most recent (settled or tentatively settled) cost report data.

Under § 412.525(a)(4)(ii), for discharges occurring on or after August 8, 2003, and before October 1, 2006, we determined the applicable LTCH PPS Statewide average CCRs using the “combined” IPPS operating and capital Statewide average CCRs (that is, adding the separate IPPS operating and capital CCRs together to determine the LTCH PPS Statewide average CCRs). Also, under § 412.525(a)(4)(ii), for discharges occurring on or after August 8, 2003, and before October 1, 2006, if a LTCH's CCR is above the applicable “combined” IPPS operating and capital ceiling (that is, adding the separate IPPS operating and capital CCR ceiling together), the applicable Statewide average CCR may be assigned to the LTCH.

As we explained in the FY 2007 IPPS final rule (71 FR 48117 through 48121), we revised our methodology for determining the annual CCR ceiling and Statewide average CCRs under the LTCH PPS because we believe that those changes are consistent with the LTCH PPS single payment rate for inpatient operating and capital costs. Therefore, under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, in that same final rule, we revised our methodology used to determine the LTCH CCR ceiling. For discharges occurring on or after October 1, 2006, we established that the LTCH CCR ceiling specified under § 412.525(a)(4)(iv)(C)(2) is calculated as three standard deviations above the corresponding national geometric mean total CCR (established and published annually by CMS). (The fiscal intermediary (FI) may use a Statewide average CCR if, among other things, a LTCH's CCR is in excess of the LTCH CCR ceiling.) The LTCH total CCR ceiling is determined based on IPPS CCR data, by first calculating the “total” (that is, operating and capital) IPPS CCR for each hospital and then determining the average “total” IPPS CCR for all IPPS hospitals. (Our rationale for using IPPS hospital data is discussed in the FY 2007 IPPS final rule (71 FR 48117) and reiterated below in this section.) The LTCH CCR ceiling is then established at 3 standard deviations from the corresponding national geometric mean total CCR. (For further detail on our methodology for annually determining the LTCH CCR ceiling, refer to the FY 2007 IPPS final rule (71 FR 48117 through 48119).) We also established that the LTCH “total” CCR ceiling used under the LTCH PPS will continue to be published annually in the IPPS proposed and final rules, and the public should continue to consult the annual IPPS proposed and final rules for changes to the LTCH total CCR ceiling that would be effective for discharges occurring on or after October 1 each year. Accordingly, in the FY 2007 IPPS final rule (71 FR 48119), we established a FY 2007 LTCH PPS total CCR ceiling of 1.321, effective for discharges occurring on or after October 1, 2006. (We note that the proposed FY 2008 LTCH PPS total CCR ceiling, that would be effective for discharges occurring on or after October 1, 2007, was presented in the FY 2008 IPPS proposed rule.)

In addition, under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we revised our methodology to determine the Statewide average CCRs under § 412.525(a)(4)(iv)(C) for use under the LTCH PPS in a manner similar to the way we compute the “total” CCR ceiling using IPPS CCR data (71 FR 48120). Specifically, under this revised methodology we first calculate the total (that is, operating and capital) CCR for each IPPS hospital. We then calculate the weighted average “total” CCR for all IPPS hospitals in the rural areas of the State and the weighted average “total” CCR for all IPPS hospitals in the urban areas of the State. (For further detail on our methodology for annually determining the LTCH urban and rural Statewide average CCRs, refer to the FY 2007 IPPS final rule (71 FR 48119 through 48121).) We also established that the applicable Statewide average “total” (operating and capital) CCRs used under the LTCH PPS will continue to be published annually in the IPPS proposed and final rules, and the public should continue to consult the annual IPPS proposed and final rules for changes to the applicable Statewide average total CCRs that would be effective for discharges occurring on or after October 1 each year. Accordingly, in the FY 2007 IPPS final rule (71 FR 48122), the FY 2007 LTCH PPS Statewide average total CCRs for urban and rural hospitals, effective for discharges occurring on or after October 1, 2006, were presented in Table 8C of the Addendum of that final rule (71 FR 48303.) (We note that the proposed FY 2007 LTCH PPS Statewide average total CCRs for urban and rural hospitals, that would be effective for discharges occurring on or after October 1, 2007, were presented in Table 8C of the FY 2008 IPPS proposed rule.)

As we explained in the FY 2007 IPPS final rule (71 FR 48117), we continue to believe it is appropriate to use IPPS operating and capital CCRs to compute the LTCH total CCR ceiling and the Statewide average CCRs because LTCHs' cost and charge structures are similar to that of IPPS acute-care hospitals. For instance, LTCHs are certified as acute care hospitals, as set forth in section 1861(e) of the Act to participate as a hospital in the Medicare program, and these hospitals, in general, are paid as LTCHs only because their Medicare ALOS is greater than 25 days as specified in § 412.23(e). Furthermore, prior to qualifying as a LTCH under § 412.23(e)(2)(i), a hospital generally is paid as an acute-care hospital under the IPPS during the period in which it demonstrates that it has an ALOS of greater than 25 days. In addition, since there are less than 400 LTCHs, which are unevenly geographically distributed throughout the United States, there may not be sufficient LTCH CCR data to determine an appropriate LTCH PPS CCR ceiling using LTCH data.

In the FY 2007 IPPS final rule, in addition to revising our methodology for determining the annual CCR ceiling and Statewide average CCRs under the LTCH PPS for discharges occurring on or after October 1, 2006, under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we revised § 412.525(a)(4)(iv) for discharges occurring on or after October 1, 2006, to codify in 42 CFR part 412, subpart O the remaining LTCH PPS outlier policy changes that were established in the June 9, 2003 IPPS HCO final rule (68 FR 34506 through 34513), including modifications and editorial clarifications to those existing policies Start Printed Page 26896established in that final rule. We made these revisions because we believe that they more precisely describe the application of those policies as they relate to the determination of LTCH CCRs because these changes are consistent with the changes to the calculation of the LTCH CCR ceiling.

Specifically, in the FY 2007 IPPS final rule (71 FR 48119), under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we established under the LTCH PPS HCO policy at § 412.525(a)(4)(iv)(C) that the FI may use a Statewide average CCR, which is established annually by CMS, if it is unable to determine an accurate CCR for a LTCH in one of the following three circumstances: (1) New LTCHs that have not yet submitted their first Medicare cost report (for this purpose, consistent with current policy, a new LTCH would be defined as an entity that has not accepted assignment of an existing hospital's provider agreement in accordance with § 489.18); (2) LTCHs whose CCR is in excess of the LTCH CCR ceiling; and (3) other LTCHs for whom data with which to calculate a CCR are not available (for example, missing or faulty data). (Other sources of data that the FI may consider in determining a LTCH's CCR included data from a different cost reporting period for the LTCH, data from the cost reporting period preceding the period in which the hospital began to be paid as a LTCH (that is, the period of at least 6 months that it was paid as a short-term acute care hospital), or data from other comparable LTCHs, such as LTCHs in the same chain or in the same region.)

Additionally, in the FY 2007 IPPS final rule (71 FR 48121), we established under § 412.525(a)(4)(iv)(B) and § 412.529(c)(3)(iv)(B) that, for discharges occurring on or after October 1, 2006, the CCR applied at the time a claim is processed will be based on either the most recently settled cost report or the most recent tentatively settled cost report, whichever is from the latest cost reporting period. Under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, in that same final rule, we also established at § 412.525(a)(4)(iv)(A) that, for discharges occurring on or after October 1, 2006, we may specify an alternative to the CCR computed under § 412.525(a)(4)(iv)(B) (that is, computed from the most recently settled cost report or the most recent tentatively settled cost report, whichever is later), or a hospital may also request that the FI use a different (higher or lower) CCR based on substantial evidence presented by the hospital. In addition, under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we revised § 412.525(a)(3) to change the plural reference from cost-to-charge “ratios” to the singular reference to a cost-to-charge “ratio” in that final rule. For a complete discussion on all these revisions to our methodology for determining a LTCH's CCR, refer to the FY 2007 IPPS final rule (71 FR 48119 through 48121). We note that in that same FY 2007 IPPS final rule, we made similar revisions to the SSO policy at § 412.529(c)(3), as discussed in V.A.1.b. of the preamble of this proposed rule.

Comment: A commenter asked that we consider making an exception to the outlier payment reconciliation requirements for the affected hospitals by Hurricane Katrina because they would have experienced an aberrant change in their CCR during the first and second cost reporting periods that began on or after August 29, 2005.

Response: In order for a hospital to meet the requirements of outlier reconciliation, a 10 percentage point change in a LTCHs CCRs from the time of payment to the time of cost report settlement is required in addition to SSO and HCO payment being greater then $500,000 for the cost reporting period being settled. Without further explanation from the commenter, it is not clear what type of aberrant changes to the CCR the commenter is referring. Changes to costs or charges can either result in reducing or increasing a CCR in any given cost reporting period. Based on the events of Katrina, we would anticipate an increase in costs and a reduction in total charges as effected hospitals probably experienced fewer discharges in the period after Katrina. These types of changes would increase a hospital's CCR, and therefore, a hospital would not owe CMS additional funds if a hospital met the criteria for reconciliation. We also note that even if a unique circumstance arose as a result of Hurricane Katrina and resulted in a situation where a hospital would be required to pay CMS as a result of a reconciliation, we believe the existing regulation may allow us to consider the unique needs of this hospital, and no changes to the existing regulations at § 412.525(a)(4)(ii), § 412.525(a)(4)(iv)(D), § 412.529(c)(3)(ii), or § 412.529(c)(3)(iv)(E).

c. Establishment of the Fixed-Loss Amount

When we implemented the LTCH PPS, as discussed in the August 30, 2002 LTCH PPS final rule (67 FR 56022 through 56026), under the broad authority of section 123 of the BBRA as amended by section 307(b) of BIPA, we established a fixed-loss amount so that total estimated outlier payments are projected to equal 8 percent of total estimated payments under the LTCH PPS. To determine the fixed-loss amount, we estimate outlier payments and total LTCH PPS payments for each case using claims data from the MedPAR files. Specifically, to determine the outlier payment for each case, we estimate the cost of the case by multiplying the Medicare covered charges from the claim by the LTCH's hospital specific CCR. Under § 412.525(a)(3), if the estimated cost of the case exceeds the outlier threshold (the sum of the adjusted Federal prospective payment for the LTC-DRG and the fixed-loss amount), we pay an outlier payment equal to 80 percent of the difference between the estimated cost of the case and the outlier threshold (the sum of the adjusted Federal prospective payment for the LTC-DRG and the fixed-loss amount).

In the RY 2007 LTCH PPS final rule (71 FR 27838), in calculating the fixed-loss amount that would result in estimated outlier payments projected to be equal to 8 percent of total estimated payments for the 2007 LTCH PPS rate year, we used claims data from the December 2005 update of the FY 2005 MedPAR files and CCRs from the December 2005 update of the PSF, as that was the best available data at that time. We believe that CCRs from the PSF are the best available CCR data for determining estimated LTCH PPS payments for a given LTCH PPS rate year because they are the most recently available CCRs actually used to make LTCH PPS payments.

As we also discussed in the RY 2007 LTCH PPS rate year final rule (71 FR 27838), we calculated a single fixed-loss amount for the 2007 LTCH PPS rate year based on the version 23.0 of the GROUPER, which was the version in effect as of the beginning of the LTCH PPS rate year (that is, July 1, 2006 for the 2007 LTCH PPS rate year). In addition, we applied the outlier policy under § 412.525(a) in determining the fixed-loss amount for the 2007 LTCH PPS rate year; that is, we assigned the applicable Statewide average CCR only to LTCHs whose CCRs exceeded the ceiling (and not when they fell below the floor). Accordingly, we used the FY 2006 LTCH PPS CCR ceiling of 1.423 (71 FR 27838). As noted in that same final rule, in determining the fixed-loss amount for the 2007 LTCH PPS rate year using the CCRs from the PSF, there were no LTCHs with missing CCRs or with CCRs in excess of the current ceiling and, therefore, there was no need for us to independently assign the applicable Statewide average CCR to any LTCHs in Start Printed Page 26897determining the fixed-loss amount for the 2007 LTCH PPS rate year (as this may have already been done by the FI in the PSF in accordance with the established policy).

Accordingly, in 2007 LTCH PPS rate year final rule (71 FR 27838), we established a fixed-loss amount of $14,887 for the 2007 LTCH PPS rate year. Thus, we pay an outlier case 80 percent of the difference between the estimated cost of the case and the outlier threshold (the sum of the adjusted Federal LTCH PPS payment for the LTC-DRG and the fixed-loss amount of $14,887).

In the RY 2008 LTCH PPS proposed rule (72 FR 4798 through 4799), for the 2008 LTCH PPS rate year, we used the March 2006 update of the FY 2005 MedPAR claims data to determine a fixed-loss amount that would result in estimated outlier payments projected to be equal to 8 percent of total estimated payments, based on the policies described in that proposed rule, because those data are the most recent complete LTCH data available. Consistent with our historical practice of using the best data available, we also proposed that if more recent LTCH claims data become available, we would to use it for determining the fixed-loss amount for the 2008 LTCH PPS rate year in the final rule. In addition, we determined the proposed fixed-loss amount based on the version of the GROUPER that would be in effect as of the beginning of the 2008 LTCH PPS rate year (July 1, 2007), that is, Version 24.0 of the GROUPER (as established in the FY 2007 IPPS final rule (71 FR 47973)).

In the RY 2008 LTCH PPS proposed rule (72 FR 4799), we proposed to use CCRs from the June 2006 update of the PSF for determining the proposed fixed-loss amount for the 2008 LTCH PPS rate year as they are currently the most recent complete available data. Consistent with our historical practice of using the best data available, we also proposed that if more recent CCR data are available, we would use it for determining the fixed-loss amount for the 2008 LTCH PPS rate year in the final rule. As we discussed in that same proposed rule, in determining the proposed fixed-loss amount for the 2008 LTCH PPS rate year, we used the current FY 2007 applicable LTCH “total” CCR ceiling of 1.321 and LTCH Statewide average “total” CCRs established under our revised methodology in the FY 2007 IPPS final rule (71 FR 48118 and 48121) such that the current applicable Statewide average CCR would be assigned if, among other things, a LTCH's CCR exceeded the current ceiling (1.321). We noted that in determining the proposed fixed-loss amount for the 2008 LTCH PPS rate year using the CCRs from the June 2006 update of the PSF, there was no need for us to independently assign the applicable Statewide average CCR to any LTCHs (as this may have already been done by the FI in the PSF in accordance with our established policy).

Accordingly, based on the data and policies described in the RY 2008 LTCH PPS proposed rule, we proposed to apply a fixed-loss amount of $18,774 for the 2008 LTCH PPS rate year. Thus, we proposed to pay an outlier case 80 percent of the difference between the estimated cost of the case and the proposed outlier threshold (the sum of the adjusted proposed Federal LTCH payment for the LTC-DRG and the proposed fixed-loss amount of $18,774).

In the RY 2008 LTCH PPS proposed rule (72 FR 4799 through 4800), we noted that the fixed-loss amount for the 2008 LTCH PPS rate year is higher than the current fixed-loss amount of $14,887. We also discussed that we were not proposing to adjust the existing 8 percent outlier target or 80 percent marginal cost factor under the current LTCH PPS HCO policy at that time. However, we explained that we continue to be interested in any comments that would support revisiting the analysis that was used to establish the existing 8 percent outlier target and the existing 80 percent marginal cost factor, using the most recent available data to evaluate whether any changes to the current HCO policy should be made, and therefore, may result in less of an increase in the fixed-loss amount for RY 2008.

Comment: While we received no comments in support of revisiting the analysis that was used to establish the existing 8 percent outlier target and the existing 80 percent marginal cost factor, using the most recent available data, to evaluate whether any changes to the current HCO policy should be made, some commenters expressed concern over the impact of raising the fixed-loss threshold for HCOs to $18,774, an increase of $3,887 over the RY 2007 threshold. According to one commenter's analysis, the proposed fixed-loss threshold would mean that 26 percent of cases would no longer meet the HCO threshold for receiving additional payments. Specifically, a commenter wrote, “reducing access to HCO payments for this many cases is not warranted.”

Response: As we explained in the RY 2008 LTCH PPS proposed rule (72 FR 4799), in addition to being based on the most recent available LTCH data to estimate the cost of each LTCH case, the proposed change in the fixed-loss amount is primarily due to the projected decrease in estimated aggregate LTCH PPS payments that is expected to result from the approach discussed for the SSO policy under § 412.529, in conjunction with the proposed changes to the area wage adjustment and the proposed changes to the LTC-DRG relative weights for FY 2007. In that same proposed rule, we also explained that we believe that an increase in the fixed-loss amount is appropriate and necessary to maintain the requirement that estimated outlier payments would be projected to be equal to 8 percent of estimated total LTCH PPS payments, as required under § 412.525(a), because of the estimated decrease in aggregate LTCH PPS payments for the 2008 LTCH PPS rate year. Based on the regression analysis that was performed when we implemented the LTCH PPS, we established the outlier target at 8 percent of estimated total LTCH PPS payments to allow us to achieve a balance between the “conflicting considerations of the need to protect hospitals with costly cases, while maintaining incentives to improve overall efficiency” (67 FR 56024). That regression analysis also showed that additional increments of outlier payments over 8 percent (that is, raising the outlier target to a larger percentage than 8 percent) would reduce financial risk, but by successively smaller amounts. Outlier payments are budget neutral, and therefore, outlier payments are funded by prospectively reducing the non-outlier PPS payment rates by projected total outlier payments. The higher the outlier target, the greater the (prospective) reduction to the base payment would need to be applied to the Federal rate to maintain budget neutrality.

Maintaining the fixed-loss amount at the current level would result in HCO payments that exceed the current regulatory requirement that estimated outlier payments would be projected to equal 8 percent of estimated total LTCH PPS payments. In fact, our analysis shows that if we were to keep the fixed-loss amount at the current amount of $14,887, we project that estimated outlier payments would be over 10 percent of total estimated LTCH PPS payments in RY 2008. As noted above, the results of our regression analysis concluded that an outlier target in excess of 8 percent would not allow us to achieve our stated goal of the HCO policy of balancing the need to protect hospitals with costly cases, while providing an incentive for hospitals to operate efficiently. Start Printed Page 26898

We also note that we received no comments in support of revisiting the regression analysis to evaluate whether current LTCH data would support a change in the current HCO policy, such as increasing (or decreasing) the outlier target. While we understand the commenter's concern that raising the fixed-loss threshold would mean that fewer cases would qualify to receive additional payments for extraordinarily high cost, as discussed above, we would have to reduce the standard Federal rate to account for the additional estimated outlier payments that exceed the current 8 percent outlier target since outlier payments are budget neutral. This would reduce payments to all LTCH cases, not just those that would receive a HCO payment based on the amount of the current fixed-loss threshold, which could result in inappropriately low payment amounts for typical LTCH cases (as shown by our analysis of payment-to-cost ratios when we developed the existing HCO policy when we implemented the LTCH PPS (67 FR 56022 through 56027)).

In the RY 2008 LTCH PPS proposed rule (72 FR 4799 through 4800) as an alternative to the proposal to raise the fixed-loss amount, we discussed adjusting the marginal cost factor (that is, the percentage that Medicare will pay of the estimated cost of a case that exceeds the sum of the adjusted Federal prospective payment for the LTC-DRG and the fixed-loss amount for LTCH PPS outlier cases as specified in § 412.525(a)(3)), which is currently equal to 80 percent, as a means of ensuring that estimated outlier payments would be projected to equal 8 percent of estimated total LTCH PPS payments. We explained that when we initially established the 80 percent marginal cost factor, our analysis of payment-to-cost ratios for HCO cases showed that a marginal cost factor of 80 percent appropriately addresses outlier cases that are significantly more expensive than nonoutlier cases, while simultaneously maintaining the integrity of the LTCH PPS (67 FR 56022 through 56027).

In that same proposed rule, we also discussed that although proposing to raise the fixed-loss amount from $14,887 to $18,774 would increase the amount of the “loss” that a LTCH must incur under the LTCH PPS for a case with unusually high costs before the LTCH would receive any additional Medicare payments, we continue to believe that the existing 8 percent outlier target and 80 percent marginal cost factor continue to adequately maintain the LTCHs' share of the financial risk in treating the most costly patients and ensure the efficient delivery of services. Accordingly, we did not propose to adjust the existing 8 percent outlier target or 80 percent marginal cost factor under the LTCH PPS HCO policy at this time. We also noted that the proposed fixed-loss amount of $18,774 is lower than the FY 2003 fixed-loss amount of $24,450 (67 FR 56023) and the 2004 LTCH PPS rate year fixed-loss amount of $19,590 (68 FR 34144), and only slightly higher than the 2005 LTCH PPS rate year fixed-loss amount of $17,864 (69 FR 25688), all of which were in effect during the time period that we estimate positive Medicare margins (as discussed in the RY 2007 LTCH PPS final rule (71 FR 27820 through 27825)).

In conclusion, for the reasons discussed above in this section, we continue to believe a marginal cost factor of 80 percent and an outlier target of 8 percent best identifies LTCH patients that are truly unusually costly cases. Furthermore, we still believe that such a policy appropriately addresses LTCH HCO cases that are significantly more expensive than non-outlier cases, which is consistent with our intent of the LTCH HCO policy as stated when we implemented the LTCH PPS. Therefore, we are not making any changes to the marginal cost factor or outlier target in that final rule. Consequently, in order to maintain that estimated outlier payments are projected to be equal to 8 percent of estimated total LTCH PPS payments, as required under § 412.525(a), under the broad authority of section 123(a)(1) of the BBRA and section 307(b)(1) of BIPA, we are establishing a fixed-loss amount of $22,954 based on the best available LTCH data and the policies presented in this final rule (as described in greater detail below). For the reasons discussed above, we believe a fixed-loss amount of $22,954 would appropriately identify unusually costly LTCH cases while maintaining the integrity of the LTCH PPS. We note that, as discussed in the RY 2008 proposed rule (72 FR 4800), we intend to revisit a budget neutral policy change in the outlier policy (among other things), which would affect future LTCH PPS payment rates, after the conclusion of the 5-year transition period when we expect to have several years of data generated after the implementation of the LTCH PPS.

In this final rule, as we proposed and consistent with our historical practice of using the best data available (as noted above), for the 2008 LTCH PPS rate year, we used the December 2006 update of the FY 2006 MedPAR claims data to determine a fixed-loss amount that would result in estimated outlier payments projected to be equal to 8 percent of total estimated payments, based on the policies described in this final rule, because these data are the most recent complete LTCH data available. Furthermore, as noted previously, we determined the fixed-loss amount based on the version of the GROUPER that would be in effect as of the beginning of the 2008 LTCH PPS rate year (July 1, 2007), that is, Version 24.0 of the GROUPER (as established in the FY 2007 IPPS final rule (71 FR 47973)).

In addition, as we proposed and consistent with our historical practice of using the best data available (as noted above), we used CCRs from the December 2006 update of the PSF for determining the fixed-loss amount for the 2008 LTCH PPS rate year as they are currently the most recent complete available data. As we discussed above in this section, we revised our methodology for our annual determination of the applicable LTCH CCR ceiling and applicable Statewide average CCRs in determining a LTCH's CCR effective for discharges occurring on or after October 1, 2006 in the FY 2007 IPPS final rule (71 FR 48117 through 48122). Accordingly, as proposed, in determining the fixed-loss amount for the 2008 LTCH PPS rate year, we used the current FY 2007 applicable LTCH “total” CCR ceiling of 1.321 and LTCH Statewide average “total” CCRs established under our revised methodology in the FY 2007 IPPS final rule (71 FR 48118 and 48121) such that the current applicable Statewide average CCR would be assigned if, among other things, a LTCH's CCR exceeded the current ceiling (1.321). We note that in determining the fixed-loss amount for the 2008 LTCH PPS rate year using the CCRs from the December 2006 update of the PSF, there was no need for us to independently assign the applicable Statewide average CCR to any LTCHs (as this may have already been done by the FI in the PSF in accordance with our established policy). (Currently, the applicable FY 2007 LTCH Statewide average CCRs can be found in Table 8C of the FY 2007 IPPS final rule (71 FR 48303).)

Accordingly, based on the data and policies described in this final rule, we are applying a fixed-loss amount of $22,954 for the 2008 LTCH PPS rate year. Thus, we will pay an outlier case 80 percent of the difference between the estimated cost of the case and the outlier threshold (the sum of the adjusted Federal LTCH payment for the LTC-DRG and the fixed-loss amount of $22,954). As discussed above, the fixed-Start Printed Page 26899loss amount for the 2008 LTCH PPS rate year is higher than the current fixed-loss amount of $14,887. In addition to being based on the most recent available LTCH data to estimate the cost of each LTCH case (as discussed in detail below in this section), this change in the fixed-loss amount is due to the projected decrease in estimated aggregate LTCH PPS payments that is expected to result from the revision to the SSO policy under § 412.529 (discussed in greater detail in section V.A.2. of this preamble), in conjunction with the changes to the area wage adjustment (discussed in greater detail in section IV.D.1. of this preamble) and the changes to the LTC-DRG relative weights for FY 2007 (as discussed in the FY 2007 IPPS final rule (71 FR 47971 through 47994)). Specifically, as discussed in greater detail in the impact analysis presented in section XV.B.4. of this final rule, we are projecting that the changes presented in this final rule will result in an estimated 3.8 percent decrease in estimated payments per discharge in RY 2008 as compared to RY 2007, on average, for all LTCHs. While we are projecting that the 0.71 percent update to the Federal rate (discussed in section IV.C. of this preamble) will result in an increase in estimated payments per discharge in RY 2008 as compared to RY 2007, this increase will be offset by the projected decrease in estimated payments per discharge from RY 2007 to RY 2008 of 0.9 percent due to the revision to the SSO policy and a projected decrease in estimated payments per discharge from RY 2007 to RY 2008 of 1.0 percent due to the changes to the area wage adjustment (including the progression of the established phase-in of that adjustment). We also project an estimated 2.5 percent decrease in estimated payments per discharge from RY 2007 to RY 2008 due to the changes in the fixed-loss amount resulting from the use of more recent LTCH data to estimate the cost of each LTCH case.

We also note that the final fixed-loss amount for RY 2008 of $22,954 is higher than the proposed fixed-loss amount for RY 2008 of $18,778. This change in the fixed-loss amount is primarily due to the updated LTCH data (that is, LTCH claims data and CCR data) used in determining the fixed-loss amount. That is, to determine the proposed fixed-loss amount for RY 2008, we used claims data from the March 2006 update of the FY 2005 MedPAR file and CCRs from the July 2006 update of the PSF, as that was the best available data at that time.

However, to determine the fixed-loss amount for RY 2008 in this final rule, the most recent available data are the December 2006 update of the FY 2006 MedPAR claims data and the CCRs from the December 2006 update of the PSF. Our analysis of the data showed that, in general, the average cost per case has increased in the FY 2006 claim data as compared to the FY 2005 claims data, which if we had kept the fixed-loss amount at $18,778 would have caused the HCO target to exceed 8 percent. In fact, our analysis shows that if we were to keep the proposed fixed-loss amount of $18,774, we project that estimated outlier payments would be over 10 percent of total estimated LTCH PPS payments in RY 2008. As discussed at length above, when we implemented the LTCH PPS, under the HCO policy we established the outlier target at 8 percent of estimated total LTCH PPS payments to allow us to achieve a balance between the need to protect hospitals with costly cases, while providing an incentive for hospitals to operate efficiently, and an outlier target in excess of 8 percent would not allow us to achieve this goal. In fact, our analysis shows that if we were to keep the proposed fixed-loss amount of $18,774, we project that estimated outlier payments would be over 10 percent of total estimated LTCH PPS payments in RY 2008. As discussed at length above in this section, when we implemented the LTCH PPS, under the HCO policy we established the outlier target at 8 percent of estimated total LTCH PPS payments to allow us to achieve a balance between the need to protect hospitals with costly cases, while providing an incentive for hospitals to operate efficiently, and an outlier target in excess of 8 percent would not allow us to achieve this goal. Consequently, the fixed-loss amount is increased to maintain the HCO target at 8 percent. Furthermore, although in the past we have found LTCHs' CCRs have been relatively stable, in establishing the fixed-loss amount for RY 2008, we noticed that the CCRs used to estimate cost per case are more volatile in recent years. This causes us concern, and therefore, we intend to monitor LTCHs' CCRs in the future. As specified at § 412.525(a)(4)(iv)(D), HCO payments are subject to the outlier reconciliation process described below in this section.

d. Reconciliation of Outlier Payments Upon Cost Report Settlement

In the June 9, 2003 HCO final rule (68 FR 34508 through 34512), we established our policy for LTCHs that provided that effective for LTCH PPS discharges occurring on or after August 8, 2003, any reconciliation of outlier payments will be based upon the actual CCR computed from the costs and charges incurred in the period during which the discharge occurs. In that same final rule, we also established that, for discharges occurring on or after August 8, 2003, at the time of any reconciliation, outlier payments may be adjusted to account for the time value of any underpayments or overpayments based upon a widely available index to be established in advance by the Secretary and will be applied from the midpoint of the cost reporting period to the date of reconciliation. (Additional information on the administration of the reconciliation process under the IPPS is provided in CMS Program Transmittal 707 (October 12, 2005; Change Request 3966). We note that we are currently developing additional instructions on the administration of the reconciliation process under the LTCH PPS that would be similar to the IPPS reconciliation process.)

In the FY 2007 IPPS final rule (71 FR 48121 through 48122), for discharges occurring on or after October 1, 2006, we codified into the LTCH PPS section of the regulations (42 CFR part 412, subpart O) the provisions governing the determination of LTCHs' CCRs, including modifications and editorial clarifications to our existing methodology for determining the annual LTCH CCR ceiling and applicable Statewide average CCRs under the LTCH PPS. (We note that we also made the same changes under the SSO policy at § 412.529(c)(3), as discussed in section V.A.1.c. of this preamble).

In the FY 2007 IPPS final rule (71 FR 48122), under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we revised § 412.525(a)(4)(iv)(D) through (E), for discharges occurring on or after October 1, 2006, to codify in subpart O of 42 CFR part 412 the provisions discussed concerning the reconciliation of LTCH PPS outlier payments, including editorial clarifications discussed in greater detail in this section, that would more precisely describe the application of those policies. Specifically, at § 412.525(a)(4)(iv)(D), we specified that for discharges occurring on or after October 1, 2006, any reconciliation of outlier payments will be based on the CCR calculated based on a ratio of costs-to-charges computed from the relevant cost report and charge data determined at the time the cost report coinciding with the discharge is settled. In addition, at § 412.525(a)(4)(iv)(E), we specified that for discharges occurring on or after October 1, 2006, at the time of any reconciliation, outlier payments may be adjusted to account for the time value of any underpayments or overpayments. We also specified that Start Printed Page 26900such an adjustment will be based upon a widely available index to be established in advance by the Secretary and will be applied from the midpoint of the cost reporting period to the date of reconciliation. We made these additional revisions to § 412.525(a)(4) because we believe that these changes are more consistent with the LTCH PPS single payment rate for inpatient operating and capital costs (as discussed in greater detail previously), and because we believe it is more appropriate and administratively simpler to include all of the regulatory provisions concerning the determination of LTCH PPS outlier payments applicable under the LTCH PPS regulations in subpart O of 42 CFR part 412 of the CFR.

Comment: One commenter requested that we clarify how we interpret the 10 percentage point criterion of the SSO and HCO reconciliation policy.

Response: We did not propose any changes to the current reconciliation policy. Therefore, we do not believe this final rule is the appropriate vehicle to address this comment. As we have stated, we intend to issue subregulatory guidance on LTCH reconciliation that would be similar to the IPPS reconciliation process and would address the commenters question at that time.

e. Application of Outlier Policy to Short-Stay Outlier (SSO) Cases

As we discussed in the August 30, 2002 final rule (67 FR 56026), under some rare circumstances, a LTCH discharge could qualify as a SSO case (as defined under § 412.529 and discussed in section V.A.1.a. of this preamble) and also as a HCO case. In this scenario, a patient could be hospitalized for less than five-sixths of the geometric ALOS for the specific LTC-DRG, and yet incur extraordinarily high treatment costs. If the costs exceeded the outlier threshold (that is, the SSO payment plus the fixed-loss amount), the discharge would be eligible for payment as a HCO. Thus, for a SSO case in the 2008 LTCH PPS rate year, the HCO payment will be 80 percent of the difference between the estimated cost of the case and the outlier threshold (the sum of the fixed-loss amount of $22,954 and the amount paid under the SSO policy).

4. Other Payment Adjustments

As indicated earlier, we have broad authority under section 123(a)(1) of the BBRA as amended by section 307(b) of BIPA to determine appropriate adjustments under the LTCH PPS, including whether (and how) to provide for adjustments to reflect variations in the necessary costs of treatment among LTCHs. Thus, in the August 30, 2002 LTCH PPS final rule (67 FR 56014 through 56027), we discussed our extensive data analysis and rationale for not implementing an adjustment for geographic reclassification, rural location, treating a disproportionate share of low-income patients (DSH), or indirect medical education (IME) costs. In that same final rule, we stated that we would collect data and reevaluate the appropriateness of these adjustments in the future once more LTCH data become available after the LTCH PPS is implemented.

As we discussed in the RY 2007 LTCH PPS final rule (71 FR 27839), we now believe that after the completion of the 5-year transition, sufficient new data that will have been generated while LTCHs are subject to the LTCH PPS may be available for a comprehensive reevaluation of payment adjustments such as geographic reclassification, rural location, DSH, and IME. The end of the 5-year transition occurs with cost reporting periods beginning on or after October 1, 2007. Therefore, in the RY 2008 LTCH PPPS proposed rule (72 FR 4801), we did not propose to make any adjustments for geographic reclassification, rural location, DSH, or IME. However, we noted that we will continue to collect and interpret new data as they become available in the future to determine if these data support proposing any additional payment adjustments. We also reiterated our belief that it is appropriate to wait for the conclusion of the 5-year transition to 100 percent of the Federal rate under the LTCH PPS, to maximize the availability of data that are reflective of LTCH behavior in response to the implementation of the LTCH PPS to be used to conduct a comprehensive evaluation of the potential payment adjustment policies (such as rural location, DSH and IME) in conjunction with our evaluation of the possibility of making a one-time prospective adjustment to the LTCH PPS rates provided for at § 412.523(d)(3).

Therefore, in this final rule, we are not making any adjustments for geographic reclassification, rural location, DSH, or IME under the LTCH PPS for RY 2008. As noted above, we will continue to collect and interpret new data as they become available in the future to determine if these data support proposing any additional payment adjustments. We plan to conduct a comprehensive evaluation of the potential payment adjustment policies (such as rural location, DSH and IME) in conjunction with our evaluation of the possibility of making a one-time prospective adjustment to the LTCH PPS rates provided for at § 412.523(d)(3) after the conclusion of the 5-year transition to 100 percent of the Federal rate under the LTCH PPS.

5. Budget Neutrality (BN) Offset To Account for the Transition Methodology

Under § 412.533, we implemented a 5-year transition, during which a LTCH is paid a total LTCH PPS payment that is comprised of an increasing percentage of the LTCH PPS Federal prospective payment rate and a decreasing percentage of its payments based on the reasonable cost-based payment principles for each discharge. Furthermore, we allow a LTCH (other than those defined as “new” under § 412.23(e)(4)) to elect to be paid based on 100 percent of the standard Federal rate in lieu of the blended methodology.

The standard Federal rate was determined as if all LTCHs will be paid based on 100 percent of the standard Federal rate. As stated earlier, we provided for a 5-year transition period that allows LTCHs to receive LTCH PPS payments in which a component incorporates reasonable cost principles. To maintain BN for FY 2003 as required by section 123(a)(1) of the BBRA during the 5-year transition period, we reduce all LTCH Medicare payments (whether a LTCH elects payment based on 100 percent of the Federal rate or whether a LTCH is being paid under the transition blend methodology) to account for the cost of the applicable transition period methodology in a given LTCH PPS rate year.

Specifically, during the LTCH PPS rate years governed under the 5-year transition policy at § 412.533(a), we reduce all LTCH Medicare payments during the 5-year transition by a factor that is equal to 1 minus the ratio of the estimated TEFRA reasonable cost-based payments that would be made if the LTCH PPS was not implemented, to the projected total Medicare program PPS payments (that is, payments made under the transition methodology and the option to elect payment based on 100 percent of the Federal rate).

In the RY 2007 LTCH PPS final rule (71 FR 27841), based on the best available data at that time, we projected that approximately 98 percent of LTCHs will be paid based on 100 percent of the standard Federal rate rather than receive payment under the transition blend methodology for the 2006 LTCH PPS rate year. Using the same methodology described in the August 30, 2002 LTCH PPS final rule (67 FR 56034), this projection, which used updated data and inflation factors, was based on our Start Printed Page 26901estimate that either: (1) A LTCH has already elected payment based on 100 percent of the Federal rate prior to the start of the 2007 LTCH PPS rate year (July 1, 2006); or (2) a LTCH would receive higher payments based on 100 percent of the 2007 LTCH PPS rate year standard Federal rate compared to the payments it would receive under the transition blend methodology. Similarly, we projected that the remaining 2 percent of LTCHs would choose to be paid based on the applicable transition blend methodology (as set forth under § 412.533(a)) because they would receive higher payments than if they were paid based on 100 percent of the 2007 LTCH PPS rate year standard Federal rate.

Also in the RY 2007 LTCH PPS final rule (71 FR 24202), based on the best available data at that time and policy revisions described in that same rule, we projected that in absence of a transition BN offset, the full effect of the final full year of the transition period (including the election option) as compared to payments as if all LTCHs would be paid based on 100 percent of the Federal rate would result in a negligible cost to the Medicare program (that is, less than $1 million in RY 2007). Because the $1 million in estimated costs to the Medicare program was such a small percentage of the estimated total LTCH payments for RY 2007 (over $5 billion), the formula that we use to establish the BN offset resulted in a factor, which we reduce all Medicare payments by to account for the additional costs of the transition methodology of zero (due to rounding). Therefore, we established a zero percent transition period BN offset to all LTCH PPS payments for discharge occurring on or after July 1, 2006 through June 30, 2007, to account for the estimated cost of the transition period methodology (including the option to elect payment based on 100 percent of the Federal rate) in RY 2007. Furthermore, in that same final rule (71 FR 27841), we explained that we are no longer projecting a small cost for the 2008 LTCH PPS rate year (July 1, 2007 through June 30, 2008) even though some LTCHs will have a cost reporting period for the 5th year of the transition period which will be concluding in the first 3 months of the 2008 LTCH PPS rate year. This is because, based on the most available data, we are projecting that the vast majority of LTCHs would have made the election to be paid based on 100 percent of the Federal rate rather than the transition blend which would result in a negligible cost to the Medicare program. In fact, as discussed in the RY 2008 LTCH PPS proposed rule (72 FR 4802), based on the most recent available data at that time from the July 2006 update of the PSF, we continue to estimate that nearly all (over 98 percent) LTCHs are currently being paid based on 100 percent of the Federal rate (rather than the transition blend methodology). Even for those few remaining LTCHs paid under the transition blend methodology set forth at § 412.533(a), the majority of their LTCH PPS payments are now based on at least 80 percent of the Federal rate and 20 percent of the reasonable cost amount (for cost reporting periods beginning during FY 2006) since there are no longer any LTCHs in their cost reporting periods that began during FY 2003 through FY 2005 (the first three years of the 5-year transition period). Therefore, in that same proposed rule, we explained that we continue to believe that there would be no measurable estimated cost to the Medicare program due to the transition period methodology (including the option to elect payment based on 100 percent of the Federal rate) in RY 2008. Accordingly, we did not propose a transition BN offset to all LTCH PPS payments for discharges occurring on or after July 1, 2007 through June 30, 2008, to account for the estimated cost of the transition period methodology (including the option to elect payment based on 100 percent of the Federal rate, since some LTCHs may still be paid under the 4th year of the transition blend methodology, specified at § 412.533, for the first 3 months of RY 2008) in RY 2008.

We received no comments on this proposal, and based on the most recent available data from the December 2006 update of the PSF, we continue to estimate that nearly all (over 98 percent) LTCHs are currently being paid based on 100 percent of the Federal rate (rather than the transition blend methodology). Therefore, we continue to believe that there would be no measurable estimated cost to the Medicare program due to the transition period methodology (including the option to elect payment based on 100 percent of the Federal rate) in RY 2008. Accordingly, in this final rule, based on updated data and using the same methodology established in the August 30, 2002 final rule (67 FR 56034), we are not implementing a transition BN offset to all LTCH PPS payments for discharges occurring on or after July 1, 2007 through June 30, 2008, to account for the estimated cost of the transition period methodology (including the option to elect payment based on 100 percent of the Federal rate, since some LTCHs may still be paid under the 4th year of the transition blend methodology, specified at § 412.533, for the first 3 months of RY 2008) in RY 2008.

6. One-Time Prospective Adjustment to the Standard Federal Rate.

As we discussed in the August 30, 2002 LTCH PPS final rule (67 FR 56036), consistent with the statutory requirement for BN in section 123(a)(1) of the BBRA, we estimated aggregate payments under the LTCH PPS for FY 2003 to be equal to the estimated aggregate payments that would be made if the LTCH PPS were not implemented. Our methodology for estimating payments for purposes of the BN calculations used the best available data at the time and necessarily reflected assumptions. As the LTCH PPS progresses, we are monitoring payment data and will evaluate the ultimate accuracy of the assumptions used in the BN calculations (for example, inflation factors, intensity of services provided, or behavioral response to the implementation of the LTCH PPS) described in the August 30, 2002 LTCH PPS final rule (67 FR 56027 through 56037). To the extent these assumptions significantly differ from actual experience, the aggregate amount of actual payments may turn out to be significantly higher or lower than the estimates on which the BN calculations were based.

Section 123(a)(1) of the BBRA as amended by section 307(b) of BIPA provides broad authority to the Secretary in developing the LTCH PPS, including the authority for establishing appropriate adjustments. Under this broad authority to make appropriate adjustments, as implemented in the existing § 412.523(d)(3) (as revised in the RY 2007 LTCH PPS final rule), we have provided for the possibility of making a one-time prospective adjustment to the LTCH PPS rates by July 1, 2008, so that the effect of any significant difference between actual payments and estimated payments for the first year of the LTCH PPS would not be perpetuated in the LTCH PPS rates for future years.

In the RY 2007 LTCH PPS final rule (71 FR 27842), based on the best available data at that time, we estimated that total Medicare program payments for LTCH services over the next 5 LTCH PPS rate years would be $5.27 billion for the 2007 LTCH PPS rate year; $5.43 billion for the 2008 LTCH PPS rate year; $5.63 billion for the 2009 LTCH PPS rate year; $5.86 billion for the 2010 LTCH PPS rate year; and $6.13 billion Start Printed Page 26902for the 2011 LTCH PPS rate year. In the RY 2008 LTCH PPS proposed rule (72 FR 4802 through 4803), based on the best available data at that time, we estimated that total Medicare program payments for LTCH services over the next 5 LTCH PPS rate years would be $4.65 billion for the 2008 LTCH PPS rate year; $4.84 billion for the 2009 LTCH PPS rate year; $5.02 billion for the 2010 LTCH PPS rate year; $5.24 billion for the 2011 LTCH PPS rate year; and $5.48 billion for the 2012 LTCH PPS rate year.

In this final rule, consistent with the methodology established in the August 30, 2002 final rule (67 FR 56036), based on the most recent available data, we estimate that total Medicare program payments for LTCH services for the next 5 LTCH PPS rate years would be as shown in Table 4.

Table 4

LTCH PPS rate yearEstimated payments ($ in billions)
2008$4.65
20094.85
20105.04
20115.25
20125.50

In accordance with the methodology established in the August 30, 2002 LTCH PPS final rule (67 FR 56037), these estimates are based on the most recent available data, including the projection that nearly all LTCHs will be paid based on 100 percent of the LTCH PPS standard Federal rate during the majority of RY 2008 (in accordance with the transition blend percentages set forth at § 412.533(a)). These estimates are also based on our estimate of LTCH PPS rate year payments to LTCHs using CMS's Office of the Actuary's (OACT) most recent estimate of the RPL market basket of 3.2 percent for the 2008 LTCH PPS rate year, 3.2 percent for the 2009 LTCH PPS rate year, 2.8 percent for the 2010 LTCH PPS rate year, 3.1 percent for the 2011 LTCH PPS rate year, and 3.2 percent for the 2012 LTCH PPS rate year. (We note that OACT develops its spending projections based on existing policy. Therefore, changes that have not yet been implemented are not reflected in the spending projections shown in this section.) We also considered OACT's most recent projections of changes in Medicare beneficiary enrollment that estimate a change in Medicare fee-for-service beneficiary enrollment of −0.1 percent in the 2008 LTCH PPS rate year, 0.7 percent in the 2009 LTCH PPS rate year, 0.3 percent in the 2010 LTCH PPS rate year, 0.6 percent in the 2011 LTCH PPS rate year, and 1.1 percent in the 2012 LTCH PPS rate year.

In the August 30, 2002 LTCH PPS final rule implementing the LTCH PPS (67 FR 55954), we set forth the implementing regulations, based upon the broad authority granted to the Secretary, under section 123 of the BBRA as amended by section 307(b) of the BIPA. Section 123(a)(1) of the BBRA required that the system “maintain budget neutrality” for FY 2003, that is, that estimated aggregate payments under the LTCH PPS would be projected to be equal to the estimated aggregate payments that would be made if the LTCH PPS would not be implemented for FY 2003. The methodology for determining the LTCH PPS standard Federal rate for FY 2003 that would “maintain budget neutrality” is described in considerable detail in the August 30, 2002 final rule (67 FR 56027 through 56037). As we discussed in that same final rule, our methodology for estimating payments for the purposes of BN calculations used the best available data and necessarily reflects assumptions in estimating aggregate payments that would be made if the LTCH PPS was not implemented. We also stated our intentions to monitor LTCH PPS payment data to evaluate the ultimate accuracy of the assumptions used in the BN calculations (for example, inflation factors, intensity of services provided, or behavioral response to the implementation of the LTCH PPS). To the extent that those assumptions significantly differ from actual experience, the estimated aggregate amount of actual payments during FY 2003 may result in significantly higher or lower estimated payments than the estimates upon which the BN calculations were based. In that same final rule, the Secretary exercised his broad authority in establishing the LTCH PPS and provided for the possibility of a one-time prospective adjustment to the LTCH PPS rates by October 1, 2006, in § 412.523(d)(3) (this deadline was revised to July 1, 2008, in the RY 2007 LTCH PPS final rule). The purpose of that provision was to prevent any significant difference between actual payments and estimated payments for the 1st year of the LTCH PPS, when we established the budget neutral Federal rate as required by the statute (discussed previously), from being perpetuated in the PPS rates for future years.

As we discussed in the RY 2007 LTCH PPS final rule (71 FR 27842 through 27844), because the LTCH PPS was only recently implemented, sufficient new data had not been generated that would enable us to conduct a comprehensive reevaluation of our BN calculations. Therefore, in that same final rule, we did not implement a one-time adjustment under § 412.523(d)(3) so that the effect of any significant difference between actual payments and estimated payments for the 1st year of the LTCH PPS would not be perpetuated in the PPS rates for future years. However, we stated that we will continue to collect and interpret new data as it becomes available in the future to determine if this adjustment should be proposed. Therefore, in the RY 2007 LTCH PPS final rule (71 FR 27842), we revised § 412.523(d)(3) by changing the original October 1, 2006 deadline (established in the August 30, 2002 final rule that implemented the LTCH PPS) to July 1, 2008, to postpone the requirement due to the time lag in the availability of Medicare data upon which this adjustment would be based.

As we discussed in the RY 2007 LTCH PPS final rule (71 FR 27843 through 27844), we now believe that after the conclusion of the 5-year transition period, sufficient new data will be generated by the LTCH PPS for a comprehensive reevaluation of our FY 2003 BN calculations. Specifically, we explained that the final year of the 5-year transition to LTCH PPS payments based on 100 percent of the Federal rate for all LTCHs will begin for cost reporting periods beginning on or after October 1, 2006 (FY 2007), and end with cost reporting periods beginning before October 1, 2007 (FY 2008). After the conclusion of the 5-year transition period (October 1, 2007), we expect to have between 3 and 4 years (FY 2003 through FY 2006) of LTCH data generated since the implementation of the LTCH PPS. We note that there is a lag time between the submission of claims data and cost report data, and the availability of that data in the MedPAR files and HCRIS, respectively. Based on a comprehensive analysis of that data, we may then propose to make a one-time prospective adjustment to the LTCH PPS rates as provided for in § 412.523(d)(3). As also explained in that same final rule, we believe that postponing the deadline of the possible one-time prospective adjustment to the LTCH PPS rates provided for in § 412.523(d)(3) to July 1, 2008, would result in the availability of additional data generated under the LTCH PPS and, therefore, our decisions regarding a possible adjustment would be based on more complete and up-to-date data. This data would be reflective of LTCH behavior in response to the implementation of the LTCH PPS.

Evaluating the appropriateness of the possible one-time prospective Start Printed Page 26903adjustment will entail a thorough review of the actual Medicare costs incurred by LTCHs during the first year of the LTCH PPS, that is, for LTCH cost reporting periods beginning on or after October 1, 2002 through September 30, 2003. When we established the FY 2003 standard Federal rate to be budget neutral, we used the most recent LTCH cost data available at that time, and trended that data forward to estimate what Medicare would have paid to LTCHs under the TEFRA payment system if the PPS were not implemented (67 FR 56033). Our methodology for estimating payments for the purposes of BN calculations, utilized the best available data and necessarily reflected assumptions in estimating aggregate payments that would have been made had the LTCH PPS not been implemented. (The methodology for determining the LTCH PPS standard Federal rate for FY 2003 that would “maintain budget neutrality” is described in considerable detail in the August 30, 2002 LTCH PPS final rule (67 FR 56027 through 56037).) In that same final rule (67 FR 56036), we also stated our intentions to monitor LTCH PPS data to evaluate the ultimate accuracy of the assumptions used in the BN calculations (for example, inflation factors, intensity of services provided, or behavioral response to the implementation of the LTCH PPS). To the extent that those assumptions significantly differed from actual experience, the aggregate amount of actual payments during FY 2003 could be significantly higher or lower than the estimates upon which the BN calculations were based.

At the outset of the LTCH PPS, we provided for the possibility of a one-time prospective adjustment at § 412.523(d)(3). Among other things, we wanted the opportunity to adjust the LTCH PPS Federal payment rate once data were available that reflected the actual cost-based payments that would have been made under the Medicare program during FY 2003 if the LTCH PPS had not been implemented, rather than perpetuate any significant difference between actual payments and estimated payments in the 1st year of the LTCH PPS used in determining the Federal rate into future years. Therefore, in the RY 2007 LTCH PPS final rule, we revised § 412.523(d)(3) to postpone the adjustment until July 1, 2008, because by that time, given the lag time typically involved in the entire cost report settlement procedure, we believe we will be able to utilize the most accurate data reflecting the actual costs incurred by LTCHs for cost reporting periods beginning during FY 2003.

As we discussed in the RY 2008 LTCH PPS proposed rule (72 FR 4804), we continue to believe that collecting and evaluating new data as it becomes available will allow us to have the best data from the first year of the LTCH PPS upon which to base an adjustment such as this. As we explained in the RY 2007 LTCH PPS final rule (71 FR 27844), there are many LTCHs with cost reporting periods from September 1 through August 30 which first became subject to the LTCH PPS on September 1, 2003. Given the lag time required for typical cost report settlement involving submission, desk review, and in some cases an audit, which can take approximately 2 additional years to complete (and we expect to audit a number of LTCH cost reports for the purpose of this analysis), we explained that the October 1, 2006 deadline established § 412.523(d)(3) was no longer reasonable or realistic. In fact, we believe that for cost reports for providers on August 2004 fiscal year ending date, we would be in possession of the most reliable cost report data, indicating the actual costs of the Medicare program of the LTCH PPS during the year in which we established the Federal payment rate by July 2007. Any proposed adjustment under § 412.523(d)(3), if finalized could then be implemented on July 1, 2008. Therefore, in the RY 2008 LTCH PPS proposed rule, we did not propose to make a one-time adjustment under § 412.523(d)(3) since we believe that we still do not have sufficient new data to enable us to conduct a comprehensive reevaluation of our FY 2003 BN calculations (as discussed in greater detail above in this section).

Comment: We received a few comments in support of waiting another year (that is, until RY 2009) to make the one-time BN adjustment to benefit from the availability of better data. However, some other commenters noted that considering all of the payment adjustments we have made to the LTCH PPS since it was implemented on October 1, 2002, there is no need for a one-time BN adjustment to ensure that aggregate payments under the LTCH PPS would equal approximately the amount that would have been paid to LTCHs under TEFRA had the LTCH PPS not been implemented.

Response: We agree with the commenters that any one-time adjustment under § 412.523(d)(3) should be based on the most complete and up-to-date data available for a comprehensive analysis of the actual Medicare costs incurred by LTCHs during the first year of the LTCH PPS. As discussed in greater detail above, given the lag time required for typical cost report settlement and the lag time in data availability, after the conclusion of the 5-year transition period (October 1, 2007), we expect to have between 3 and 4 years (FY 2003 through FY 2006) of LTCH data generated since the implementation of the LTCH PPS. Specifically, we expect that we will be in possession of the most reliable cost report data, indicating the actual costs of the Medicare program of the LTCH PPS during the year in which we established the standard Federal base payment rate by July 2007, and any proposed adjustment under § 412.523(d)(3), if finalized could then be implemented on July 1, 2008.

We recognize that there have been many changes to the payment rates and policies under the LTCH PPS since its implementation over 5 years ago. Many of these changes have been implemented as a result of our on-going monitoring of LTCH data and changes in LTCHs' behavior in response to the implementation of the LTCH PPS. As discussed above, the purpose of the one-time adjustment under § 412.523(d)(3) is to prevent any significant difference between actual payments and estimated payments from the first year of the LTCH PPS, when we established the budget neutral Federal rate as required by the statute, from being perpetuated in the PPS rates for future years. As discussed above, our methodology for estimating payments for the purposes of BN calculations when the LTCH PPS was implemented used the best available data and necessarily reflects assumptions in estimating aggregate payments that would be made if the LTCH PPS was not implemented. To the extent that those assumptions significantly differ from actual experience, the aggregate amount of actual payments may result in significantly higher or lower payments than the estimates upon which the BN calculations were based. Therefore, we established in regulations at § 412.523(d)(3) the possibility of a one-time prospective adjustment to the LTCH PPS rates to prevent any significant difference between actual payments and estimated payments from being perpetuated in the LTCH PPS rates for future years (as described in greater detail above in this section). Among the changes that have been made to the LTCH PPS since its implementation include updates to the standard Federal rate as set forth under § 412.523(c)(3). We note that we will take into consideration such changes Start Printed Page 26904when we evaluate the most recent complete available data for the purposes of determining whether to propose a one-time prospective adjustment to the LTCH PPS rates under § 412.523(d)(3) in the RY 2009 proposed rule.

For the reasons discussed in this section, we believe that we still do not have sufficient new data to enable us to conduct a comprehensive reevaluation of our FY 2003 BN calculations. Accordingly, in this final rule, we are not making a one-time adjustment under § 412.523(d)(3) at this time.

V. Other Policy Changes for the 2008 LTCH PPS Rate Year

A. Short Stay Outlier (SSO) Cases

1. Background

In the Prospective Payment System for LTCHs: Implementation and FY 2003 Rates final rule (67 FR 55954, August 30, 2002) (hereinafter referred to as the FY 2003 LTCH PPS final rule), under § 412.529, we established a special payment policy for SSO cases, that is cases with a covered LOS that is less than or equal to five-sixths of the geometric average LOS for each LTC-DRG. When we established the SSO policy, we explained in the FY 2003 LTCH PPS final rule that “[a] short-stay outlier case may occur when a beneficiary receives less than the full course of treatment at the LTCH before being discharged.” (67 FR 55995) Also in the FY 2003 LTCH PPS final rule, we stated that when we first described the policy, in the Prospective Payment System for LTCHs: Implementation and FY 2003 Rates proposed rule (67 FR 55995, March 27, 2002), “* * * we based the proposed policy on the belief that many of these patients could have been treated more appropriately in an acute hospital subject to the acute care hospital inpatient prospective payment system”. Therefore, under the LTCH PPS, we implemented a special payment adjustment for SSO cases. Under the original SSO policy, for LTCH PPS discharges with a covered LOS of up to and including five-sixths the geometric average LOS for the LTC-DRG, we adjusted the per discharge payment under the LTCH PPS by the least of 120 percent of the estimated cost of the case, 120 percent of the LTC-DRG specific per diem amount multiplied by the covered LOS of that discharge, or the full LTC-DRG payment 67 FR 55995 through 56000).

As noted previously, generally LTCHs are defined by statute as having an ALOS of greater than 25 days. We stated that we believed that the SSO payment adjustment results in more appropriate payments, since these cases most likely did not receive a full course of a LTCH-level of treatment in such a short period of time and the full LTC-DRG payment would generally not be appropriate. Payment-to-cost ratio analyses indicated that if LTCHs received a full LTC-DRG payment for those cases, they would have been significantly “overpaid” for the resources they have actually expended in treating those patients (67 FR 55995 through 56000).

Furthermore, in establishing the SSO policy, we stated that we believed that providing a reduced payment for SSO cases would discourage hospitals from admitting these patients. We also believed that the policy did not severely penalize providers that, in good faith, had admitted a patient and provided some services before realizing that the beneficiary could receive more appropriate treatment at another site of care. As we explained in the FY 2003 LTCH PPS final rule, establishing a SSO payment for these types of cases addresses the incentives inherent in a discharge-based PPS for LTCHs for treating patients with a short LOS (67 FR 55995 through 56000).

2. Additional Discussion of the SSO Payment Formula

In the FY 2003 LTCH PPS final rule, when we first presented our rationale for establishing the SSO policy, we had proposed an adjustment to ensure appropriate payment for cases that we believed may have been transferred from an acute hospital prematurely. Even if a patient was an appropriate admission to the LTCH, we also believed that a short stay case at a LTCH most likely did not receive a full course of medical treatment during the short stay and that a full LTC-DRG payment would therefore, be inappropriate (67 FR 55995 through 56000).

In keeping with these concerns, and based on an evaluation of data from more than 3 years of the LTCH PPS, which revealed that a large percentage of SSOs had a covered LOS of 14 days or less, we revised our payment policy for SSO cases in the RY 2007 LTCH PPS final rule for subclause (I) LTCHs (71 FR 27845 through 27870).

Consistent with the Secretary's broad authority “to provide for appropriate adjustments to the long-term hospital payment system * * *” established under section 123 of the BBRA as amended by section 307(b)(1) of BIPA, for RY 2007, we reduced the cost-based option of the SSO policy adjustment to 100 percent of the estimated costs of the case for discharges occurring on or after July 1, 2006. We believed that by reducing the Medicare payment to a LTCH for a specific SSO case so that it would not exceed the estimated costs incurred for that case, we would be removing what we believed could be a financial incentive to admit and treat SSO cases that the then existing policy had established for LTCHs. We did not change the payment option of 120 percent of the per diem for a specific LTC-DRG multiplied by the covered LOS for that case because as described in detail in the FY 2003 final rule LTCH PPS, when we first established the SSO policy, we found that by adjusting the per discharge payment by paying at 120 percent of the per diem LTC-DRG payment, once a stay reaches five-sixths of the geometric average LOS for the LTC-DRG, the full LTC-DRG payment will have been made (67 FR 55999). We continue to believe that this specific methodology, which results in a gradual increase in payment as the LOS increases without producing a significant payment “cliff” at any one point, provides a reasonable payment option under the SSO policy.

However, an analysis of the FY 2004 MedPAR data indicated that even under the existing SSO policy, LTCHs were admitting short stay patients that we believe could have continued treatment at the acute care hospitals (paid for under the IPPS) but could have been actually being prematurely discharged to LTCHs. Therefore, in the RY 2007 LTCH PPS final rule, we added a fourth payment option. This fourth payment alternative, a blend of an LTCH PPS amount that is comparable to the IPPS per diem payment amount, and 120 percent of the LTC-DRG per diem payment amount, as described below in this section, reflects our belief that as the length of a SSO stay increases, the case begins to resemble a more “typical” LTCH stay and, therefore, it is appropriate that incrementally, payment should be based more on what would otherwise be payable under the LTCH PPS and less on the IPPS-comparable amount. (Specifics of calculating the IPPS-comparable amount are set forth in considerable detail in the RY 2007 LTCH PPS final rule (71 FR 27852 through 27853).

We noted at the outset of the LTCH PPS for FY 2003, that the LTCH standard rate was calibrated based on LTCH resources expended in treating a patient population requiring long stays. Therefore, in establishing the SSO policy at the beginning of the LTCH PPS, we determined that it was appropriate that we not pay a full LTC-DRG payment for a patient stay not requiring those resources (67 FR 55995 through 56000). Our revision of the payment formula for SSOs for RY 2007 reflected our belief that where a case Start Printed Page 26905met our definition of a SSO at § 412.529(a), as the covered LOS increased, the case began to more closely resemble a characteristic LTCH case (and less like a short term acute care hospital case). Therefore, it was appropriate to base an increasing percentage of payment for SSOs on the LTC-DRG payment amount and a decreasing percentage of the LTCH PPS payment amount based upon the IPPS-comparable amount.

We continue to believe that in defining a LTCH as a hospital with an inpatient ALOS of greater than 25 days in section 1886(d)(1)(B)(iv)(I) of the Act, that the Congress was focusing on LOS as the essential characteristic of this provider category. Furthermore, we believe that the statutory change requiring the establishment of the LTCH PPS emphasized that the payment system should reflect the different resource use related to inpatient hospital services provided by hospitals specified by section 1886(d)(1)(B)(iv) of the Act, that is, by LTCHs (71 FR 27865). Specifically, we believe that the language of the statute indicates that the Congress believed that LTCHs treat or should be treating patients with different medical needs which results in those patients having a significantly longer LOS than those acute care hospital patients that we pay for under the IPPS.

In section 4422 of the BBA of 1997, which required that the Secretary develop a legislative proposal for the establishment of a PPS for LTCHs, the Congress specified that the system “shall include an adequate patient classification system that reflects the differences in patient resource use and costs among such hospitals.” Section 123 of the BBRA of 1999, which required implementation of a PPS for LTCHs for cost reporting periods beginning on or after October 1, 2002, specified, among other things, that the system be a per discharge payment system, based on diagnosis-related groups (DRGs), and “reflects the differences in patient resource use and costs” of LTCH patients. Section 307(b) of the BIPA of 2000 required the Secretary “to examine the feasibility and the impact of basing payment under such a system on the use of existing (or refined) hospital DRGs that have been modified to account for different resource use of LTCH patients.”

When we developed the LTCH PPS for FY 2003, the most recently available MedPAR data (generally, for FYs 1998 and 1999) revealed that 52 percent of the Medicare patients at LTCHs nationwide had a LOS of less than two-thirds of the ALOS for the LTC-DRG to which they were grouped. Of these cases, 20 percent had stays of less than 8 days. Since payments under the LTCH PPS were based on the resources necessary for treatment requiring long term hospital-level stays, beginning with the start of the LTCH PPS, we established the SSO policy, to provide appropriate payment for stays that were significantly shorter than the ALOS for each specific LTC-DRG.

The original SSO policy focused on our concerns that a SSO patient would generally receive less than the full course of treatment at the LTCH before being discharged and a full LTC-DRG payment would not be appropriate (67 FR 55943, 55995 through 55996). As we noted in the RY 2007 LTCH PPS final rule, when we revised the SSO policy based on our analysis of the nearly 3 years of data since we designed the LTCH PPS, we believed that our SSO policy should reflect our conviction that many SSO patients could otherwise have continued to receive appropriate care in the acute care hospital from which they were admitted. Had these patients not been discharged from the acute care hospital, the additional days of treatment would have continued to have been paid for under the IPPS (71 FR 27845 through 27865).

Section 123 of the BBRA, as amended by section 307(b) of the BIPA, confers broad authority on the Secretary to implement a PPS for LTCHs, including provisions for appropriate adjustments to the payment system. This broad authority gives the Secretary flexibility to fashion a LTCH PPS based on both original policies, as well as concepts borrowed from other payment systems that are adapted, where appropriate to the LTCH context. In the RY 2007 LTCH PPS final rule, we formulated a payment adjustment under the LTCH PPS that we believed would result in an appropriate payment adjustment for those inpatient stays that we believe are not characteristic of LTCHs but could more appropriately be treated in another setting.

Subsequent to the RY 2007 LTCH PPS final rule, we have performed additional analysis of more recent data FY 2005 MedPAR data, and have determined that 42 percent of LTCH SSO discharges, or approximately 19,750 cases, had lengths of stay that were less than or equal to the average LOS plus one standard deviation of an IPPS discharge that is the same DRG as the LTC-DRG to which the case was assigned. (One standard deviation is a statistical test which measures the certainty of the average of a set of measurements for the purpose of data analysis. The standard deviation is the quantity commonly used by statisticians to measure the variation in a data set.) We believe that it is appropriate to compare the covered LOS of a LTCH case grouped to a particular LTC-DRG to the ALOS plus one standard deviation for the corresponding DRG under the IPPS. At one standard deviation, we have identified approximately 68 percent of the IPPS cases within that DRG that were discharged from acute care hospitals and paid for under the IPPS. Using the statistical test of one standard deviation of the ALOS for each DRG under the IPPS, identifies the majority of IPPS discharges in any DRG.

We believe that the 42 percent of LTCH SSO cases in the RY 2005 MedPAR files with lengths of stay that are equal to or less than the IPPS ALOS plus one standard deviation for the same DRGs under the IPPS appear to be comparable to typical stays at acute care hospitals.

Although LTCHs are certified by Medicare as acute care hospitals, we believe that the Congress intended for the higher LTCH PPS payments to be made to LTCHs that treat patients requiring prolonged hospital-level care. Payments under the LTCH PPS, in compliance with the statutory mandates, have been calibrated based on “the different resource use” of LTCHs. We believe that we are “overpaying,” under the LTCH PPS, for those SSO cases in LTCHs with covered lengths of stay that are equal to or less than the typical IPPS ALOS (that is, a LOS that is less than or equal to the average IPPS LOS plus one standard deviation for the same DRG under the IPPS).

We further believe that in excluding LTCHs from being paid under the IPPS, the Congress also recognized several types of hospital-level providers that offered a different type of treatment than could reasonably be paid for under the IPPS. Specifically, in the FY 2002 LTCH PPS final rule, we reviewed the history of LTCHs as hospitals excluded from the IPPS. At that time we quoted the legislative history of the 1983 Social Security Amendments which stated, with regard to LTCHs, that the “DRG system was developed for short-term acute care general hospitals and as currently constructed does not adequately account for special circumstances of diagnoses requiring long stays'' (Report of the Committee on Ways and Means, U.S. House of Representatives, to Accompany HR 1900, H.R. Rept. No. 98025, at 141 (1983) (67 FR 55957)). Therefore, from the very outset of the IPPS, the Congress distinguished LTCHs from short term acute care hospitals by patients' lengths Start Printed Page 26906of stay. The PPS for LTCHs that we implemented in FY 2003, complied with the statutory mandate, cited above in this section, that payments under the LTCH PPS be calibrated based on “the different resource use” of these long-stay LTCH patients. Consequently, as we stated in the RY 2007 LTCH PPS final rule, we believe that “LTCHs that admit SSO patients with lengths of stay more typical of an acute care hospital may be, in fact, behaving like acute care hospitals'' (71 FR 27847), and we also believe that it is reasonable for payments under the LTCH PPS for such cases to reflect this behavior.

MedPAR data indicate that for the approximately 350 LTCHs in existence during FY 2005 that discharged approximately 130,000 cases, 46,600 discharges were SSO patients. During that same period, the approximately 3,600 acute care hospitals throughout the United States discharged approximately 12.7 million Medicare beneficiaries. At the approximately 3,600 acute care hospitals, treatment for Medicare patients is paid for under the IPPS, including those cases with a LOS that is the same as the LOS for SSO treated at a LTCH. However at a LTCH, even under the blend payment option of the SSO policy that we established for RY 2007, a percentage of the payment for those short stay patients at LTCHs may be based on a payment rate that was calculated to reflect the “different resource use” at LTCHs as compared to payment based on DRGs at acute care hospitals paid for under the IPPS. We believe that based on this analysis under the existing SSO policy for short stay patients where the patient's LOS is less than or equal to the average LOS plus one standard deviation for the same DRG at an acute care hospital, paid for under the IPPS, our blended payment methodology could result in an excessive payment.

Our data further indicates that typically LTCHs admit approximately 80 percent of their patients from acute care hospitals where their urgent conditions have been diagnosed, treated, and stabilized. We believe that when these patients are admitted to a LTCH for an extremely short stay, the LTCH appears to be serving as a step-down unit of the acute care hospital (71 FR 27857 through 27858). (Section 1886(d)(1)(B) of the Act, provides for the establishment of rehabilitation and psychiatric units of section 1886(d) hospitals (that is, acute care hospitals paid for under the IPPS) but not LTCH units.)

As we stated in the RY 2007 LTCH PPS final rule, “* * * an analysis of the CY 2004 MedPAR files revealed that for specified DRGs for acute care cases following ICU/CCU days, there were significantly fewer ‘recuperative’ days (nearly 50 percent) for acute care outlier patients that were discharged from the acute care hospital and then admitted to a LTCH than for those patients that were discharged from the acute care hospital and not subsequently admitted to a LTCH. For example, under the IPPS for DRG 475 (Respiratory system diagnosis with ventilator support) and DRG 483 (Trach with mechanical vent 96+ hours or PDX except face, mouth and neck diagnosis), the number of ‘recuperative’ days were considerably shorter at the acute care hospital if there was a discharge at the acute care hospital followed by an admission to a LTCH.” (71 FR 27857) The data in Table 5 is consistent with our belief that many LTCHs appear to be admitting some SSO patients that could have received the care at the acute care hospital.

Table 5.—HCO LOS, ICU/CCU LOS, and Post-ICU/CCU LOS for Selected Inpatient DRGs by Post-discharge Status

[Live discharges only]

DRGCasesLOSOutlier ICU/CCU daysPost ICU/CCU days
475 (no LTCH)3,88732.520.512
475 (with LTCH)51529.622.67
483 (no LTCH)3,25773.653.620
483 (with LTCH)2,35345.7414.7

In our analysis of what we believe are excessive payments under the existing LTCH PPS for the shortest SSOs, we focused on those SSO cases where a LTCH patient's covered LOS at the LTCH is less than or equal to the ALOS plus one standard deviation for the same DRG at acute care hospitals (the “IPPS comparable threshold”) and distinguishing between those SSO cases with lengths of stay that are less than or equal to the “IPPS comparable threshold” from those that exceed that threshold.

For the purposes of this discussion, whether the LTCH SSO case is within the “IPPS comparable threshold” is determined by comparing the covered LOS of that SSO case which has been assigned to a particular LTC-DRG to the ALOS for the same DRG under the IPPS. For example, if the covered LOS of the LTCH SSO case is equal to or less than the average LOS plus one standard deviation for the same DRG under the IPPS, the LTCH SSO case would be within the “IPPS comparable threshold.” In the RY 2008 LTCH PPS proposed rule, we stated that an alternative payment option would be appropriate for such a case. We indicated that we were considering the following approach: in cases where the covered LOS was equal to or less than the “IPPS comparable threshold” (defined above in this section) of the same DRG under the IPPS, the SSO payment methodology could be revised so that payment would be based upon the least of 100 percent of estimated costs of the case as determined under § 412.529(d)(2); 120 percent of the LTC-DRG per diem multiplied by the covered LOS of the case as determined under § 412.529(d)(1); the Federal prospective payment for the LTC-DRG as determined under § 412.529(d)(3); or an LTCH PPS amount comparable to the IPPS per diem amount as defined at § 412.529(d)(4), not to exceed the full IPPS comparable amount.

We noted that the RTI Report discussed in Section XI. of the RY 2008 LTCH PPS proposed rule (72 FR 4818) included an RTI recommendation that “* * * for LTCH cases whose LOS is within 1 standard deviation of the IPPS average LOS, LTCHs should be paid the IPPS rate. When this occurs, it suggests that LTCH is providing general acute care for these patients. This will allow LTCHs to treat these cases but be paid on an equitable basis with other acute hospitals since the shorter length stay would suggest general acute treatment is being provided.” (Recommendation 11, p. 139) (We also included the Executive Summary of the RTI Report as Start Printed Page 26907Addendum B in the RY 2008 LTCH PPS proposed rule (72 FR 4884).)

Under the approach that we discussed in the RY 2008 LTCH PPS proposed rule, SSO cases with covered lengths of stay exceeding the “IPPS comparable threshold” would continue to be paid under the existing SSO payment policy at § 412.529(c)(2) which is the least of: 100 percent of the estimate cost of the case as determined under § 412.529 (d)(2); 120 percent of the per diem of the LTC-DRG multiplied by the covered LOS of the case as determined under § 412.529(d)(1); the Federal prospective payment for the LTC-DRG as determined under § 412.529(d)(3); or a blend of the 120 percent of the LTC-DRG specific per diem amount and an amount comparable to the IPPS per diem amount as set forth in § 412.529 (c)(2)(iv). (The methodology for the calculation of these amounts is specified at § 412.529(d).)

However, for the shortest SSO cases (that is, if the LTCH patient's covered LOS is less than or equal to the “IPPS-comparable threshold”), the IPPS comparable per diem amount, capped at the full IPPS comparable amount that is used under the blend option of the current SSO policy, could be the fourth payment option in the SSO payment formula, replacing the blend option in the adjusted LTCH PPS payment formula at existing § 412.529(c)(2)(iv). We indicated that we believed this approach to be appropriate because it would continue to ensure that the LTCH PPS payments are appropriate for all cases; including those with a LOS that resemble cases typically treated at acute care hospitals.

However, we also indicated that, in considering this policy direction, we did not believe that this approach for SSOs would be appropriate for the specific situation of a subsection (II) LTCH (that is, a LTCH meeting the definition specified in section 1886(d)(1)(B)(iv)(II) of the Act). We have addressed the uniqueness of this type of LTCH in several notices ((62 FR 45966, 46016, and 46026), (67 FR 55954 and 55974), (68 FR 34147 through 34148) (71 FR 27863)). We believe that subclause (II) LTCHs operate under a unique Congressional mandate which, as set forth in section 1886(d)(1)(B)(iv)(II) of the Act, circumscribes such a LTCHs' admission policies to the extent that it is being identified as a LTCH in order to provide a particular type of service (for which the ALOS is greater than 20 days) to a particular population (at least 80 percent have a principal diagnosis of neoplastic disease) (68 FR 34147). Therefore, in the RY 2008 LTCH PPS proposed rule (72 FR 4807), we indicated that exempting subsection (II) LTCHs under this approach is consistent with positions regarding the application of SSO policies to subclause (II) LTCHs. For example, in RY 2004, we provided a distinctive phase-in formula for subclause (II) LTCHs (§ 412.529(e)), and in the RY 2007 LTCH PPS final rule, we did not apply SSO policy revisions for subclause (I) LTCHs (§ 412.529(c)(2)) to subclause (II) LTCHs ((68 FR 34122, 34147 through 34148) (71 FR 27798, 27863)).

To encourage a thorough and accurate evaluation of this approach, we included a column in Table 3 of Addendum A of the RY 2008 LTCH PPS proposed rule (72 FR 4872 through 4884), which set forth the IPPS-comparable threshold for each LTC-DRG. We noted that to determine the “IPPS Comparable Threshold” for some DRGs it was sometimes necessary to supplement IPPS hospital statistical data due to a low volume of IPPS cases grouped to those DRGs. In addition, although IPPS hospital statistical data for the six transplant DRGs (103, 302, 480, 495, 512 and 513) and two error DRGs (469 and 470) may be available, we noted that we could assign a value of zero for the “IPPS Comparable Threshold” for these LTC-DRGs. This approach was consistent with our on-going policy under the LTCH PPS to assign a value of 0.0000 to the relative weights for these LTC-DRGs, as discussed in section III.D of this final rule.

As we detailed in this discussion, we are concerned as to whether it is appropriate to pay cases that have a covered LOS in the LTCH that is less than or equal to the IPPS ALOS plus one standard deviation for the same DRG more than would be paid under the IPPS for a similar case. In the RY 2008 LTCH PPS proposed rule, we solicited comments on the approach described above, as well as suggestions as to alternative ways in which to address our concerns.

We received many comments on the possible revision to the SSO policy that we discussed in the proposed rule. The commenters expressed the views of trade associations representing LTCHs, both for-profit and not-for-profit LTCH groups, medical corporations that include LTCHs, State medical societies, a Chamber of Commerce, legislators, physicians and other hospital staff, and several interested citizens. In general, commenters did not support the policy approach that we discussed and the payment effects that would result for LTCHs if the policy were adopted.

Comment: A number of commenters stated that the IPPS-comparable option that we discussed for payment under the SSO policy would be a violation of the express will of the Congress in establishing the category of hospitals that were excluded from the IPPS under section 1886(d)(1)(B) of the Act. In addition, these commenters stated that under that provision the Congress acknowledged that these excluded hospitals (that is, LTCHs, IRFs, IPFs, childrens hospitals, and cancer hospitals) could not reasonably be paid under a PPS system that had been designed to pay for treatment in acute care hospitals. Further, these commenters stated that the approach we discussed would violate the intent of the Congress (that is, as expressed in the BBRA of 1999 and the BIPA of 2000) to establish a unique PPS that is specific to LTCHs.

Some of these commenters claimed that the proposed IPPS-comparable option to the SSO payment policy would be forbidden under the statute because such a payment option would ignore the “differences in patient resource use and cost” at LTCHs. Some commenters criticized our use of the phrase “a payment otherwise comparable to what would have been paid under the IPPS” as a disingenuous attempt to “side-step” the Congressional mandate that the LTCHs not be paid based on the acute care IPPS. Generally, commenters expressed the view that, if we adopted the approach described in the RY 2008 LTCH PPS proposed rule, we would be violating the statutory intent that LTCHs be excluded from the IPPS in adopting the proposed IPPS-comparable payment adjustment under the revised SSO policy.

Some commenters specifically cited the Court's two-prong test for validity of a regulation established under Chevron U.S.A., Inc. v. Natural Resources Defense Counsel, Inc. 467 U.S. 837, 842-843 (1984), and asserted that the policy we discussed would fail to pass that test. Under the ruling, the Court asks whether the Congress addressed, in clear language, the issue in question and, if the answer is affirmative, the effect is given to the “unambiguously expressed intent of the Congress.” If the “statute is silent or ambiguous with respect to the specific issue,” the Agency's interpretation is allowed to stand as long as it is based on a permissible construction of the statute.” Id. at 843. Deference to the Agency's interpretation is “only appropriate when the agency has exercised its own judgment” and is not based upon an erroneous view of the statute. Commenters asserted that the adoption Start Printed Page 26908of the revised SSO policy that we discussed would clearly violate the statutory requirement to pay LTCHs under a PPS separate and distinct from the IPPS.

Response: We disagree with commenters' contention that the LTCH PPS SSO policy that we described in the RY 2008 LTCH PPS proposed rule, based on an IPPS comparable payment amount, constitutes payment under the IPPS. Rather, the policy that we discussed adapts methodologies and approximate payment amounts from the IPPS to specific cases under the LTCH PPS. We have adapted many different features originally developed under the IPPS for use in the LTCH PPS, including the DRG structure, wage index adjustments (and wage index values), outlier payments, and many others. We believe that none of these adaptations constitute establishment of payment under the IPPS for LTCH hospitals.

In addition, section 123 of the BBRA, as amended by section 307(b)(1) of the BIPA, confers broad discretionary authority on the Secretary to develop and implement a PPS for LTCHs, specifically mandating a few specific features of the new system including “a per discharge prospective payment system” that includes an “adequate payment classification system” based on diagnosis-related groups (DRGS) that reflects the differences in patient resource use and costs, and shall maintain budget neutrality.” Section 307(b)(1) of the BIPA further provides that the Secretary “may provide for appropriate adjustments to the long-term hospital payment system, including * * * outliers * * * ” We believe that these statutory provisions provide broad authority and allow the Secretary great flexibility to fashion a LTCH PPS based on both original policies, as well as concepts borrowed from other payment systems that are adapted, where appropriate, to the LTCH context. In the instant case, the SSO policy that we discussed in the RY 2008 LTCH PPS proposed rule utilizes principles from the IPPS payment methodology and builds upon those concepts to create a LTCH PPS payment adjustment that results in an appropriate payment for those inpatient stays that we believe do not necessarily belong in LTCHs but could be treated in another setting. In this final rule, we are adopting the approach we discussed to supplement our existing SSO policy. Therefore, we disagree with commenters that the Secretary is acting in contradiction of the statute and inconsistently with the Chevron doctrine. On the contrary, we believe that this policy is consistent with the direction given to the Secretary by the Congress in the BBRA. The Congress specifically provided for the adoption of appropriate adjustments to the LTCH PPS.

Comment: Several commenters similarly objected that adopting the policy we discussed in the proposed rule would constitute a violation of the Administrative Procedures Act (APA). Specifically, these commenters objected that our discussion of the policy failed to satisfy the APA's requirement that a notice of proposed rulemaking include “the terms or substance of the proposed rule” because we did not provide “specific regulatory language to implement” the policy. Commenters contended that, in the absence of this specific regulatory language, interested parties are “improperly limited in the degree to which they are able to participate in the rulemaking process,” even if CMS receives comments on the policy discussed.

Response: We do not agree that adopting the policy approach discussed in the proposed rule, in this final rule, would constitute a violation of the APA. Specifically, we believe that we have complied with all the applicable requirements in 5 U.S.C. 553. Among the requirements of section 553, the notice shall include the terms or substance of the proposed rule, or a description of the subjects or issues involved. Our comprehensive discussion in the proposed rule set forth the substance of the final SSO policy we are adopting in this final rule and provided a complete description of the subject and issues involved. Therefore, we believe we satisfied this and all other applicable APA requirements. Our discussion of the policy in the RY 2008 LTCH PPS proposed rule that we are adopting in this final rule was detailed and specific, and even detailed the impact the change would have on payments to LTCHs, despite the absence of regulatory language. We received 270 comments on the RY 2008 LTCH PPS proposed rule. As is evident in our detailed discussion of these comments, commenters were able to provide complex, specific, and pertinent discussion of “the terms or substance” and “description of the subjects and issues involved” of the policy that we discussed.

It may be worth noting that, despite the absence of proposed, formal regulatory text, a number of commenters (including some who raised this objection) referred to the revised SSO policy that we discussed in the proposed rule with terms such as “proposal,” “proposed change,” “proposed SSO payment methodology,” and “proposed policy.” We believe that commenters clearly understood both the substance of the possible revised policy, and the fact that we might adopt the revised policy in the final rule after review of the comments.

Comment: Several commenters stated that adopting the policy discussed in the RY 2008 LTCH PPS proposed rule would be premature, since the existing SSO policy only became fully effective on October 1, 2006. Specifically, the commenters believe that there has not been sufficient time to evaluate the impact and effectiveness of the policy change adopted last year to provide for a blend of unadjusted LTCH payment rates and IPPS-comparable LTCH PPS payment rates as one of the formulas for determining payment of SSOs. Some commenters stated that, as a result of last year's change, LTCHs no longer have an incentive to knowingly admit these kinds of patients.

Response: While we understand the concerns of the commenters, we believe that it is not premature to implement this revision to the SSO policy. We have been studying these cases intensively since the implementation of the LTCH PPS (which was fully effective for cost reporting periods on or after October 1, 2002, contrary to the implications of some commenters) and remain concerned that, in a considerable number of cases, LTCHs may be receiving higher payment than is warranted for cases that are also treated with similar lengths of stay at IPPS hospitals. We have a responsibility to ensure that Medicare trust fund is appropriately spent, and therefore, we do not believe that we should delay adoption of a provision to preserve the program's resources. However, if the commenters are indeed correct that last year's policy change removed any incentive to admit these kinds of SSO patients, the actual effect of the policy that we are now adopting may be relatively small and we believe that it is the CMS’s responsibility to conserve the Medicare program's resources to the maximum extent that is appropriate. Therefore, we are finalizing the policy in this final rule.

Comment: Several commenters supported our goal of analyzing the role of LTCHs as one of several treatment settings among post-acute providers for Medicare beneficiaries. However, they urged us not to finalize the SSO policy that we discussed in the proposed rule that would include the alternative payment option for an SSO payment comparable to the IPPS payment amount. These commenters believe that finalizing this policy would result in drastic payment reductions and Start Printed Page 26909consequential losses to the LTCHs. These commenters noted that our discussion related to serious issues about the proper place for LTCHs along the continuum of care for Medicare beneficiaries. The commenters urged us not to address these issues through payment mechanisms, but to arrive at “clinically-based” answers to these issues. Commenters also recommended that we wait until Research Triangle Institute (RTI) completes the next phase of its work, which includes a review of proposed and existing criteria to restrict admission to LTCHs to medically complex cases.

Response: The commenters are correct that the issue involves the role of LTCHs in the continuum of beneficiary care. As a provider category, LTCHs were created by section 1886(d)(1)(B)(iv)(I) of the Act and defined by the statute as “a hospital which has an average inpatient length of stay (as determined by the Secretary) of greater than 25 days.” (Subclause (II) LTCHs, discussed below in these responses, which were established under the BBA of 1997, qualify as LTCHs under highly specific requirements.) As a “prudent purchaser of care,” we believe that we have the mandate to pay appropriately for the hospital-level services provided to Medicare beneficiaries. The RTI study, as discussed in section XI. of the preamble to this final rule, represents a highly significant step in evaluating the clinical role for LTCHs. In addition to the RTI study, there is considerable attention being focused by CMS on issues of substitution of services among provider types, and the potential for the development of a uniform assessment tool across post-acute providers. As RTI evaluates the feasibility of identifying clinically-based criteria for LTCH patients, we are concerned that patients with the same general medical profile as the same types of patients that constitute some SSO cases in the LTCH setting are also being treated at acute care hospitals, often as HCO cases. Therefore, we are finalizing this specific revision to the SSO policy, as discussed in the RY 2008 LTCH PPS proposed rule, because we are concerned about the significant number of very short stay patients currently receiving treatment at LTCHs. These are patients with a LOS that is comparable to the LOS for many patients (under the same DRG) treated in acute care hospitals and paid under the IPPS. LTCHs in actuality are also acute care hospitals, they are a provider type that is distinguished solely by its focus on long-stay hospital-level care as compared to patients paid under the IPPS.

Comment: We received numerous comments that praised the quality care given to Medicare beneficiaries by the LTCHs in their areas and commenters urged us not to make significant cuts in Medicare payments which they fear would result in reduced services. The commenters asserted that the revision of the payment adjustment for SSO patients as discussed in the RY 2008 LTCH PPS proposed rule will be detrimental to the industry as costs of providing care will exceed payment. Further, the commenters stated that underpayment to LTCHs will cause patients with complex medical conditions to lose access to appropriate care and increase costs to acute care hospitals which will be forced to continue caring for these sicker patients. The commenters believed that the proposed revisions to the SSO payment policy would have a profound impact on the entire health care system of their communities since their LTCHs are a critical component of the State health care delivery system. They stated that since LTCHs offer specialized services not available elsewhere, severe cutbacks for LTCHs could resonate throughout the entire health care system.

One commenter noted that CMS made a statement that it does not expect any changes in quality of care or access to services for Medicare beneficiaries under the LTCH PPS based on proposed rule policies. However, one of the commenters stated that a decrease in payments will have pervasive effects on LTCHs. Moreover, the commenter stated that the impact of changes in our payments to LTCHs because of the proposed SSO policy revisions will not only affect services offered to “the most vulnerable patients,” but also will have an impact on the staff of the LTCHs. Several of the commenters specified that they envision that acute care hospitals will be overtaxed and incur additional costs without being able to provide ICU beds for patients requiring short-term acute care services. They also stated that the acute care hospitals in their communities may not be able to meet patient needs for those needing LTCH services.

One commenter cited the experience of a local faith-based, not-for-profit LTCH system that admits only very high acuity, long-term patients and realizes exceptional quality, outcomes, and cost effectiveness. But other LTCHs within the industry admit low acuity patients. The commenter stated, “* * * many LTCH providers seek to admit chronically ill ‘slow-recovery’ patients as a primary target population. These patients have little difficulty meeting the 25-day LTCH ALOS criteria, and while these patients may meet continued stay criteria, we believe many could be cared for in a less acute setting.”

Response: We understand the serious concerns expressed by the commenters and, although we are finalizing the SSO policy revisions as were discussed in the RY 2008 LTCH PPS proposed rule, we want to assure the commenters that we are aware of their concerns. We agree that if a Medicare beneficiary is appropriately referred, and admitted, to one of the approximately 400 LTCHs in the United States for a complex medical condition, the beneficiary could receive excellent medical care from a highly-trained and committed professional staff. However, we do not believe that the revisions to the SSO policy that we are finalizing will result in LTCHs going out of business or that significant services would have to be curtailed with dire consequences for beneficiaries, staff or the local medical care system. As noted elsewhere, our data indicates the aggregate margins for LTCHs were 7.8 percent for FY 2003 and 12.7 percent for 2004. When we proposed the RY 2007 change to the SSO policy, commenters also warned that the policy would result in the closure of LTCHs with disastrous effects on the health care delivery system in those areas of the country. However, after implementing the proposed changes, we have not observed any significant reduction in the number of available LTCH beds in the country. On the contrary, we continue to observe that LTCHs are opening new LTCHs. Therefore, we believe that even with decreased Medicare payments for SSO patients, such as we are envisioning based on this finalized payment policy and detailed in the Impact (see section XV. to this final rule), we believe that LTCHs will generally be able to continue delivering high quality medical care to their patients. However, we continue to believe that acute care hospitals should not be discharging patients to LTCHs without having provided a full episode of care and we also continue to have concerns about LTCHs admitting those relatively short stay patients who could otherwise be treated in acute care hospitals.

Comment: Many commenters stated that our proposed IPPS-comparable payment option under the SSO policy could discourage physicians from discharging patients from acute care hospitals and admitting them to LTCHs. Thus, they charged that we were establishing a system in which clinical judgment is trumped by determinations based solely on payment. The commenters further stated that since Start Printed Page 26910physicians discharge patients to LTCHs because it is in the patients' best interests, we would be substituting our judgment for a physician, setting a very dangerous precedent. The commenters also noted that there is available data supporting the medical determination that physicians are discharging patients to the LTCH setting because the patient's needs are better served in the LTCH setting than in an acute care hospital setting.

Response: Our objective for the revised SSO policy discussed in the RY 2008 LTCH PPS proposed is to preclude LTCHs and physicians from taking advantage of a system that significantly “overpays” (that is, relative to what would be paid for the same DRG under the IPPS) for patients that do not require the extensive resources that such high payments are intended to support. As discussed subsequently in this final rule, we recognize that some SSO cases are unavoidable due to death or an unexpected clinical improvement and early discharge. However, we have noted that in a community where both acute care and LTCH beds are available, patients are routinely transferred from the acute care hospital to the LTCH for the remainder of care because the LTCH resource is available.

As we discuss below in this section, we further compared MedPAR data on acute care hospitals regarding their LOS during CY 2003 to their LOS during CY 2005 in markets where LTCHs opened in CY 2004. We compared 304,650 acute care cases in CY 2004 to 316,816 cases in CY 2005. In CY 2003, there were 7,586 outliers, and in CY 2005, there were 5,858. The percentage of outliers in the acute care hospitals decreased from 2.5 percent to 1.8 percent and the numbers of patients that were admitted to LTCHs in those communities increased from 2,128 in CY 2003 to 6,597 in CY 2005. Furthermore, the percentage of acute care hospital discharges to LTCHs increased from 0.7 percent in CY 2003 to 2.1 percent in CY 2005. The percentage decline in total outliers between the CY 2003 and CY 2005 was −25.7 percent. The increase in LTCH discharges from CY 2003 to CY 2005 was 198.1 percent.

We are concerned that this trend has increased exponentially because it provides an acceptable disposition of the patient for the physician, and because it is an expeditious means of lowering the acute hospital's LOS and costs. We understand that the multidisciplinary approach for certain complex patients (for example, ventilator weaning) is appropriate. However, we are very concerned that the LTCH is assuming the role of the acute care hospital for many patients, at a far higher cost, which it is possible to do as long as the LTCH continues to maintain an ALOS of 25 days for purposes of qualifying for payments under the LTCH. Moreover, we do not believe that the payment policy option that we are finalizing for SSO discharges will deter physicians from delivering appropriate care to beneficiaries or from making appropriate referrals in the interests of their patients to LTCHs. Furthermore, LTCHs remain free to accept these patients. In finalizing this payment policy, we are seeking to remove any financial incentive that could encourage a LTCH to admit a patient from an acute care hospital prior to that patient having received a full episode of care at the acute care hospital.

Comment: Several commenters cited a study centered at Barlow Respiratory Hospital that charted the course of ventilator weaning treatment for 1419 medically unstable patients at 23 LTCHs from March 2002 through February 2003. The study reported that more than 50 percent of this group of patients were weaned from the ventilators and showed improvement, both neurologically and functionally. The commenters asserted that this study exemplifies the excellent level of care for such patients at LTCHs.

Response: We agree with the commenters that the results of the “Barlow” study indicate a significant rate of very positive outcomes for the very sick LTCH patients who were included in the study. In the late 1990s, we sponsored a ventilator demonstration study which included, among other acute care settings the Mayo Clinic and Temple University Hospital that also reported impressive results. Furthermore, we understand that the results of the Barlow study were used for the establishment of national ventilator-weaning protocols issued by the National Institutes of Health (NIH) and utilized by all acute care hospitals. We also understand that input from the Temple University program continues to be critical in formulating national standards. We believe that these programs established a level of excellence that should be emulated by all hospital-level facilities that treat ventilator-dependent patients, including acute care hospitals, LTCHs, and IRFs. Accordingly, we believe it is not simply the fact that the patient is treated at a LTCH that is critical to predicting positive results. Rather, it is the type of clinical intervention that is furnished to the patient at the hospital. In many cases that intervention is currently exemplified at acute care IPPS hospitals, as well as at LTCHs.

Comment: Several commenters claimed that even for what we would term “appropriate” admissions, our proposed payment option under the SSO policy that could generate an IPPS-comparable payment will erect barriers to the use of LTCHs. One commenter asserted that typical LTCH patients (described by the commenter as elderly patients with persistent multiple-system failures who are de-conditioned and protocol-resistant) respond impressively to the aggressive blending of therapeutic interventions, interdisciplinary teams, and medical intervention that is not otherwise available in the community or tertiary hospital setting. The commenter stated that from “a case rate reimbursement perspective,” grouping such a “treatment-resistant” population with the rest of the general acute care population is highly inappropriate. Some commenters asserted that even when adjusted for HCOs, acute care hospitals are not designed or intended to provide service to long-term care-type patients. The commenters emphasized that acute care hospitals are not designed to provide extended care services, unlike LTCHs, with their specially-trained expert staff and clinicians and multi-disciplinary approaches. One commenter noted that LTCHs are like acute care hospitals but must sustain a high level of care for longer periods.

Response: We disagree with the contention that acute care hospitals are not capable of providing extended hospital level care services such as the care provided in LTCHs. Although there may be communities with LTCHs where the acute care hospitals may have functionally “restricted” their services because of the presence of these LTCHs, as well as because of the financial advantages and clinical niche that they have sought to fill, acute care hospitals are equipped to provide services to the same population, and the IPPS under which they are paid, is calibrated based on the resources needed to treat those patients. Moreover, because there are over 3,500 acute care hospitals and approximately only 400 LTCHs, which are not distributed uniformly throughout the U.S. (for example, few are located in California), currently many acute care hospitals are providing care for the vast majority of Medicare beneficiaries requiring the type of care described by the these commenters. Our FY 2005 MedPAR files indicate that 20 percent of cases treated at acute care hospitals nationwide have lengths of stay between 7 and 14 days (that is, 2,386,057 out of a total of 11,855,205 Start Printed Page 26911cases). Additionally, 5.2 percent of acute care hospital cases (617,219) or have LOS greater than 14 days. In those acute care hospitals, we believe that during these longer periods those patients are receiving the same high level of care in an acute care hospital paid under the IPPS as they would receive as patients at a LTCH.

Comment: Several commenters claimed that we based our proposed revision of the SSO policy that could have resulted in an IPPS-comparable payment for a particular SSO case, on the incorrect assumption that “short stay” LTCH patients are clinically similar to short term acute care hospital patients. They stated that the SSO thresholds (5/6 of the geometric ALOS for each LTC-DRG) were never intended to be a measure of the appropriateness of a LTCH admission, but rather, were mathematically-derived from the per diem payment amounts, which were based on a methodology that would produce a payment-to-cost ratio for SSO cases close to one. Furthermore, a commenter stated the presence of a SSO patient does not indicate a premature discharge from an acute care hospital, and cited that 11 percent of the patients had previously qualified as HCOs at the referring acute care hospital.

Additionally, the commenters asserted that we are mistaken in our claim that LTCHs can foresee the LOS for patients admitted to LTCHs or predict likely deaths, where in actuality, upon admission, there is generally no substantial clinical difference between long stay and “short stay” patients. Commenters found it to be incongruous that a patient in LTC-DRG 475 (Respiratory System Diagnosis with Ventilator Support) would still be an SSO patient (for example, 28 days for LTC-DRG 475) and could be hospitalized in a LTCH for greater than 25 days (the definition of a LTCH). A case such as this could be appropriately treated in a LTCH. The commenters noted that physicians cannot and should not be asked to predict the LOS or the likely death of severely ill patients.

Commenters further asserted that we have made an erroneous assumption that LOS equates to “severity of illness” (SOI) and is a proxy for the appropriateness of an admission. However, the commenters assert that this is not the case. They outlined another incorrect belief in the proposed rule that LTCHs function like acute care hospitals when they have patients for the same LOS. On the contrary, the commenters asserted that SSO patients are being admitted because they look just like “inliers,” and we have proposed that LTCHs absorb payment rates that bear no relationship to the costs of furnishing patient care at the LTCH level.

Furthermore, based on claims analysis, using the APR-DRGs, the medical complexity and mortality rates of SSO patients, as measured by the SOI and “risk of mortality” (ROM) standards are very similar to that of the LTCH “inlier” patient population. The commenters further presented comparisons between these measures for SSO patients and for patients with the same DRGs in acute care hospitals, indicating that 52 percent of all patients admitted to LTCHs were in the highest APR-DRG ROM categories, whereas only 24 percent of acute care patients are in those same categories, resulting in a total percentage of APR-DRGs 3 and 4 at LTCHs among the SSO population that is approximately double that of acute care hospitals. The commenters noted that higher patient acuity correlates to higher utilization of facility resources, and hence, higher costs, which argues against our proposed policy that would significantly lower reimbursements for SSO cases. Several commenters also provided a comparison of case mix indices (CMI) for LTCH SSO cases and cases at acute care hospitals. The commenters asserted that SSOs at LTCHs have a relative CMI that parallels the CMI of LTCH “inlier” cases at LTCHs and which is 72 percent higher than the comparable CMI at acute care hospitals.

Response: We understand that not every SSO patient can be so identified at the time of admission to a LTCH. Further, we recognize that many patients who will eventually be defined as SSO patients because their LTCH stay is equal to or less than 5/6 of the geometric ALOS for their particular LTC-DRG, may, upon admission, present the same severity of illness and risk of mortality as “inlier” LTCH patients. As we discuss subsequently in this final rule, we selected the threshold of one standard deviation above the average LOS of an IPPS discharge as an appropriate measure to select the subset of SSO cases that are typically treated in acute care hospitals. We agree that the general SSO threshold (5/6 of the geometric ALOS for each LTC-DRG) was never meant to be a measure of the appropriateness of a LTCH admission, but rather, was mathematically-derived from the per diem payment amounts. We believe this enabled us to arrive at a reasonable payment policy at the outset of the LTCH PPS for cases that had lengths of stay significantly shorter than those patients fitting the typical profile of those who are treated at LTCHs. We recognize that a LTCH admission could be a medically-complex admission (an appropriate LTCH admission) with a relatively long LOS and still be considered an SSO case. We also acknowledge that, in some cases, LTCH admissions could also have qualified as HCOs at the referring acute care hospital. However, we still have concerns that patients in LTC-DRGs with significantly shorter stays than the ALOS for that particular DRG might have been unnecessarily admitted to the LTCH rather than receiving their care at an acute care hospital. In addition, we are adjusting the LTCH PPS to appropriately pay for those SSO stays that have a LOS that is comparable to the LOS for that DRG under the IPPS and consume far less than a full array of services in the LTCH for the particular LTC-DRG.

We believe this policy is appropriate since our data indicates a correlation between the LOS at an acute care hospital for a patient following treatment at the highest level of intensity (ICU or CCU), that is, the number of “recuperative” days, and whether or not the patient was admitted to a LTCH upon discharge from the acute care hospital. An analysis of the CY 2004 MedPAR files revealed that for the specified DRGs for acute care cases following ICU/CCU days, there were significantly fewer “recuperative” days for acute care HCO patients that were discharged and admitted to a LTCH than for those patients that were discharged directly from the acute care hospital. For example, for acute care cases in DRGs 475 (Respiratory system diagnosis with ventilator support) and DRG 483 (Trach with mechanical vent 96+ hours or PDX except face, mouth and neck diagnosis), the number of “recuperative” days were considerably shorter at the acute care hospital if there was a discharge followed by an admission to a LTCH. We believe that this data confirms MedPAC's assertion in the June 2004 Report to Congress that “patients who use LTCHs have shorter acute hospital lengths of stay than similar patients” (p. 125).

Furthermore, we agree that some SSO patients become so by virtue of death or a faster than expected recovery and early discharge, and that in certain LTC-DRGs, the SSO threshold still requires a relatively long hospital stay (for example, DRG 475, Respiratory System Diagnosis with Ventilator Support). However, in the absence of better admission criteria, we are concerned that LTCHs are admitting some SSO patients that could have received their full care at the acute care hospital or SNF-level facility. Start Printed Page 26912

We disagree with comparisons made by some commenters concerning the SOI and ROM of LTCH SSO patients to those of acute care patients based on similar lengths of stay and case-mix indices. Generally, LTCH patients that had been previously hospitalized in an acute care hospital received the diagnostic work up and major interventional treatment during that initial stay. Assuming that the patient continued to need hospital-level care after being somewhat stabilized and was discharged to a LTCH, the discharge to a LTCH could have been determined as clinically appropriate. The clinical status of this patient at this point cannot be reasonably compared to a typical patient who is treated in the acute care hospital and who is grouped to the same DRG. This is the case because the original patient has already been treated at that initial level and has required additional hospital-level care either by remaining at the acute care hospital, which would be paid for under the IPPS (perhaps as a HCO), or by being admitted to a LTCH where the stay could either be a SSO or an “inlier.” The only valid comparison of the SOIs and ROMs of two such patients in the context of the commenter's concerns would be to contrast the SOI and ROMs of the patient at the LTCH with the patient who, following the same initial intervention at the acute care hospital, continued treatment at the acute care hospital. In addition, it is not appropriate to compare the average CMI at acute care hospitals to the average CMI at LTCHs. The acute care hospital CMI is affected by a broad range of cases, so that the only appropriate comparison is between DRGs in acute care settings and DRGs in LTCHs, which is the approach we have adopted in the revised SSO policy we are finalizing in this final rule. In regions of the country where LTCHs are scarce, acute care hospitals treat the same cases that are treated in LTCHs where those facilities are available. In those areas, acute care hospitals do indeed treat the most severe cases, and the calibration of the DRG weights takes into account the resource requirements for such cases. In the light of this fact, we do not believe that it is necessary or appropriate to pay LTCHs more for cases that can be successfully treated in acute care hospitals. We understand that the option that we are finalizing, paying for some SSO stays based on the IPPS-comparable amount, will result in significant payment reductions to LTCHs for some SSO cases. However, we still believe that this modification to the SSO policy is appropriate since it ensures that payments to the LTCH are not greater than the program would pay in a different setting of care, where these patients can also be successfully treated. At the outset of the LTCH PPS, we established the SSO payment adjustment to address this distinction which we continue to believe is a valid and reasonable consideration for Medicare payments to LTCHs (67 FR 55995, August 30, 2002).

Comment: Many commenters asked that we not finalize the proposed SSO policy revisions, stating that the SSO payment option that could pay the LTCH based on an amount comparable to what would otherwise have been paid under the IPPS was not based on solid data analysis and supportable conclusions. In fact, a number of commenters asserted that the proposed policy was not based on data but rather on “erroneous and unsubstantiated assumptions” that all SSO patients are inappropriately admitted to LTCHs and inappropriately discharged from acute care hospitals. The commenters noted that, because of the way in which the policy was formulated, the percentage of LTCH cases that are paid under the SSO payment policy was a function of the SSO threshold and the dispersion of cases above and below the ALOS for the LTC-DRGs. That is, statistically, the SSO definition at 5/6 of the geometric ALOS would necessarily produce approximately 37 percent of cases as SSOs. Therefore, under the commenters belief that given the regulatory 5/6 definition of SSOs, which we had not proposed to change, the percentage of SSO cases was not amenable to change just based upon LTCHs admission policies. One commenter noted that for a significant number of patients to fall below 5/6 ALOS for a LTC-DRG is expected in a LTCH. Additionally, commenters noted that a case may qualify as a SSO because the patient has run out of covered days, regardless of the actual LOS in the LTCH and that in establishing our policy for qualifying as a LTCH (that is, meeting the average greater than 25-day LOS for a particular cost reporting period), we have recognized the “appropriateness” of including “total” rather than just “covered” days of a stay, since regardless of the payer, if the patient is still receiving hospital-level care, the facility is functioning like a LTCH. For this reason, these commenters urged us to remove such cases from the calculations we used to develop a SSO payment policy. Some commenters expressed concerns about the reliability of the data that underlay our policy proposals and asserted that our proposals are based on faulty assumptions, insufficient data, and a fundamental lack of understanding of the valuable care LTCHs provide. Moreover, the commenters asserted that LTCH patients are just not the same type of patients as acute patients; they believe that our proposed policies indicate that we are unaware of the distinction between acute care patients and patients at LTCHs. They further stated that they did not believe that the public was able to submit meaningful comments to our proposed policies because of our data flaws, our biases, and the resulting policies that we proposed.

Response: As we have stated previously, we are aware that the vast majority of LTCH patients are admitted following treatment at acute care hospitals. The patient's stay at the acute care hospital generated a Medicare payment under the IPPS, and the subsequent admission to a LTCH, an acute care hospital with an ALOS of greater than 25 days, will generate an additional Medicare payment. To protect the Medicare Trust Fund from what may be inappropriate and unnecessary payments, and to ensure that the program is not paying twice for the same episode of care, we believe it is essential that we evaluate those cases that are admitted for an unusually short stay following an initial treatment at another acute care hospital to acute care hospitals that specialize in long-stay care, since that second stay will generate another Medicare payment. In MedPAC's June 2004 Report to the Congress, the Commission stated that, “* * * Living near a LTCH increases a beneficiary's probability of using such a facility. For example, living in a market area with a LTCH quadruples the probability of LTCH use. Being hospitalized in an acute hospital with a LTCH located within the hospital also quadruples the probability that a beneficiary will use a long-term care hospital” (page 125).

Although we acknowledge that our establishment of the 5/6th of the geometric ALOS threshold, from a statistical standpoint, will result in approximately 37 percent of LTCH cases being defined as SSOs, we are extremely concerned with the number of cases that are being treated in LTCHs that fall considerably below the geometric ALOS for any given LTC-DRG. In fact, as stated previously, in the commenters' specific suggestions for how to reasonably and fairly pay SSOs, the commenters themselves drew a distinction between those cases that fall within the definition of a SSO but are more in keeping with the LOS generally Start Printed Page 26913associated with a LTCH (for example, a case assigned to LTC-DRG 482 with SSO threshold of 32.1 days, would still be paid as a SSO if the patient was treated in the LTCH for 25 days) and those cases that many commenters referred to as “very short stay outliers (VSSO)” or “very short stay discharges (VSSD).” In our revised SSO policy, the payment formula particularly takes into account our very strong belief that LTCHs are acute care hospitals that specialize in treating patients requiring “long-stay” hospital-level care.

The LTCH PPS has been designed and calibrated to pay specifically for that type of care. Since the inception of the LTCH PPS, when we established the SSO adjustment (67 FR 5594 through 55995, August 30, 2002) at § 412.529, we have provided that if a LTCH treats patients not requiring a long stay for that DRG, Medicare pays the LTCH based on the applicable payment adjustment option. Furthermore, as we revise the payment options in this final rule for the SSO policy, we continue to believe that such a payment adjustment is reasonable for all short stay patients, including those that die shortly after their admission to the LTCH. The FY 2004 MedPAR data indicates that 43 percent of all patients that die in LTCHs are deaths that occur within the first 14 days of the stay, with 35 percent of SSO deaths occurring within the first 7 days following admission. As we have since the inception of the LTCH PPS, we continue to believe that Medicare payments for those death cases occurring within the SSO threshold should be determined under the SSO policy since the length of the patient's treatment in the LTCH did not utilize the full measure of hospital resources for which the full LTC-DRG payment was calibrated.

Conversely, MedPAR data indicate that of all SSO cases, approximately 60 percent of the discharges are 14 days or less and also that acute care hospitals treat a significant percentage of patients for longer than the 5-day ALOS. (In acute care hospitals, paid under the IPPS, over 20 percent, in the aggregate, of patients that are treated have a LOS of between 14 and 7 days.) Therefore, as described below, we believe that the SSO policy that we are finalizing under the LTCH PPS provides a fair and reasonable payment, in light of our stated concerns that the short-term hospital-level care that LTCHs provide for many SSO cases may be substituting for care that could otherwise be delivered at acute care hospitals and for which at best, Medicare would otherwise pay under the IPPS.

Under § 412.507(b), Medicare will pay for inpatient care delivered only on those days that the beneficiary has coverage until the LOS exceeds the SSO threshold and becomes an inlier stay. Therefore, since the inception of the LTCH PPS, we established the distinction between “covered days” and “total days” of a LTCH stay. At the point when a patient's benefits exhaust, the patient is “discharged for payment purposes” and even though the patient may continue to be hospitalized at the LTCH, Medicare will pay only for the covered days, with the patient (or the patient's secondary insurance) being responsible for the remaining days' LTCH costs. For example, even though a patient could have been treated in an LTCH for 40 days, if upon admission, the patient only had 20 covered days remaining, for Medicare payment purposes, the stay could qualify as a SSO, unless the 20 covered days exceeded the 5/6th threshold for the LTC-DRG to which the case was grouped, at which point, the stay would become an inlier stay and a full LTC-DRG payment would be generated. Several commenters urged us to remove SSO cases occurring as a result of such lapses of Medicare coverage from our revised SSO policy but based on our data analysis, we will not be excluding benefit exhausted cases from the policy. According to FY 2005 MedPAR data, these cases constitute only 3.31 percent of SSO cases. It has been our policy since the beginning of the LTCH PPS to count those stays during which benefits are exhausted as SSOs if the covered portion of the stay is less than 5/6th of the geometric ALOS for the DRG. In this way, we appropriately determine payment based on the part A-covered stay. At the same time, we continue counting the total days of the stay for purposes of qualification as a LTCH, because that calculation is intended to reflect the length of care provided to Medicare beneficiaries. However, our policy of including total days for Medicare patients to identify hospitals qualifying (or continuing to qualify) as LTCHs indicates our recognition that conceivably, a beneficiary may be appropriately treated in a LTCH for example, for 40 days; and yet because the beneficiary had only 5 remaining benefit days, would be reported in our claims data as a 5-day SSO case. We may revisit this issue in the future and, at that time, would solicit comments to that end. However, at present, since a very small percentage of SSO cases are caused by beneficiaries exhausting benefits, the “short” SSO cases discussed above in this section, will continue to be governed by the SSO policy finalized in this rule.

Comment: One commenter expressed concern that the SSO policy would penalize LTCH providers in a situation where a patient developed a new or unexpected complication during his or her LTCH stay and required treatment that can only be provided by the referring acute care hospital.

Response: The situation to which the commenter is referring is possible and may result in a sudden discharge from a LTCH and a readmission to the acute care hospital. In such a case, if the total covered length of stay at the LTCH is less than5/6 of the LOS for the LTC-DRG to which the case is assigned, payment would be made under the SSO policy. Consequentially, the additional payment option that we are finalizing could also be applicable if the covered LOS at the LTCH fell within the IPPS-comparable threshold prior to discharge. Such payment would be appropriate because the patient would have received less than a full episode of care at the LTCH prior to being discharged back to the acute care hospital. We note that should the patient subsequently be discharged from the acute and readmitted to the LTCH to continue treatment begun before the acute episode, Medicare payment to the LTCH would be governed under our interrupted stay policy at § 412.531. We would also note that this stay could also be subject to adjustment under the SSO policy (including the payment option that we are finalizing) depending upon the total covered length of stay (both prior to and following the acute episode).

Comment: Many commenters stated that their objections to the policy discussed in the proposed rule extended to the existing SSO payment policy with which they have expressed disagreement in the past. Several of these commenters asserted that the current SSO threshold (5/6 of the geometric ALOS for each LTC-DRG) is not statistically justifiable. These commenters recommended that, if we are going to employ LOS as the only criterion for determining SSOs, we should logically select a threshold that better identifies cases that are dissimilar to the median or average, such as the 5th percentile through 10th percentile.

Response: We believe that the policy we are adopting in this final rule is a consistent extension of the principles that we have employed in developing the SSO payment policy. In this rulemaking cycle, we have not introduced any discussion or proposals concerning the existing SSO threshold, and therefore, we are not implementing Start Printed Page 26914the commenters' recommendation that we establish a dramatically-revised threshold level. However, we did provide an exhaustive discussion of the reasons for adopting this threshold in the FY 2003 LTCH PPS final rule (67 FR 55995), which included statistical analysis, various simulations, regressions, and consideration of various options.

Comment: Several commenters stated that the objective of the SSO policy that we discussed in the RY 2008 LTCH PPS proposed rule is to establish a de facto exclusionary policy, prohibiting the admission of these patients to LTCHs by means of a payment mechanism rather than careful clinical review.

Response: We disagree that we are establishing an exclusionary policy. On the basis of analysis that we presented in the RY 2008 LTCH PPS proposed rule and previously in this final rule, we believe that many of these cases may represent “premature and inappropriate discharge from the acute care hospital and inappropriate admission to the LTCH” (72 FR 4840). The intent of this policy is to establish an appropriate payment level for this class of cases. Hospitals remain free to accept these patients. As we stated in the RY 2008 LTCH PPS proposed rule, * * * a short stay case at a LTCH most likely did not receive a full course of medical treatment during the short stay and* * * a full LTC-DRG payment would therefore, be inappropriate” (72 FR 4804).

Comment: Several commenters objected that the policy we discussed could apply to cases whose length of stay exceeds 25 days, the ALOS required for a hospital to qualify as an LTCH. Commenters indicated that at least 9 IPPS DRGs have an ALOS plus one standard deviation that is greater than 25 days, and at least 26 other IPPS DRGs have an ALOS plus one standard deviation that exceed 20 days. Commenters contended that cases exceeding the 25-day threshold for qualifying as an LTCH should not be considered short stay cases.

Response: We do not believe that it is inappropriate for individual cases that exceed the ALOS threshold for LTCH status to be considered SSOs. In fact, we have treated some such cases as SSOs since the establishment of the SSO policy. For a number of LTC-DRGs, the SSO threshold,5/6 of the geometric ALOS, significantly exceeds 25 days. These include DRGs 498, 499, 520, and others. Similarly, a number of IPPS DRGs have an ALOS plus one standard deviation that is greater than 25 days. As a result, many cases with lengths of stay shorter than 25 days receive payment under the SSO methodology, and a subset of those cases will be paid specifically under the formula that we are adopting in this final rule for certain cases: For SSO cases with a length of stay less than ALOS plus one standard deviation of the IPPS DRG, payment will be no greater than the IPPS comparable amount that we have defined. These results are appropriate because the respective thresholds serve different purposes. The 25-day threshold defines an ALOS established by the statute to define a LTCH. The respective outlier thresholds (the basic SSO threshold of5/6 of the geometric LTC-DRG ALOS, and the threshold that we are now adopting to identify every SSOs) serve to identify subsets of LTCH cases for appropriate payment treatment, based on comparisons to relevantly similar cases. We have explained the basis for adopting the SSO threshold in the FY 2003 LTCH PPS final rule (67 FR 55995). The threshold that we are adopting in this final rule, the geometric ALOS plus one standard deviation of the IPPS DRG, selects a subset of SSOs that are similar to cases successfully treated in short-stay acute care hospitals. Since these cases have received a course of treatment similar to the typical course of treatment in an IPPS hospital, we are limiting payment for them to an amount no greater than the comparable payment under the IPPS.

Comment: Several commenters stated that we had not presented any conclusive financial or clinical evidence to support the policy discussed in the RY 2008 LTCH PPS proposed rule, but that we instead rely merely on statements such as: “many LTCHs appear to be admitting some SSO patients that could have received the care at the acute care hospital.” (72 FR 4806) (Emphasis supplied by commenter.) Furthermore, a commenter stated that our own expert consultant, RTI, had failed to find evidence conclusively illustrating that the typical LTCH SSO patient could be treated as effectively in an acute care hospital. Some of these commenters also maintained that, contrary to our suggestions, the care received by patients at LTCHs is often unique and not available at acute care hospitals. Commenters cited physicians who were consulted on the clinical aspects of transfer from an acute care hospital to a LTCH. These physicians provided numerous explanations and scenarios detailing how LTCHs provide different kinds of services even if the DRG for a case is nominally the same.

Response: As we have discussed elsewhere in this final rule, LTCHs are certified as acute care hospitals and acute care hospitals paid under the IPPS are throughout the country treating beneficiaries requiring hospital-level care lengths of stay comparable to those that are typical of LTCHs. We disagree with commenters who imply that there is a clear distinction between the patients that are appropriate for successful treatment at LTCHs and patients that are appropriately and successfully treated at acute care hospitals. Across the United States, the nearly 3,600 acute care hospitals that discharge approximately 12.7 million Medicare beneficiaries treat the full range of medical issues that the commenters identify as LTCH cases. We do not question that many LTCHs have highly regarded reputations for their success in treating respiratory and ventilator cases (MS-LTC-DRGs 207 and 208). However, as detailed in the RTI report, the 2004 MedPAR files indicate that where LTCHs treated 13,394 cases assigned to DRG 475 in 2004, acute care hospitals treated 18,727 Medicare patients with an additional 7,072 HCOs in DRG 475. For DRG 88, Chronic obstructive pulmonary disease (COPD), LTCHs treated 4,894 cases where acute care hospitals treated 37,523 cases. Data on other common DRGs treated in LTCHs as compared to the same DRGs treated in acute care hospitals reflect a similar pattern, particularly among the DRGs that could fall into the broad category of “medically complex” patients, which are the majority of LTCH patients (Table 3-2, RTI report, p. 35. We understand that MedPAC and RTI have noted that many LTCHs deliver a high level of care to very sick Medicare beneficiaries, with fine doctors, exemplary nursing care, and top-notch rehabilitation therapists, but we also know that many acute care hospitals throughout the nation are treating the same patients and similarly delivering excellent care, especially where there are few LTCHs. We also know that some LTCHs specialize in a particular subset of patients and achieve a noteworthy success in their treatment (for example, of patients requiring ventilator weaning or wound care). However, similar patients are also receiving care in acute care hospitals. Therefore, we cannot agree with commenters implying that acute care hospitals are incapable of competently treating Medicare beneficiaries that happen to fall within the DRGs that LTCH identify as their specialties and that any patients falling into such categories would receive “substandard” care at an acute care hospital. Start Printed Page 26915

Comment: Several commenters stated that our proposed policy should not apply to cases that were HCOs at an acute care hospital prior to transfer to a LTCH. Since such cases received the full complement of services at the acute care hospital, and the acute care hospital actually incurred significant losses before receiving an outlier payment from the Medicare program, it cannot be stated that any discharge and transfer to a LTCH was premature and inappropriate.

Response: We agree that, in such cases, the transfer to a LTCH is unlikely to be premature and inappropriate. In fact, typically, HCO cases in the acute care setting represent a full course of treatment in that setting. However, as our discussion in the RY 2008 LTCH PPS proposed rule indicates, this is not the only, or even the primary, factor that deserves consideration in determining an appropriate SSO payment level. Regardless of whether a case had reached outlier status in an acute care hospital prior to transfer to a LTCH, the course of treatment at the LTCH could more closely resemble the normal course of treatment at an acute care hospital than the normal course of treatment for cases at a LTCH. We stated in the RY 2008 LTCH PPS proposed rule that cases “with lengths of stay that are equal to or less than the IPPS ALOS plus one standard deviation for the same DRGs under the IPPS appear to be comparable to typical stays at acute care hospitals'' and “LTCHs that admit SSO patients with lengths of stay more typical of an acute care hospital may be, in fact, behaving like acute care hospitals'' (72 FR 4806 citing 71 FR 27847). For purposes of the SSO policy discussed in the RY 2008 LTCH PPS proposed rule, the issue is primarily the course of treatment actually received at the LTCH, rather than the course of treatment at the acute care hospital prior to transfer to a LTCH. Of course, one reason the course of treatment at a LTCH may resemble the normal course of treatment at an acute care hospital may be that an acute care hospital has prematurely and inappropriately transferred a patient to a LTCH. However, in cases where a patient has received a high level of treatment at an acute care hospital, including levels of treatment that qualify for outlier payments, a subsequent stay in an LTCH may still “be comparable to typical stays at acute care hospitals.” (72 FR 4806) In these cases, since we believe the Congress excluded LTCHs from the IPPS because cases with longer lengths of stay (as compared to acute care hospitals paid under the IPPS) tend to be costlier than cases with shorter stays, we do not believe that it would be appropriate for the program to pay an LTCH an unadjusted LTCH PPS payment for case with such an abbreviated stay that it did not receive the full course of treatment particularly when we would pay a much lower amount in to an acute care hospital for a similar course of treatment.

Comment: Several commenters urged us not to apply the policy we discussed to cases in which patients die in the hospital. These commenters noted that physicians and hospitals are not able to predict which patients will die subsequent to admission to an LTCH. In addition, many of these patients are high cost, requiring significant medical resources in the last days of life. One LTCH commenter determined that about 50 percent of its extreme SSOs were discharged due to death. The commenter notes that it may not be appropriate for these cases to receive a full LTCH payment, but that it is equally unfair for CMS to assume “sinister intent” and to financially penalize LTCHs operating in good faith. Some commenters emphasized generally that adoption of the revised SSO policy that we discussed would be unfair to LTCHs because they cannot predict in advance who will become SSO cases. There are several reasons why a patient could become an SSO including the patient dying or leaving against medical advice. Many of these commenters noted that if this policy is adopted, LTCHs will only receive, at best, costs for SSO cases. Other commenters recommended that, if we adopt this policy, it should incorporate outlier payments when determining an equivalent IPPS payment amount in the SSO payment methodology.

Response: We certainly acknowledge that hospitals and physicians are not able to predict with certainty at admission which patients will die during an inpatient stay in a LTCH, or whether a patient will leave against medical advice. However, the issue with regard to these cases, as with the cases discussed in the previous comment, is that “lengths of stay that are equal to or less than the IPPS ALOS plus one standard deviation for the same DRGs under the IPPS appear to be comparable to typical stays at acute care hospitals.” The point is not to penalize LTCHs, but rather, to pay appropriately for cases that receive less than the full course of treatment at a LTCH. Even when a patient dies in a LTCH, whether unexpectedly or not, cases with lengths of stay more typical of an acute care hospital are not receiving the full course of treatment in a LTCH, and resemble more the course of treatment in acute care hospitals. It is therefore appropriate to limit the payment for such cases accordingly. We would also like to note that where a LTCH is finding that nearly half of its patients are discharged due to death, if in fact many of these patients are SSO cases, the LTCH may need to consider whether those patients were too fragile to be transferred from the acute care hospital to the LTCH. Transfer trauma is a serious issue that must be considered whenever a hospital considers transferring a patient to another facility.

With respect to the recommendation that we take outlier payments into account when determining the equivalent IPPS payment amount in the SSO payment methodology, under existing LTCH PPS policy, a SSO case that meets the criteria for a LTCH PPS HCO payment at § 412.525(a)(1) (that is, if the estimated costs of the case exceed the adjusted LTC-DRG SSO payment plus the fixed loss amount) would receive an additional payment under the LTCH PPS HCO policy at § 412.525(a) (67 FR 56026, August 30, 2002). For purposes of HCOs under the proposed SSO policy, we would continue to use a fixed-loss amount calculated under § 412.525(a), and not a fixed-loss amount based on § 412.80(a). Medicare would pay the LTCH 80 percent of the costs of the case that exceed the sum of the applicable option of the least of the four proposed payment options, described above, and the fixed-loss amount determined under § 412.525(a).

Comment: Several commenters stated that the payment reductions associated with the very short SSO policy discussed in the RY 2008 LTCH PPS proposed rule violate the principles of a PPS in which some cases are expected to cost less than others.

Response: We disagree that these policies violate the principles of averaging found in a PPS. As we stated in the RY 2007 LTCH PPS final rule, “* * *we believe it is very important to evaluate the adjustment in light of the fact that in a PPS there are numerous principles that we try to balance simultaneously when making policy decisions. Among these principles are appropriate payment, predictability, averaging, beneficiary access to appropriate care, and equity so that while the averaging principle is an important one in PPSs, it is not the only principle that guides our policy decisions. For example, in the case of SSOs and HCOs, we must determine how to appropriately to pay for aberrant cases that are much shorter (that is, SSOs) and much costlier (that is, HCOs) when compared to typical cases in the Start Printed Page 26916relevant LTC-DRG. In the case of short stays, if we failed to adjust the payment to reflect that the case did not receive the full resources of a typical LTCH stay for the particular DRG, the PPS would be greatly “overpaying” for the stay, could serve as an incentive to game the system, and would also waste valuable Medicare Trust Fund dollars. Similarly, in the case of HCOs, if we did not adjust the payment to reflect the extraordinary high costs that LTCH was incurring for treating a particular patient when compared to a typical case in the respective LTC-DRG, we would be “underpaying” significantly for the case. We have stated that providing additional money for HCOs strongly improves the accuracy of the payment system as well as reduces the incentive to under serve these patients. Since we do not pay SSOs or HCOs an amount paid to “inliers'/cases that have length of stays or costs commensurate with other cases in the respective but instead make payment adjustments to reflect the unique circumstances of these cases, the averaging principle is less heavily emphasized under these circumstances to achieve equity, appropriate payments that accurately reflect resource costs at the patient and hospital level, and beneficiary access to medical care.”

We believe that, given that LTCHs are defined as acute care hospitals that have an average inpatient LOS of greater than 25 days, the payment policies under the LTCH PPS appropriately reflect the averaging principle. That is, where some cases, within the “inlier” range will have generated relatively lower costs, other cases will generate higher costs and Medicare will pay a LTCH the same for both less and more costly cases. The SSO policy, along with the HCO policy addresses payments for cases that fall outside of the normal types of averaging in the inlier range in the PPS and ensures that payment for SSO cases is not greatly in excess of the resources required to treat those cases. (71 FR 27866 through 27867)

Comment: Some commenters asked that we comment on why the IPPS post-acute transfer policy does not appropriately adjust for payment when transferred cases ultimately become SSO discharges in the LTCH setting. Another commenter suggested that, we provide policies under the acute IPPS side to address inappropriate, or early discharges and asked that post-acute transfer rules, readmission rules and DRGs for acute care hospitals should be used to minimize the issue instead of penalizing LTCHs.

Response: We note that we addressed the effect of the post-acute transfer policy on SSOs previously in the RY 2003 LTCH PPS final rule, but will reiterate that the IPPS post-acute transfer provision was created to address cases in which the transferring acute hospital provides less than the full spectrum of care for the qualified DRG and to avoid providing an incentive for a hospital to transfer a patient to another hospital early in the patient's stay to minimize costs while still receiving the full DRG payment. The post-acute transfer policy only addresses the appropriate level of payments for the course of treatment received in an acute care hospital. It does not address the appropriate level of payments at the facility to which the patients are then transferred.

We note that the post-acute care transfer policy only affects DRGs that meet the criteria at § 412.4. Although we expect the post-acute transfer policy to have some impact on the discharge behavior of acute care hospitals because of the reduced payments that they will receive for qualified discharges, the post-acute transfer policy does not necessarily affect the issues being addressed by the SSO policy change. Both the IPPS post-acute transfer policy and the revised SSO policy being finalized in this rule are designed to ensure that Medicare payments are appropriate given the types of treatment provided in each setting; we note that in the instance of an acute transfer (that is subject to the post-acute transfer policy) to an LTCH that discharges the patient as an SSO, neither the acute nor the LTCH facility provided the full episode of care to the patient and it would not be appropriate to pay either facility a full DRG payment. We believe that the revised payment formula for SSO patients that we are finalizing will appropriately pay LTCHs for delivering services to patients who do not otherwise require the lengths of stay that are characteristic of LTCHs. The SSO policy will address payments to LTCHs for patients discharged from the acute care hospital even after the IPPS geometric ALOS, who are subsequently discharged from the LTCH as a short SSO.

Comment: Two commenters suggested that rather than challenging the cases that are admitted from acute care hospitals, we should be more concerned about inappropriate admittances from nonhospital settings such as SNFs or elsewhere.

Response: After analyzing recent data, we note that approximately 80 percent of the patients admitted to the LTCHs come from the short term acute care hospitals and only 20 percent are admitted from other nonhospital settings. Since SNFs do not offer hospital-level care but are still serving patients with compromised health, we believe that a decision to transport a SNF patient to a hospital would generally be made because the patient appears to the medical professionals at the SNF to be in need of a higher level of medical treatment or care than is available at the SNF. (In fact, such patients would typically be admitted to the acute care hospital rather than to a LTCH.) However, both an acute care hospital and a LTCH offer acute hospital-level care. As discussed previously in this final rule, we are very concerned about the treatment of a short-stay patient who could reasonably and effectively continue to be treated in an acute care hospital and paid for under the IPPS, being admitted unnecessarily to a LTCH, which specializes in treating patients requiring long-term hospital-level care and paid for under a PPS which has been calibrated based upon the high resource use associated with long patient stays. Furthermore, admission of such a patient could also result in an unnecessary and inappropriate LTCH hospitalization, which would also result in a second Medicare payment under the LTCH PPS for what was essentially, one episode of care.

Comment: Several commenters believe that we are incorrect that LTCHs could be admitting patients not requiring long stays, noting that LTCHs actually have a disincentive to admit short stay patients because LTCH certification status can be at risk if the hospital does not maintain an ALOS of more than 25 days.

Response: Under the TEFRA system, all inpatient days (whether covered by Medicare or not) were included in the LOS computation, and the mathematical determination was based upon the number of patient days, during the cost reporting period when they occurred, divided by discharges occurring during that same period of time (67 FR 55954, 55971). With the establishment of the per discharge LTCH PPS, we restricted the patient count for purposes of qualifying as a LTCH solely to Medicare patients (67 FR 55971), and we implemented the policy of ‘days following the discharges,’ under which, if a patient's stay crosses two cost reporting periods, the total days of that stay (both covered and non-covered days) would be included in the computation during the cost-reporting period that the discharge occurred (69 FR 25706).

LTCH cost report data reveal that the general ALOS of most LTCHs varies only slightly. Generally, LTCHs maintain an ALOS that is just over 25 Start Printed Page 26917days, meeting the statutory definition of a LTCH, that is, having an ALOS of greater than 25 days. Furthermore, we understand that LTCHs closely monitor their yearly ALOS and that one extremely long-stay case can mathematically offset for a number of short-stay cases. After studying the hospital-specific data, we believe that this is indeed the case for many LTCHs. We also believe that the payment policy that has been utilized since the start of the LTCH PPS for FY 2003 has not operated as a financial disincentive for the admission of patients who will not ultimately require long-stay hospital-level care. In fact, we note that MedPAR data show approximately 27,000 SSO cases with a LOS of 14 days or less. This indicates that even with over 20 percent of their discharges having such a short ALOS, LTCHs have maintained their greater than 25-day statutory ALOS. Therefore, we believe that it is both possible for a LTCH to maintain its designation and also admit many very short stay cases.

Comment: Several commenters maintained that the SSO policy we discussed would have unintended effect of lengthening patients stay. Some of these commenters specifically noted that this effect could be the result of a payment “cliff” where payments rise abruptly once the threshold for the application of this policy (the ALOS of the IPPS DRG plus one standard deviation) is reached. The commenters believe that the proposed rule introduced “backwards” incentives associated with the old “cost-based” system. Policies will result in encouraging a profit for longer stays, which could raise costs to the Medicare program.

Response: We acknowledge that there could be such a cliff effect in some cases as a result of the policy that we are adopting. However, we believe that the merits of adopting this limitation on outlier payments in certain cases outweighs the risks of some possible, unintended consequences. We will monitor experience under the new policy to detect whether there is an inappropriate increase in lengths of stay that are slightly greater than the ALOS plus one standard deviation of the comparable IPPS DRGs. As part of our program integrity responsibilities, we may ask the FIs to review the medical necessity of the last few days of a LTCH stay that just exceeds the threshold, and if some days are determined not to be “medically necessary,” then if the remaining days result in a LOS lower than the threshold, the stay may be paid at the IPPS comparable rate.

Comment: Some commenters contended that the concerns behind the possible revision to the SSO policy could be more appropriately addressed by establishing patient criteria and QIO review of medical necessity for admissions, as has been recommended by MedPAC and RTI.

Response: Under our QIO program, QIOs review services to determine whether services are reasonable and medically-necessary, whether the quality of services meets professionally-recognized standards, and whether services in an inpatient hospital or other inpatient health care facility could, consistent with the provision of appropriate medical care, be effectively provided more economically on an outpatient basis or in an inpatient facility of a different type. We have not historically interpreted any of these areas of review to involve determinations of which kind of acute care facility would be appropriate, and QIOs do not regard short term acute care hospitals and LTCHs as facilities “of a different type.” A QIO uses criteria, based on typical patterns of practice. The QIOs also consult with (a) physician(s) and practitioner(s) actively engaged in practice in that State and to the extent possible, in the same specialty, when making the determination that care was or was not medically-necessary. Although a QIO review can detect whether or not the patient requires an acute level of care or whether care in a SNF would have been appropriate, since both acute care hospitals and LTCHs are certified as acute care hospitals, QIOs do not make the distinction between whether a patient should be hospitalized at an acute care hospital or at a LTCH, so long as the patient requires an acute level of care.

QIOs are authorized by statute to determine whether, in case such services and items are proposed to be provided in a hospital or other health care facility on an inpatient basis, such services and items could, consistent with the provision of appropriate medical care, be effectively provided more economically on an outpatient basis or in an inpatient health care facility of a different type as specified in section 1154(a)(1)(C) of the Act. Therefore, QIOs have authority to determine the appropriate hospital-level setting in the face of objective criteria. But there is no objective criteria distinguishing between settings where acute care is delivered. Since the statute states “a facility of a different type,” and because short term acute care hospitals and LTCHs are very similar and provide the same level of care, we have at no time interpreted “a facility of a different type” in section 1154(a)(1)(C) of the Act to mean that QIOs must distinguish between them.

In a memorandum issued to the Regional Offices, Chief Executive Officers, and all QIOs, from the Director of the Quality Improvement Group of the CMS Office of Clinical Standards on October 28, 2004, among other matters, the following policy was further clarified:

Note:

there are different provider types that may offer the same level of intensity of inpatient care. QIOs do not specify which provider type should be used when the level of intensity is the same. For example, a patient requires an acute level of care that could be delivered in a short—term acute care PPS hospital, a long-term care hospital or an acute rehabilitation hospital. The QIO determines what intensity of care is appropriate (that is, the patient requires an acute level of care) but would not specify as a matter of admission necessity which provider type the patient should be admitted to. If the QIO determines that there is a quality of care concern implicated, that issue should be addressed through the quality review process.

Under current contracts, QIOs review LTCH cases under the following circumstances: When a claim is selected for purposes of determining or lowering the payment error rate; if there is a QIO-identified need to perform additional review based on their contractual responsibilities; if there is an immediate appeal of certain beneficiary notices; as a result of the referral of a case or cases; or when there is a beneficiary complaint or other quality of care concern.

Since one of the recommendations made by MedPAC in their June 2004 Report to Congress was for an increased role for the QIOs in monitoring criteria to assure that LTCHs are treating appropriate patients, researchers from RTI have been in contact with several QIOs nationwide in order to evaluate their role. However, involving QIOs in the on-going determination of the appropriateness of admissions, continuing stay or discharge for a significant proportion of LTCH patients was never envisioned when the QIO program was established. There will not be a reassignment of Medicare funds to QIOs from the LTCH PPS. However, we are currently developing the next Quality Improvement Organization Scope of Work. These comments will be considered in that process.

After consideration of the numerous comments submitted on this issue, we are finalizing the policy that we discussed in the proposed rule. That is, in SSO cases where the covered LOS is equal to or less than the “IPPS Start Printed Page 26918comparable threshold” (defined above in this section) of the same DRG under the IPPS, the SSO payment methodology will be based upon the least of the following: 100 Percent of estimated costs of the case as determined under § 412.529(d)(2); 120 percent of the LTC-DRG per diem multiplied by the covered LOS of the case as determined under § 412.529(d)(1); the Federal prospective payment for the LTC-DRG as determined under § 412.529(d)(3); or an LTCH PPS amount comparable to the IPPS per diem.

Technical Correction

We are making a technical correction to existing § 412.529(a) which would add the term “covered” immediately before the phrase “length of stay” in the initial definition of a SSO case. This technical correction is not a substantive policy change but rather corrects the regulatory definition of a SSO case so that it is consistent with policy determinations that we have made since the FY 2003 implementation of the LTCH PPS. We would note that utilizing only Medicare covered days for payment purposes has been our policy from the outset of the LTCH PPS, as is specified at § 412.503 where we defined “discharge” for purposes of payment, as “* * * when the patient stops receiving Medicare-covered long-term care services * * *.” Furthermore, in subsequent revisions of our SSO policy, we included the term “covered” at § 412.529(c)(2)(iv)(A), § 412.529(d)(1) and § 412.529(d)(4)(i)(B). We are making this technical correction to conform all references at § 412.529 to our existing policy regarding a SSO discharge which is determined based on the number of “covered” days in the patient stay.

3. Determination of Cost-to-Charge Ratios (CCRs)

In the FY 2007 IPPS final rule (71 FR 48117 through 48121), similar to the revisions to the HCO policy as discussed in IV.D.3.d. of the preamble of this final rule, we revised our methodology for determining the annual CCR ceiling and Statewide average CCRs under the LTCH PPS because we believe that those changes are more consistent with the LTCH PPS single payment rate for inpatient operating and capital costs. Under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, for discharges occurring on or after October 1, 2006, the LTCH CCR ceiling specified under § 412.529(c)(3)(iv)(C)(2) is calculated as three standard deviations above the corresponding national geometric mean total CCR (established and published annually by CMS). (As discussed in greater detail in this section, the fiscal intermediary (FI) may use a Statewide average CCR if, among other things, a LTCH's CCR is in excess of the LTCH CCR ceiling.) The LTCH total CCR ceiling is determined based on IPPS CCR data, by first calculating the “total” (that is, operating and capital) IPPS CCR for each IPPS hospital and then determining the average “total” IPPS CCR for all hospitals. The LTCH CCR ceiling is then established at 3 standard deviations from the corresponding national geometric mean total CCR. (For further detail on our methodology for annually determining the LTCH CCR ceiling, refer to the FY 2007 IPPS final rule (71 FR 48117 through 48119).) We also established that the LTCH “total” CCR ceiling used under the LTCH PPS will continue to be published annually in the IPPS proposed and final rules, and the public should continue to consult the annual IPPS proposed and final rules for changes to the LTCH total CCR ceiling that would be effective for discharges occurring on or after October 1 each year. Accordingly, in the FY 2007 IPPS final rule (71 FR 48119), we established a FY 2007 LTCH total CCR ceiling of 1.321, effective for discharges occurring on or after October 1, 2006.

In addition, under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, for discharges on or after October 1, 2006, we revised our methodology to determine the Statewide average CCRs under § 412.529(c)(3)(iv)(C) for use under the LTCH PPS in a manner similar to the way we compute the “total” LTCH CCR ceiling using IPPS CCR data (71 FR 48120). Specifically, under this revised methodology, we first calculate the total (that is, operating and capital) CCR for each IPPS hospital. We would then calculate a weighted average “total” CCR for all IPPS hospitals in the rural areas of the State and weighted average “total” CCR for all IPPS hospitals in the urban areas of the State. (For further detail on our methodology for annually determining the LTCH urban and rural Statewide average CCRs, refer to the FY 2007 IPPS final rule (71 FR 48119 through 48121).) We also established that the applicable Statewide average “total” (operating and capital) CCRs used under the LTCH PPS will continue to be published annually in the IPPS proposed and final rules, and the public should continue to consult the annual IPPS proposed and final rules for changes to the applicable Statewide average total CCRs that would be effective for discharges occurring on or after October 1 each year. Accordingly, in the FY 2007 IPPS final rule (71 FR 48122), the FY 2007 LTCH PPS Statewide average total CCRs for urban and rural hospitals, effective for discharges occurring on or after October 1, 2006, were presented in Table 8C of the Addendum of that final rule (71 FR 48303).

Additionally, in the FY 2007 IPPS final rule (71 FR 48119), under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we established under the LTCH PPS SSO policy at § 412.529(c)(3)(iv)(C) that the FI may use a Statewide average CCR, which is established annually by CMS, if it is unable to determine an accurate CCR for a LTCH in one of the following three circumstances: (1) New LTCHs that have not yet submitted their first Medicare cost report (for this purpose, a new LTCH would be defined as an entity that has not accepted assignment of an existing hospital's provider agreement in accordance with § 489.18); (2) LTCHs whose CCR is in excess of the LTCH CCR ceiling; and (3) other LTCHs for whom data with which to calculate a CCR are not available (for example, missing or faulty data). Other sources of data that the FI may consider in determining a LTCH's CCR included data from a different cost reporting period for the LTCH, data from the cost reporting period preceding the period in which the hospital began to be paid as a LTCH (that is, the period of at least 6 months that it was paid as a short-term acute care hospital), or data from other comparable LTCHs, such as LTCHs in the same chain or in the same region.

Furthermore, in the FY 2007 IPPS final rule (71 FR 48121), we established under § 412.529(c)(3)(iv)(B) that, for discharges occurring on or after October 1, 2006, the CCR applied at the time a claim is processed will be based on either the most recently settled cost report or the most recent tentatively settled cost report, whichever is from the latest cost reporting period. Under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, in that same final rule, we also established at § 412.529(c)(3)(iv)(A) that, for discharges occurring on or after October 1, 2006, we may specify an alternative to the CCR computed under § 412.529(c)(3)(iv)(B) (that is, computed from the most recently settled cost report or the most recent tentatively settled cost report, whichever is later), or a hospital may also request that the FI use a different (higher or lower) CCR based on substantial evidence presented by the hospital. A complete discussion of these revisions to our methodology for determining a LTCH's CCR is Start Printed Page 26919discussed in the FY 2007 IPPS final rule (71 FR 48119 through 48121).

4. Reconciliation of SSO Cases

In the FY 2007 IPPS final rule (71 FR 48121 through 48122), under the broad authority of section 123 of the BBRA and section 307(b)(1) of BIPA, we revised § 412.529(c)(3)(iv) (D) through (E), for discharges occurring on or after October 1, 2006, to codify in subpart O of 42 CFR part 412 the provisions concerning the reconciliation of LTCH PPS outlier payments, including editorial clarifications discussed in greater detail below in this section, that would more precisely describe the application of those policies.

Specifically, at § 412.529(c)(3)(iv)(D), similar to our current policy, we specified that for discharges occurring on or after October 1, 2006, any reconciliation of outlier payments will be based on the CCR calculated based on a ratio of costs to charges computed from the relevant cost report and charge data determined at the time the cost report coinciding with the discharge is settled. In addition, at § 412.529(c)(3)(iv)(E), we specified that for discharges occurring on or after October 1, 2006, at the time of any reconciliation, outlier payments may be adjusted to account for the time value of any underpayments or overpayments. Such an adjustment will be based upon a widely available index to be established in advance by the Secretary and will be applied from the midpoint of the cost reporting period to the date of reconciliation. We made these additional revisions to § 412.529(c)(3) because we believe that these changes would be more consistent with the LTCH PPS single payment rate, and because we believe it would be more appropriate and administratively simpler to include all of the regulatory provisions concerning the determination of LTCH PPS outlier payments applicable under the LTCH PPS regulations at subpart O of 42 CFR part 412. (For a complete discussion on the revisions made to the SSO reconciliation policy, refer to the FY 2007 IPPS final rule (71 FR 48121 through 48122).)

Comment: One commenter requested that we clarify how we interpret the 10 percentage point criterion of the SSO and HCO reconciliation policy.

Response: We did not propose any changes to the current reconciliation policy. Therefore, we do not believe this final rule is the appropriate vehicle to address this comment. As we have stated, we intend to issue subregulatory guidance on LTCH reconciliation that would be similar to the IPPS reconciliation process and would address the commenter's question.

B. Expansion of Special Payment Provisions for LTCH Hospitals Within Hospitals (HwHs) and LTCH Satellites: Expansion of the 25 Percent Rule to Certain Situations Not Currently Covered Under Existing § 412.534

In the FY 2005 IPPS final rule, we established the special payment provisions at § 412.534 for LTCHs that are HwHs and for satellites of LTCHs that are co-located with host hospitals. In developing that policy, we were particularly concerned with patient shifting between the host acute care hospitals and the co-located LTCH HwH or satellite for financial rather than for medical reasons, a scenario that we believed was encouraged by physical proximity, and that resulted in inappropriate increased cost to the Medicare program (69 FR 49191). We specified that the payment adjustment for co-located LTCHs at § 412.534 was also applicable to host hospitals other than acute care hospitals that served as hosts to LTCH HwHs or satellites of LTCHs since we had similar concerns to those stated above regarding patient shifting between such hosts and their co-located LTCHs. However, the vast majority of host hospitals continue to be acute care hospitals (69 FR 49198).

In the FY 2005 IPPS final rule, we quoted the FY 1995 IPPS final rule where we first discussed our concern that LTCH HwHs were, in effect, operating as step-down units of acute care hospitals. We explained that this was inconsistent with the statutory framework and that such a configuration could lead to Medicare making one payment to the acute care hospital and another under LTCH PPS for what was essentially one episode of care (69 FR 49191 through 49192, and 59 FR 45389).

When we first established the separateness and control criteria for LTCH HwHs at § 412.22(e) in the FY 1995 IPPS final rule, our main objective was to address the shifting of costly, long-stay patients from the host to the on-site LTCH, resulting in two hospital stays which would result in a financial windfall for both providers. We sought to protect the integrity of the IPPS by ensuring that those costly, long-stay patients who could reasonably continue treatment in an acute care hospital would not be unnecessarily discharged to an onsite LTCH, a behavior that would undermine the Medicare IPPS DRG payment system for acute care hospitals. We explained that the Federal standardized payment amount for the IPPS was based on the average cost of an acute care patient across all acute care hospitals for the base year. This is premised on the assumption that, on average, both high-cost and low-cost patients are treated at hospitals. Although Medicare may pay a hospital less than was expended by the hospital for a particular costly case, the hospital could also receive more than it expended for other, less costly cases. However, an acute care hospital that consistently discharges higher cost patients to a post-acute care setting for the purpose of lowering its costs, undercuts the foundation of the IPPS DRG payment system which is based on averages, as noted above. Because the course of acute treatment had not been completed, the hospital inappropriately would have incurred lower costs under the IPPS. It did not incur additional costs for what would have been the remainder of the patient's stay at the IPPS acute care hospital. We were concerned that once that patient was discharged from the IPPS acute care hospital, the patient, still under active treatment for the same condition, would be admitted to a LTCH, thereby generating a second admission and Medicare payment that often would not have taken place but for the availability of the LTCH (59 FR 45389 through 45393).

With the growth of satellites of excluded hospitals, another category of co-located facilities, we established “separateness and control” policies applicable to satellites, which we defined at § 412.22(h) as “a part of a hospital that provides inpatient services in a building also used by another hospital or in one or more entire buildings located on the same campus as buildings used by another hospital.” In the FY 2003 IPPS final rule at § 412.22(h), we finalized additional regulations governing the satellites of hospitals (64 FR 41532 through 41535 and 67 FR 50105 through 50106).

As detailed in the FY 2005 proposed and final rules for the IPPS (69 FR 28323 through 28327, 69 FR 49191 through 49214), with the explosive growth in the number of LTCH HwHs and concomitant cost to the Medicare program, we reevaluated the effectiveness of existing policies regarding HwHs. (OSCAR data showed that there were 105 LTCHs in 1993 of which 10 were HwHs. By October 2005, there were 373 LTCHs of the majority which were HwHs.) We considered whether our regulations sufficiently protected the Medicare program from the problems that we envisioned in the FY 1995 IPPS final rule. We also questioned the effectiveness of the “performance of basic hospital functions” aspect of the “separateness Start Printed Page 26920and control” requirements alone because we were aware that some co-located providers had been establishing complex arrangements among corporate affiliates, and had obtained services from those affiliates, masking true corporate identities, and therein, diluting or impairing the effectiveness of the separateness criteria in determining whether both hospitals were interrelated. While technically remaining within the parameters of the rule, these arrangements intermingled corporate interests so that the corporate distinctness was lost, thus side-stepping the intent of our regulations. (Although we have had similar concerns regarding patient movement between host hospitals and their satellites, there had never been any “performance of basic hospital functions” criteria established in § 412.22(h) because satellites are part of another hospital, and therefore, share a Medicare provider number with “the hospital of which they are a part” thus making it administratively burdensome to distinguish between the inpatient operating costs of the main hospital and its satellite(s).)

In the FY 2005 IPPS final rule, following serious consideration of the public comments that we received on our proposed policy revisions for LTCH HwHs and satellites (69 FR 28323 through 28327) and further evaluation of the issues, regulatory changes were finalized for HwH separateness and control policies at § 412.22(e) and a new payment adjustment was established for LTCH HwHs and satellites of LTCHs, at § 412.534. (We wish to note that the term “satellite facility” in this section refers to satellites of excluded hospitals, in particular, LTCHs, and does not include satellites of excluded units at § 412.25.)

Specifically, in the FY 2005 IPPS final rule (69 FR 49091 through 49214), effective for cost reporting periods beginning on or after October 1, 2004, for LTCHs we eliminated the performance of basic hospital functions test under § 412.22(e)(5)(i), the 15 percent test under existing § 412.22(e)(5)(ii), and the 75 percent of admissions from other than the host criteria at § 412.22(e)(5)(iii). A LTCH that met administrative separateness and control requirements at § 412.22(e)(1)(i) through (e)(1)(iv), under our finalized policy, satisfied the LTCH HwH requirements. (As noted above in this section, the performance of basic hospital functions test does not exist for satellites; therefore, we did not similarly revise § 412.22(h).) However, we established a new payment adjustment at § 412.534 based upon annual threshold criteria for LTCH HwHs or LTCH satellites of 25 percent (or an applicable percentage) for LTCH discharges who were admitted from their host hospitals.

Section 412.534, Special payment provisions for long-term care hospitals within hospitals and satellites of long-term care hospitals, provides that if a LTCH HwH or LTCH satellite's discharges that were admitted from its host hospital exceed 25 percent (or the applicable percentage) of its total Medicare discharges for the LTCH HwH or LTCH satellite's cost reporting period, an adjusted payment would be made at the lesser of the otherwise payable amount under the LTCH PPS or the amount payable under the LTCH PPS that would be equivalent to what Medicare would otherwise pay under the IPPS. In determining whether a hospital met the 25 percent (or applicable percentage) criterion, patients transferred from the host hospital that had already qualified for outlier payments at the host would not count as a discharge that had been admitted from the host. (We commonly refer to this throughout the preamble and regulations text as the discharge not being counted towards the applicable threshold.)

It is important to note that if the hospital exceeds its threshold, LTCH discharges admitted from the host before the LTCH exceeds the 25 percent threshold would be paid an otherwise unadjusted payment under the LTCH PPS.

We also finalized additional adjustments to the 25 percent policy for specific circumstances. For an LTCH HwH or LTCH satellite located in a rural area, there is no payment adjustment applied under § 412.534 if no more than 50 percent, rather than 25 percent, of the Medicare patients discharged from the LTCH or satellite were admitted from the host. In addition, in determining the percentage of patients admitted from the host, any patients that had been Medicare outliers at the host and then discharged to the rural LTCH HwH or LTCH satellite would be considered as if they were admitted to the LTCH or satellite from a non-host hospital. In addition, in the case of a LTCH or LTCH satellite facility that was co-located with the only other hospital in the MSA or with an MSA-dominant hospital, as defined at § 412.534(e)(4), a payment threshold was established that we believed responded to “the unique needs of these communities” (69 FR 49207). Under § 412.534(e)(2), we do not adjust payments to those LTCH HwHs or LTCH satellite facilities as long as the percentage of Medicare patients discharged from the LTCH HwH or LTCH satellite that were admitted from the urban single or MSA dominant host hospital, did not exceed the percentage of the total Medicare discharges in the MSA in which the hospital is located that were discharged from the host hospital, for the cost reporting period for which the adjustment would be made, but in no case is the percentage less than 25 percent or more than 50 percent. In addition, in determining the percentage of patients admitted to the LTCH from the urban single or MSA dominant host hospital, any patients that had been Medicare outliers at the host and then transferred to the LTCH HwH or LTCH satellite would be considered as if they were admitted to the LTCH from a non-host hospital. (When we refer to “the 25 percent (or applicable percentage)” patient threshold throughout this final rule, the “applicable percentage” refers to these special adjustments that we have provided for the special circumstances of rural, urban-single, or MSA-dominant LTCHs or to the percentage associated with the transition policy, discussed below in this section.)

When implementing this policy, we also provided for a 4-year transition for existing LTCH HwHs or LTCH satellites that met the applicable criteria outlined in the regulations to allow these LTCHs a reasonable period during which hosts and co-located LTCH HwH or LTCH satellites and specific “LTCHs under formation” would be able to adapt to the requirements of the new policy. For cost reporting periods beginning on or after October 1, 2004, through September 30, 2005, these transitioned hospitals were to be grandfathered, with the first year as a “hold harmless” year. However, even for facilities that were being phased-in to the full payment adjustment, in the first cost reporting period, the hold harmless year, the percentage of discharges admitted from the host hospital to the LTCH could not exceed the percentage of discharges admitted from the host hospital to the LTCH HwH or LTCH satellite in its FY 2004 cost reporting period. (For the purposes of § 412.534, the hospital's cost reporting period during FY 2004, the last cost reporting period prior to the implementation of § 412.534, is the “base period” for purposes of establishing the gradual phase-in of the full payment threshold adjustment (69 FR 49196).)

After the first grandfathered cost reporting period, these LTCH HwHs and LTCH satellite facilities were required to meet a percentage transition over the 3-year period beginning in FY 2006. For cost reporting periods beginning on or after October 1, 2005, but before October Start Printed Page 269211, 2006, the percentage of Medicare discharges that may be admitted from the host with no adjustment may not exceed the lesser of the percentage of their discharges admitted from their host during its FY 2004 cost reporting period or 75 percent. For cost reporting periods beginning on or after October 1, 2006 but before October 1, 2007, the percentage of Medicare discharges that may be admitted from the host with no adjustment may not exceed the lesser of the percentage of its Medicare discharges admitted from its host during its FY 2004 cost reporting period or 50 percent, and finally, 25 percent (or other applicable percentage) beginning with cost reporting periods beginning on or after October 1, 2007. Additionally, the 25 percent policy for co-located LTCHs is currently implemented in a location-specific manner. That is, the computation of the percentage of LTCH HwH or LTCH satellite discharges admitted from a host is based solely on the admissions from the physically co-located host and not from other campuses or remote locations which may share a common Medicare provider number with the host.

Although the payment adjustment at § 412.534 focused on LTCH HwHs and satellites of LTCHs and its host hospitals, the relationship between a receiving provider and any referring hospital has been an issue of concern for the Medicare program, even in the absence of co-location. Under section 1886(d)(5)(J) of the Act, added by section 4407 of the BBA of 1997, the Congress provided for a post-acute transfer policy which addressed certain patient discharges from acute care hospitals that subsequently received additional treatment delivered by a second Medicare provider. We believe that the Congress enacted this legislation to discourage acute care hospitals from prematurely discharging patients to another treatment setting in order to increase Medicare payment.

The Congress' enactment of the legislation authorizing the post-acute transfer policy is indicative of its serious concerns about patient shifting between acute and post-acute providers. In the case of the post-acute transfer policy, described above in this section, we focused on overpayment, under the IPPS, to the transferring hospital when a patient is prematurely discharged to another provider during the same episode of illness.

The payment adjustment for co-located LTCHs at § 412.534 was based on concerns similar to those underlying the post-acute transfer policy at § 412.4, that is, an inappropriately truncated hospitalization at a host facility and an admission to another provider, specifically a LTCH, for which an additional Medicare payment would be generated. However, the payment adjustment at § 412.534 is not applied to the transferring hospital but rather, to discharges from the co-located LTCH to which the presumably prematurely discharged patient has been admitted. Moreover, although the referring hospital under the post-acute transfer policy must be an acute care hospital, for the purposes of the payment adjustment at § 412.534, any hospital is a potential host if it is co-located with a LTCH HwH or LTCH satellite.

When we proposed the 25 percent (or applicable percentage) payment adjustment for co-located LTCHs in the FY 2005 IPPS proposed rule, MedPAC expressed concern that the 25 percent patient threshold policy would have a significant impact and could possibly lead to an inequitable situation for co-located LTCHs, as compared to freestanding LTCHs. Among their concerns were the following: Freestanding LTCHs also have strong relationships with acute care hospitals, and that where on average LTCH HwHs receive 61 percent of their patients from their hosts, on average freestanding LTCHs receive 42 percent of their patients from their primary referring hospital; a 25 percent rule that only applied to LTCH HwHs and not to freestanding LTCHs could be inequitable; and if this policy approach applied the adjustment only to HwHs and satellites it could be circumvented by an increase in the number of freestanding LTCHs instead of LTCH HwHs (69 FR 49211).

In the RY 2007 LTCH PPS final rule, we also stated that according to a commenter, the data indicated “* * * that it is common practice for LTCHs * * * to admit patients from a single-source acute care hospitals” and that 71.2 percent of freestanding LTCHs admit more than 25 percent of their patients from a single source acute-care hospital (71 FR 27878).

Additionally, in comments received on the FY 2005 IPPS proposed rule to preclude common ownership of a host and a HwH (which was not finalized), two commenters asserted that the financial incentive to accept inappropriate patients from an acute care hospital could exist only when the acute care hospital and the LTCH were commonly owned and when there was common governance, a situation that “can exist even without co-location, that is, a freestanding LTCH, exempt from the requirements of § 412.22(e) could be owned and governed by the hospital from which it receives the majority of its referrals' (69 FR 49202). Despite the commenters' assertions, we do not believe that either common ownership or co-location are the only circumstances under which financial incentives exist for acute care hospitals to prematurely discharge Medicare patients to LTCHs for additional treatment during the same episode of patient care. In fact, we are aware of the existence of “arrangements” between Medicare acute and post-acute hospital-level providers that may not have any ties of ownership or governance relating to patient shifting that appear to be based on mutual financial gain rather than on significant medical benefits for the patient. This could be the case if an acute care hospital discharges a Medicare beneficiary who continues to require hospital-level care primarily to preclude that patient's case from reaching outlier status at the acute care hospital, to an LTCH for additional treatment. Under this scenario, Medicare would pay the acute care hospital under the IPPS for the beneficiary's care but the hospital would be able to avoid both losing the “fixed loss” amount and absorbing 20 percent of the remaining costs for the outlier patient's care, as established under the IPPS outlier policy at subpart F of part 412. Medicare would also be responsible for a payment, to the LTCH, under the LTCH PPS upon the patient's discharge from the LTCH. Accordingly, we believe that additional regulation in this area is both necessary and appropriate to protect the Medicare Trust Fund when generating two payments under two different payment systems for what was essentially one episode of beneficiary care.

When we finalized the payment adjustment at § 412.534, which focused solely on co-located LTCHs, that is, LTCH HwHs and satellites of LTCHs, and as we subsequently noted in the RY 2007 LTCH PPS final rule, we took considerable note of these comments and we have continued since that time to monitor the relationships between referring hospitals and LTCHs (71 FR 27878). Specifically, at that time we also analyzed patient claims data from the FY 2004 MedPAR files for acute care patients who are admitted to freestanding LTCHs. We have analyzed the discharge and LOS information from this data to evaluate whether there was a significant difference in patient shifting behavior between co-located LTCHs and their host acute care hospitals and those freestanding LTCHs that admit a majority of their patients from particular referring acute care hospitals. (As stated previously, for the Start Printed Page 26922purposes of the payment adjustment at existing § 412.534, any inpatient hospital-level provider is a potential host if it is co-located with a LTCH HwH or LTCH satellite (69 FR 49198). Similarly, freestanding LTCHs also admit patients from sources other than acute care hospitals. However, our data reveals that approximately 80 percent of all LTCH admissions are from acute care hospitals. Therefore, our data analysis discussed below in this section, focuses on the relationship between a referring acute care hospitals and LTCHs.)

We also analyzed more recent data on relationships between LTCHs and acute care hospitals from which they received a significant percentage of referrals. The RY 2005 MedPAR files indicate that only 73 of the then 200 freestanding LTCHs admitted 25 percent or less of their Medicare discharges from an individual acute care hospital; for 82 of those freestanding LTCHs, the percentage was between 25 and 50 percent; for 33 it was between 50 and 75 percent, and for 6 percent of those freestanding LTCHs it was between 75 and 100 percent of their Medicare discharges that were admitted from one acute care hospital. Thus, the data indicates that for over 60 percent of all freestanding LTCHs, over 25 percent of their discharges were for patients admitted from an individual acute care hospital.

Generally, the data reveals minimal differences for cases grouped to the same DRG between the ALOS at the acute care hospital prior to an admission to a co-located LTCH and the ALOS at a referring acute hospital prior to admission to a freestanding LTCH. For example, when we finalized the 25 percent threshold payment adjustment for co-located LTCHs at § 412.534, we evaluated data from CY 2004 MedPAR files regarding LTC-DRG 475, Respiratory System Diagnosis with Ventilator Support, for both LTCH HwHs with more than 25 percent of their discharges admitted from their host hospital and freestanding LTCHs with more than 25 percent of their discharges admitted from an individual referring hospital. The ALOS for patients stays that have not reached outlier status at the host prior to being discharged to the co-located LTCH was 12.7 days and for freestanding LTCHs, the average LOS at their individual referring hospital was 12.9 days. Similarly, for LTC-DRG 416, Septicemia, the ALOS at the host acute care hospital was 9.8 days prior to admission to the co-located LTCH and the prior ALOS at the individual referring acute care hospital was 9.6 days prior to admission to the freestanding LTCH. Even though we finalized the percentage threshold payment adjustment only for co-located LTCH HwHs and satellites at that time, we believed that this data indicates considerable similarity between the patient-shifting behavior at acute care hospitals with co-located LTCHs and acute care hospitals with LTCHs with which they are not co-located. We would have expected the LOS at the acute care hospital that discharged patients to non-co-located LTCHs to be longer.

Furthermore, as noted above in this section, we have concentrated on the relationships between acute care hospitals and non-co-located LTCHs in this discussion, because approximately 80 percent of Medicare patients in LTCHs are admitted from acute care hospitals. However, we believe that the same concerns, articulated above, would also exist when the patient source is not an acute care hospital. There could still be a financial incentive on the part of the referring hospital (for example, an IRF, to prematurely discharge a beneficiary to a LTCH for additional post-acute treatment in order to avoid absorbing high treatment costs under the IRF outlier policy at § 412.624(e)(5)) that would result in two Medicare payments, one to the initial provider and the other under the LTCH PPS for, what is actually, a single episode of beneficiary care. (We recognize that a patient could experience a medical crisis while an inpatient at an IRF, but typically, the most appropriate setting for such urgent care would be a general acute care hospital, rather than a LTCH.)

We believe that this data gives further credence to concerns articulated by MedPAC and the assertions made by the Lewin Group in their comments on our FY 2005 IPPS proposed rule regarding the “strong relationships” for referral purposes that exist between many acute care hospitals and freestanding LTCHs. Although, our decade-old concerns, about LTCHs functioning as long-stay or step-down “units” of acute care hospitals, focused on co-located LTCHs (HwHs and LTCH satellites), we believe that this data indicates that many freestanding LTCHs may also be serving the same purpose as those that are co-located, that is, as functional step-down units of their primary referring acute care hospital.

We are also concerned about other attempts to evade our regulations at § 412.534. In implementing the HwH regulations at § 412.22(e) and the satellite regulations at § 412.22(h), we have consistently utilized the definition of “campus” that was established in the provider-based regulations at § 413.65(a)(2) which specifies that a campus is “the physical area immediately adjacent to the provider's main buildings, other areas and structures that are not strictly contiguous to the main buildings but are located within 250 yards of the main buildings, and any other areas determined on an individual basis, by the CMS regional office, to be part of the provider's campus.” We have become aware of certain LTCH companies that have both established new LTCHs and are considering relocating existing HwHs or LTCH satellites so that they are at least 300 yards from the acute care hospital, thus side-stepping the intent of existing § 412.534. We believe that extending the existing payment policy will also address the type of “gaming,” described above in this section.

We first noted in the RY 2006 LTCH PPS final rule (71 FR 27878) our concern that in many cases that the line of “functional separateness” between freestanding LTCHs and their major referral sources appears to have been erased. We believe that our analysis of patient movement between these facilities supports these concerns.

Therefore, under the broad authority conferred on the Secretary by section 123 of the BBRA, as amended by section 307(b) of the BIPA to implement a prospective payment system for LTCHs, including authority to provide for appropriate adjustments to the payment system, we proposed the extension of the payment adjustment at § 412.534, presently applicable to co-located subclause (I) LTCHs, to all subclause (I) LTCHs (section 1886(d)(1)(B)(iv)(I) of the Act), as explained below in this section. (For the purposes of the discussion of this policy, a “subclause (I) LTCH” is also intended to include satellites of these LTCHs. Our proposal regarding subclause (II) LTCHs, that is those LTCHs that meet the definition at section 1886(d)(1)(B)(iv)(II) of the Act, is discussed below in this section.) Specifically, at § 412.536, we proposed regulations that govern payments under the LTCH PPS for LTCH and LTCH satellite Medicare discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH.

The proposed policy provisions of the 25 percent (or applicable percentage) payment adjustment apply to any subclause (I) LTCH or LTCH satellite regardless of the physical proximity to the hospital from which it is accepting admissions. In order to apply this policy at all subclause (I) LTCHs and LTCH satellites, we proposed to additionally revise existing § 412.534 to include a new provision at § 412.534(h) that Start Printed Page 26923would extend the 25 percent (or applicable percentage) payment threshold to those grandfathered co-located subclause (I) LTCH HwHs and LTCH satellites at § 412.22(f) and § 412.22(h)(3)(i), respectively, for Medicare discharges that had been admitted from the grandfathered LTCH or LTCH satellite facility's host for cost reporting periods beginning on or after July 1, 2007. (We address the issue of satellites of subclause (II) LTCHs below in this section.) We proposed adding § 412.536 that applies a comparable payment adjustment governing Medicare discharges from subclause (I) LTCHs and LTCH satellites that were admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH.

The proposed payment adjustment at § 412.536 applies to those Medicare discharges from co-located subclause (I) LTCHs (HwHs and LTCH satellite facilities) that have been admitted from hospitals other than those with which they are co-located. We believe that this policy addresses our concerns with LTCHs and LTCH satellites that in many cases appear to be functioning like step-down units of acute care hospitals.

Furthermore, we believe it is appropriate that the same analytical standards and payment policies be applied by Medicare to all subclause (I) LTCHs. Therefore, we proposed amending existing § 412.534 to include subclause (I) grandfathered LTCH HwHs and LTCH satellite facilities, as well as using the same thresholds applicable to co-located LTCH HwHs and LTCH satellite facilities for subclause (I) LTCHs and LTCH satellite facilities that admit Medicare patients from referring hospitals not co-located with the LTCH or the satellite of a LTCH, under § 412.536.

Specifically under the proposed policy, for cost reporting periods beginning on or after July 1, 2007, as we specified in revised § 412.534(h), this proposed payment adjustment would have included those subclause (I) LTCH HwHs and satellites that had been “grandfathered” under § 412.22(f) and § 412.22(h)(3)(i), respectively, and that are presently exempted from the existing payment adjustment for co-located LTCHs. As noted previously, both grandfathered HwHs at § 412.22(f) and satellite facilities at § 412.22(h)(3)(i) would be permitted to retain their exclusions from the IPPS despite not meeting “separateness and control” policies with regard to their relationships with their host hospitals, as long as they continued to comply with applicable Medicare requirements. This inclusion of grandfathered LTCH HwHs and LTCH satellites in the proposed 25 percent (or applicable percentage) threshold policy would not effect their ability to continue to be “grandfathered” and excluded from the IPPS. Moreover, as noted above, the 25 percent (or the applicable percentage) threshold policy governing discharges from subclause (I) LTCHs that had been admitted from any individual referring hospital not co-located with the LTCH or the satellite of a LTCH, at § 412.536, would also apply in determining payments under the LTCH PPS for Medicare discharges from LTCH HwHs and LTCH satellites, including grandfathered HwHs and LTCH satellites, that had been admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH (that is, referring hospitals other than their hosts).

Under the policies applicable to grandfathered subclause (I) LTCH HwHs and LTCH satellites, we proposed to pay an adjusted amount for those discharged Medicare patients that were admitted from their co-located host, under § 412.534(h) or from any other referring hospital under § 412.536, in excess of the applicable percentage threshold. The grandfathered LTCHs and LTCH satellite facility's Medicare discharges that reached outlier status at the host, at § 412.534(h), or at the referring hospital not co-located with the LTCH or the satellite of a LTCH, at § 412.536, would not count towards the applicable threshold.

We believed that since we proposed expanding the 25 percent policy to all subclause (I) LTCHs and LTCH satellite facilities it was appropriate to include LTCH HwHs and LTCH satellites grandfathered respectively under § 412.22(f) and § 412.22(h)(3)(i). We proposed that the provisions at § 412.534(h) would apply for Medicare discharges from grandfathered LTCH and LTCH satellite facilities admitted from co-located hospitals and the provisions at § 412.536 would apply for discharges admitted from the referring hospital not co-located with the LTCH or the satellite of a LTCH. As we noted in our RY 2007 LTCH PPS final rule regarding grandfathered HwHs, “[W]e do not believe that it is reasonable to assume that by creating a limited exception for these hospitals, the Congress was immunizing these facilities from any further regulation by the Secretary as to their growth and financial impact on the Medicare program. We do not believe the Congress was establishing a separate class of providers” (71 FR 48109).

As noted in the proposed rule, when we implemented the existing 25 percent (or applicable percentage) for cost reporting periods beginning on or after October 1, 2004, we opted to implement on a “location-specific” basis rather than based on Medicare provider numbers. That is, we applied the percentage threshold payment adjustment only to discharges from a specific location of a LTCH HwH or LTCH satellite that was admitted from the host hospital with which they share a building or campus. However, since implementing this policy, we have been contacted by numerous representatives of LTCH chains whose questions appear to indicate that the site-specific implementation of the threshold percentage had resulted in patient-shifting between hospital locations that shared a Medicare provider number and even between separately owned LTCHs (for their mutual advantage) that side-stepped the intent of our policy. Specifically, we offer the following example of a situation that was occurring: a host hospital at Location A was discharging patients to a LTCH HwH or satellite at Location B while the host hospital at Location B discharged patients to the LTCH HwH or satellite at Location A.

We also proposed that for those co-located LTCHs already subject to the 25 percent (or applicable percentage) payment adjustment at existing § 412.534, the policy expansion at § 412.536 would apply to payments under the LTCH PPS for patients discharged from co-located LTCHs (HwHs and satellites) that were admitted from referral sources other than their host hospital(s).

Therefore, under the proposed policy, for cost reporting periods beginning on or after July 1, 2007, a subclause (I) LTCH or LTCH satellite that discharges more than 25 percent (or applicable percentage) of Medicare patients admitted from any individual referring hospital not co-located with the LTCH or the satellite of a LTCH. (that had not already reached outlier status, as discussed above) would be subject to the payment adjustment at § 412.536 for Medicare discharges from that hospital in excess of the applicable threshold. Furthermore, we believe that with the application of our proposed policy at § 412.536 to Medicare discharges from subclause (I) LTCH HwHs and LTCH satellites that were admitted from any individual referring hospital not co-located with the LTCH or the satellite of a LTCH., we are closing the “location-specific loophole” established by the implementation of § 412.534. The change would affect all LTCHs or LTCH satellite Medicare discharges that were Start Printed Page 26924admitted from hospitals that are located on a different campus.

We proposed that the payment adjustment at § 412.534(h) for grandfathered LTCH HwHs and LTCH satellite facilities, discussed above in this section, would track the applicable provisions of the existing payment adjustment at § 412.534. Therefore, we proposed, at § 412.534(h), for cost reporting periods beginning on or after July 1, 2007, the provisions of § 412.534 will also apply to grandfathered subclause (I) LTCH HwHs and LTCH satellite facilities. Accordingly, under revised § 412.534, if the percentage of the grandfathered LTCH or LTCH satellite's discharged Medicare inpatient population that were admitted from its co-located host exceeds the applicable percentage of the LTCH's Medicare discharges for that cost reporting period, an adjusted payment will be made for those discharges that were admitted from that hospital beyond the applicable percent threshold, at the lesser of the otherwise payable amount under 42 CFR part 412, subpart O or the amount payable under subpart O that would be equivalent to what Medicare would otherwise pay under the rules at subpart A, § 412.1(a). (The specifics of this payment formula are explained in considerable detail in the RY 2007 LTCH PPS final rule (71 FR 27879).) Furthermore, as with our initial payment adjustment at § 412.534, we proposed additional adjustments for LTCHs and LTCH satellites that would be affected by the new regulations and that are located in rural areas, or that admit Medicare patients from urban single or MSA-dominant referring hospitals (discussed below).

We did not propose extending the payment adjustment in § 412.534(h) and § 412.536 to those LTCHs and LTCH satellite facilities that we refer to as subclause (II) LTCHs and LTCH satellites, established by section 1886(d)(1)(B)(iv)(II) of the Act. The policy for subclause (I) LTCHs and LTCH satellites would be based on a calculation of the percentage of Medicare discharges that a LTCH admits from an individual hospital during a cost reporting period as compared to the LTCH's total Medicare discharges during that cost reporting period. Because of a significant policy distinction that we made at the start of the LTCH PPS for FY 2003, at this time we do not believe that this policy should be applied to subclause (II) LTCHs and LTCH satellite facilities. With the implementation of the LTCH PPS, we revised the § 412.23(e)(2)(i) and (e)(3)(i) to calculate the ALOS based solely on Medicare patients who required long-stay hospitalizations at subclause (I) LTCHs defined by section 1886(d)(1)(B)(iv)(I) of the Act; however, we did not change the formula for calculating the ALOS for a LTCH governed by section 1886(d)(1)(B)(iv)(II) of the Act, implemented at § 412.23(e)(2)(ii), for a “subclause (II)” LTCH. We believed that in establishing a “subclause (II)” LTCH, the Congress provided an exception to the general definition of LTCHs under subclause (I). We had no reason to believe that the change in methodology for determining the average inpatient LOS would better identify the hospitals that the Congress intended to exclude under subclause (II) (67 FR 55974). Similarly, when we established the existing 25 percent or applicable percentage payment adjustment at § 412.534, we determined that its application to subclause (II) LTCHs was inappropriate because the designation of a subclause (II) LTCH was not solely dependent upon Medicare discharges (69 FR 49205). Therefore, we are not applying the expansion of the 25 percent policy at § 412.536 and amended § 412.534 to LTCHs and LTCH satellite facilities defined under section 1886(d)(1)(B)(iv)(II) of the Act. The existing and amended payment threshold adjustments at § 412.534 and at § 412.536 for subclause (I) LTCHs and LTCH satellites are based solely on percentages of LTCH Medicare discharges. As stated above in this section, we continue to believe that since we include both Medicare and non-Medicare discharges in our calculations for defining a subclause (II) LTCH at § 412.23(e)(2)(ii) that applying a payment adjustment that is based solely on Medicare discharges may not be appropriate. Furthermore, consistent with our policy not to include satellites of subclause (II) LTCHs which were specifically grandfathered at § 412.22(h)(3)(ii) in § 412.536, we have excluded subclause (II) LTCH satellites in the application of the 25 percent payment adjustment for co-located grandfathered LTCHs at § 412.534(h).

We received 270 comments on the RY 2008 LTCH PPS proposed rule. Several of these comments pertained to the extension of the expansion of the 25 percent rule to certain situations not currently covered under existing § 412.534. The following is a summary of these comments and our responses.

Comment: One commenter expressed concern about the President's budget that has submitted to the Congress the savings to be affected by this proposed rule are already “scored” and claimed as savings. In light of this, the commenter questioned the legitimacy of the comment process.

Response: We disagree with the commenter that the inclusion of anticipated savings from the LTCH PPS in the President's Budget invalidates the legitimacy of notice and comment rulemaking. Projections for expenditures and savings are a necessary and expected step in the budgetary process for the Federal Government. The budget only represents the President's expectations or projections of what may happen in the future. It may make assumptions as to policies that have been proposed (or are being evaluated for this purpose) as a representation of will happen. But at most, the Budget should not be viewed as a final blueprint because the Administration cannot anticipate policy modifications in response to public comments. We fully consider all comments received during the comment period and modify proposed policies in response to public comment. Furthermore, we would urge the commenter to review the last several years of LTCH PPS and IPPS proposed and final rules and focus on the differences between the policies that we proposed and those that we finalized (for example, the interrupted stay policy (67 FR 13416, 13455 through 13462, and 67 FR 55954, 56003 through 56006); qualifications for LTCH HwH status (69 FR 23306, 28323 through 28327, and 69 FR 48916, 49191 through 49214); and revisions in the grandfathering of HwHs and satellites (71 FR 23996, 24124 through 24126 and 71 FR 47870, 48106 through 48117)) in order to more clearly appreciate the impact that comments have on the development of our final policies.

Comment: Several commenters questioned our authority in proposing a payment adjustment for LTCHs that is based on an IPPS payment. These commenters assert that the Congress excluded LTCHs from the IPPS in 1983 and enacted legislation that mandated a separate PPS for LTCHs that specifically required that payments to LTCHs should reflect the resource use and costs of treating LTCH patients. The commenters believe we are violating the statutory requirement that payments to LTCHs be on a per discharge basis “that reflects the reasonable and necessary cost of providing services in a hospital having an average LOS of greater than 25 days.” The commenters assert that a payment “equivalent to” or “comparable to” payments under the IPPS are actually payments under the IPPS, violating Congressional intent. Several commenters acknowledge our belief that the IPPS-equivalent is not a Start Printed Page 26925payment under the IPPS but the “thrust of the rationale” for imposing the rule is that these cases still belong in the acute care hospital and payment should mirror payment under the IPPS. One commenter stated that the Congress “established LTCHs as a distinct and separate level of care.”

Several commenters believe we are violating section 1801 of the Act (“Nothing in this title shall be construed to authorize any Federal Officer or employee to exercise supervision or control over the practice of medicine or the manner in which medical services are provided”) and section 1802(a) of the Act (“Any individual entitled to insurance benefits under [Medicare] * * * may obtain health services from any institution, agency, or person qualified to participate * * * [in the Medicare program] if such institution, agency, or person undertakes to provide him such services”). These commenters stated that we have no authority to pay for services provided at a LTCH under the IPPS. Statutory authority for the establishment of the LTCH PPS indicates the Congress believed that LTCH care is more costly than acute because it requires the Secretary “to account for different resource use of LTCH patients.” The commenters believe that the policies in the RY 2008 LTCH PPS proposed rule would strip away the special status given by the Congress to LTCHs, thus undermining the purpose of the LTCH PPS because a significant portion of payments would be reimbursed under the IPPS.

Response: Following further data and policy analysis, we believe that the policies that we are finalizing in this rule fairly address circumstances that we have become aware of as the LTCH PPS matures. We do not believe that we violated Congressional intent in either the BBRA of 1999 or the BIPA of 2000 in establishing a payment adjustment under the LTCH PPS that addresses our concerns about paying for a substantial number of short stay patients, particularly those with extremely short stays, under a payment system designed to treat long stay patients.

As indicated previously, section 123 of the BBRA, as amended by section 307(b)(1) of the BIPA, confers broad discretionary authority on the Secretary to implement a PPS for LTCHs, including providing for appropriate adjustments to the payment system. This broad authority gives the Secretary great flexibility to fashion a LTCH PPS based on both original policies, as well as concepts borrowed from other payment systems that are adapted, where appropriate, to the LTCH context. In the instant case, our finalized policy utilizes, in large part, principles from the IPPS payment methodology and builds upon those concepts to create a LTCH PPS payment adjustment that results in an appropriate payment under the LTCH PPS for those inpatient stays that we believe could be more appropriately treated in another setting.

We disagree with commenters that our proposed expansion of the 25 percent policy that provides for a payment based on an “IPPS comparable payment amount” is a payment under the IPPS. We want to emphasize that such a payment is not an IPPS payment, but rather, given the fact that these patients are comparable to patients treated in acute care hospitals and that the statute precludes the existence of LTCH units, it is an appropriate payment adjustment under the LTCH PPS that is equivalent to a payment that would be derived from the IPPS payment methodology. Moreover, the authority extended to the Secretary by the BIPA included the discretion to “provide for appropriate adjustments to the long-term hospital payment system.” Our final policy is one such adjustment made within the authority conferred under the statute. From the inception of the LTCH PPS for FY 2003, we have interpreted the above cited statutory provision to authorize the establishment of payment adjustment policies including short stay outliers (§ 412.529), interrupted stays (§ 412.531), and discharges from LTCHs. We also believe that the authority extended to the Secretary by the BIPA includes the discretion to develop a payment adjustment based upon establishing a percentage threshold for LTCH discharges that we believe are comparable to discharges from acute care hospitals under circumstances where we believe that a full episode of care has not been delivered at the referring hospital and that the LTCH is functioning like a step-down unit of the referring hospital.

We believe that further refining the 25 percent policy actually captures Congressional intent since it addresses the situation of a LTCH which by all appearances is serving as a unit of another hospital.

Comment: Some commenters maintain that we have no authority to restrict admissions through payment reductions to LTCHs that have no relationship to the referring acute care hospitals. One commenter stated that in proposing the extension of the 25 percent policy to non-co-located LTCHs, we have violated the Court's two-prong test for validity of a regulation established under Chevron U.S.A., Inc. v. Natural Resources Defense Counsel, Inc., 467 U.S. 837, 842-843 (1984). Under the ruling, the Court asks whether the Congress addressed, in clear language, the issue in question and, if the answer is affirmative, the effect is given to the “unambiguously expressed intent of Congress.” If the “statute is silent or ambiguous with respect to the specific issue,” “the Agency's interpretation is allowed to stand as long as it is based on a permissible construction of the statute.” Id. at 843. Deference to the Agency's interpretation is “only appropriate when the agency has exercised its own judgment” and is not based upon an erroneous view of the law. Id.

Response: We disagree that we have imposed criteria that would restrict admissions through payment reductions to LTCHs that have no relationship to the referring acute care hospitals. The payment adjustment we are implementing is not the equivalent to setting “admissions criteria” for treatment at a LTCH. An LTCH may admit as many hospital-level patients as it can safely treat and from whatever source(s) it chooses. However, we believe that LTCHs that discharge greater than the applicable percentage of patients admitted from a particular source that had not reached high cost outlier status, may be understood to be functioning similarly to a co-located LTCH (HwH or satellite), and therefore, more like a step-down unit of the acute care hospital. Under such a circumstance, we believe that the Medicare program would be generating a second payment under the LTCH PPS for a single episode of care for patient who, had not completed his or her episode of care and, is discharged to a LTCH for the remaining portion of the original episode of care. Thus, we believe that it is appropriate to adjust the payment to be made to the LTCH under the LTCH PPS.

Section 123 of the BBRA, as amended by section 307 (b) of the BIPA, confers upon the Secretary tremendous discretion in creating the LTCH PPS. We believe that the expansion of the 25 percent policy is in accordance with the authority granted to the Secretary under 123 of the BBRA as amended by section 307 of the BIPA to make adjustments under the LTCH PPS and is consistent with the statute which precludes the establishment of LTCH units at section 1886(d)(1)(B) of the Act and is also consistent with the Secretary's authority under sections 1102 and 1871 of the Act. Therefore, we disagree with commenters that the Secretary is acting in contradiction of the statute and inconsistently with the Chevron doctrine. Start Printed Page 26926

As a result of our monitoring efforts, we have become increasingly aware that the intent of our existing payment adjustment policy at § 412.534 aimed at combating LTCHs functioning as long-stay “units” of the referring hospitals is being circumvented by creative patient-shifting and admission practices, in addition to, a spiked increase in the number of freestanding LTCHs. We have been monitoring the patient shifting patterns of LTCHs and referring hospitals that are not co-located with one another and have detected behavior that is not significantly different from that of co-located LTCHs and their host hospitals. Therefore, we do not believe that co-location is a prerequisite to inappropriate patient-shifting between an acute care hospital and a LTCH.

We believe that the danger of LTCHs functioning as “units” appears to be occurring not only in LTCH HwHs and LTCH satellites, but also with freestanding LTCHs, and that in many cases, these non-co-located LTCHs and their referral sources may be functioning in ways that appear to have erased the line of “functional separateness” between these LTCHs and their referring acute care hospitals. If patient-shifting between the referring hospital and a LTCH exceeds a specific threshold prior to the patient reaching outlier status at the referring hospital (that is, prior to receiving a full episode of care) the LTCH appears to be functioning as a de facto step down unit of the acute care hospital, a configuration not permitted by section 1886(d)(1)(B) of the Act, which authorizes rehabilitation and psychiatric units but not LTCH units of acute care hospitals. We believe that if the patient is in effect, being treated in a “unit” of the acute care hospital, it is reasonable to revise the payment methodology and take this into account.

Comment: We received several comments supporting our inclusion of grandfathered LTCH HwHs in the 25 percent threshold payment adjustment. These commenters stated that such inclusion would “level the playing field” among LTCHs. A number of commenters disagreed with applying the 25 percent threshold payment adjustment for co-located LTCH HwHs and satellites. Other commenters urged us to “continue the grandfathering exemption.” Several commenters stated that including grandfathered LTCH HwHs with other LTCHs “evades the Congressional mandate for grandfathering” and also contradicts regulatory statements that we have made since the start of the LTCH PPS. One commenter stated that grandfathered LTCHs HwHs have “operated in reasonable reliance on CMS statements that it [would] not apply the HwH requirements to [grandfathered LTCHs]” and requested that we continue to exempt grandfathered LTCHs from the proposed 25 percent rule. The commenter noted that since grandfathered LTCH HwHs were exempt from the original 25 percent policy that had been codified at § 412.22(e)(5)(iii) and since § 412.534 is based on that requirement, we should continue to exempt grandfathered LTCH HwHs from this policy. One commenter noted that grandfathered LTCH HwHs were protected against being paid under the IPPS even though they did not comply with the “separateness and control” regulations but if they are required to comply with the 25 percent threshold payment adjustment, the “result will be the same” because the grandfathered LTCH HwH would be paid under the IPPS. Another commenter cited that LTCH HwHs are precluded from growing under our regulations, and therefore, they should be exempted from the 25 percent policy. One commenter agreed that HwH, freestanding, and grandfathered LTCHs should be subject to the extension of the 25 percent threshold rule, but believes that the threshold should be 35 percent for this group of LTCHs instead of 25 percent because it would still allow CMS to achieve its stated goal and would also be more realistic for LTCH providers that operate in small urban markets which are very similar to rural areas.

Response: We appreciate those commenters who endorsed our inclusion of grandfathered LTCH HwHs in the 25 percent threshold payment adjustment. (We would also note that satellites of LTCHs at § 412.22(h)(4) will also be affected by the policy change.) The payment adjustment that we are finalizing, will affect all subpart (I) LTCHs, including those LTCHs and LTCH HwHs and satellites that were already regulated under § 412.534 for discharges that had been admitted from their co-located hosts. It addresses our concern regarding Medicare patients who are discharged from referring hospitals prior to the delivery of a full episode of care, to LTCHs. In keeping with our fiduciary responsibility to protect the Medicare program against duplicative and inappropriate payments, we are finalizing the proposed policy at § 412.534(h) under which all subclause (I) LTCHs, including grandfathered LTCH HwHs and satellites, will be subject to the 25 percent (or applicable percentage) threshold payment adjustment with regard to Medicare discharges that they admit from their co-located host. (We are also providing for conforming changes to § 412.534(a), (c)(1), (c)(2), (d)(1), and (e)(1) to include grandfathered HwHs and satellites, in existing provisions.) Furthermore, under new § 412.536, Medicare discharges from grandfathered LTCH HwHs and satellites that were admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH that exceed the applicable threshold, will be subject to the payment adjustment described in detail above in this section. (Elsewhere in these responses, we discuss the 3-year transition period to the full threshold adjustment that we are also providing for all LTCHs and LTCH satellites including grandfathered LTCHs and satellites affected under § 412.536.)

We disagree with commenters who stated that we are “evading Congress’ mandate, and contradicting regulatory statements that we have formerly made.” Section 4417(a) of the BBA of 1997 amended 1886(d)(1)(B) of the Act to provide that “[a] hospital that was classified by the Secretary on or before September 30, 1995 as a hospital described in clause (iv) [a LTCH] shall continue to be so classified notwithstanding that it is located in the same building as or on the same campus as another hospital.” We believe this provision was intended to prevent grandfathered LTCHs that were unable to satisfy our HwH regulations from losing their LTCH status. By finalizing the 25 percent (or applicable percentage) payment threshold policy to include grandfathered LTCHs HwHs, in no way are we countermanding their exemption from the separateness and control regulations at § 412.22(e). LTCHs that exceed the applicable threshold do not lose their LTCH status. Rather, the new policy only affects the payment level for all LTCHs that exceed the threshold. We further believe that including grandfathered LTCH HwHs (and satellites) within the scope of the percentage payment threshold that we have established to ensure that Medicare is not generating two full payments one under the IPPS and another under the LTCH PPS for one episode of care, is well within the authority of section 123 of the BBRA, as amended by section 307(b)(1) of the BIPA, which confers broad discretionary authority on the Secretary to develop and implement a PPS for LTCHs and further provides that the Secretary “may provide for appropriate adjustments to the long-term hospital payment system.”

We do not believe that it is reasonable to assume that by creating a limited Start Printed Page 26927exception for these hospitals that the Congress intended to immunize these facilities from any further regulation by the Secretary as to their growth and financial impact on the Medicare program. “We do not believe Congress was establishing a separate class of providers” (71 FR 48109). Grandfathered LTCHs and LTCH satellite facilities are paid under the LTCH PPS and the revised payment adjustment under § 412.534 and new § 412.536 is merely another feature of the LTCH PPS.

One commenter believes we contradicted our own statements by including a partial quote from the FY 2007 IPPS final rule about grandfathered LTCH HwHs' “reasonable reliance” on the fact that we would not apply the HwH requirements. In that final rule, we explained that “[t]he purposes of our grandfathering certain existing HwHs and satellites was to reflect reliance interests and settled expectations that existed on the part of these facilities at the time the separateness and control requirements were created” (71 FR 48107). We believe this statement is consistent with our belief that including grandfathered HwHs in the extension of the 25 percent (or applicable percentage) payment threshold policy does not violate the Congress' intent. The expansion of the 25 percent policy will not affect the “reliance interests and settled expectations” of grandfathered HwHs (and also on LTCH satellites) since they will continue to be exempt from meeting the separateness and control requirements that are required by non-grandfathered co-located LTCHs. Moreover, the concerns that we hold regarding premature patient shifting from host hospitals or referring hospitals to LTCHs and the consequences of such patterns for Medicare payment purpose, may even be more relevant with regards to grandfathered LTCH HwHs because since they are exempted from the separateness and control policies they may even more closely resemble step-down units of their host hospitals.

Several commenters noted that the 25 percent threshold payment adjustment originated as one of the three options (the 75/25 test) with which HwHs could comply to meet the separateness and control requirements at (then) § 412.22(e)(v)(C). They stated that since grandfathered LTCH HwHs were exempted from this requirement when it was a “certification issue,” or “control requirement,” these facilities should similarly be exempted from the policy when it is a payment adjustment. We note that even though the percentages in these policies are the same, there is a critical difference between them. Because the effect of section 1886(d)(1)(B) is that grandfathered LTCH HwHs may continue to be classified as LTCHs even if they fail to meet with the “separateness and control” requirements that we had established at § 412.22(e), among which was the 75/25 test as one of the three options for indicating independent “performance of basic hospital functions” between the host and the LTCH HwHs, grandfathered HwHs continued to be excluded from the IPPS despite their unquestioned organizational and functional linkage to their host hospitals. A non-grandfathered LTCH HwH that was not in compliance with the separateness and control requirements would have lost its IPPS exclusion. Therefore, since loss of IPPS-excluded status is not a feature of the payment adjustments that we are finalizing at revised § 412.534 and § 412.536, we would disagree with the commenter that the “result will be the same because the grandfathered LTCH HwH would be paid under the IPPS.” Under § 412.534(h), which makes grandfathered LTCH HwHs (and LTCH satellites) subject to revised § 412.534(h) and to § 412.536, for cost reporting periods beginning on or after July 1, 2007, there is no risk of losing IPPS-excluded status. Grandfathered LTCHs would continue to be paid under the LTCH PPS, albeit, an adjusted payment amount, even if they exceed the applicable percentage threshold under our finalized policy.

As with all other subclause (I) LTCHs, Medicare payments to grandfathered LTCH HwHs (and satellites) for discharges in excess of the applicable threshold that were admitted from an individual referring hospital will be based on a payment under the LTCH PPS at the lesser of the otherwise unadjusted amount under the LTCH PPS or a payment equivalent to what would otherwise have been paid under the IPPS. As with all LTCHs and LTCH satellites that are subject to this payment policy, discharges that exceed the applicable threshold that had reached outlier status at the referring (or host) hospital, will not be subject to the payment adjustment and will therefore be eligible for otherwise unadjusted payment under subpart O.

Since we are applying the 25 percent policy even to freestanding LTCHs, it would be inconceivable to treat grandfathered HwHs as being in a unique class that exempts them from the policy while applying the policy to LTCHs that are totally separate from the referring hospital. We believe that the Congress intended to allow grandfathered HwHs to maintain their LTCH status but in no way intended for this group of LTCHs to receive an exclusion from payment policies applicable to freestanding LTCHs.

We further disagree with the commenters that since grandfathered LTCH HwHs (and satellites) are precluded from “growth” under our existing regulations, that they should not be subject to the 25 percent (or applicable percentage) payment adjustment. We have allowed grandfathered LTCH HwHs and satellites to modernize their facilities as necessary and appropriate even if modernization required an increase in square footage. Specifically, in the FY 2007 IPPS final rule, we revisited previous policies that limited grandfathered LTCH HwHs (and satellite facilities, including satellite units) from changing the “terms and conditions” under which they operated at the time of their grandfathering and we revised § 412.22((f)(3) (and § 412.22(h)(4) for satellites), and finalized a policy which would allow them to increase or decrease their square footage or decrease their number of beds without risking their grandfathered status. In that same final rule, we revised this policy for all HwHs, satellites, and satellite units of all excluded hospitals, not only LTCHs, because we were persuaded by comments received on our FY 2007 IPPS proposed rule (71 FR 23996) that these facilities needed to be able to expand in order to modernize (for example, to accommodate new medical equipment, record requirements, and new Federal, State, and local safety requirements). However, we did not allow grandfathered facilities to increase their number of beds because we believed that all grandfathered co-located facilities already held a significant advantage over such facilities that were not grandfathered, because they were not required to comply with separateness and control rules. Therefore, we believed that not only would allowing them to increase their bed count convey an additional unfair advantage to these facilities, but also that such an increase would lead to additional costs for the Medicare program (71 FR 48106 through 48115). We similarly believe that continued exemption of grandfathered LTCH HwHs and satellites from the payment threshold adjustment to which all other subclause (I) LTCHs are subject is both fair and appropriate, and in the words of our commenter, helps to “level the playing field” among LTCHs.

Regarding the commenter's suggestion that even as we extend the 25 percent Start Printed Page 26928threshold payment adjustment to all LTCHs including grandfathered HwHs, we should raise the threshold to 35 percent as a more reasonable goal, particularly for small urban and rural areas, we would call the commenter's attention to the 3-year transition to the full threshold adjustment that we are providing (described in greater detail in the next response) which establishes a 75 percent threshold but not to exceed the percentage in the base year at § 412.536(f)(1) for all impacted LTCHs and LTCH satellites for cost reporting periods beginning on or after July 1, 2007, through June 30, 2008 and a 50 percent but not to exceed the percentage in the base year threshold for all impacted LTCHs and LTCH satellites for cost reporting periods beginning on or after July 1, 2008, through June 30, 2009. For cost reporting periods beginning on or after July 1, 2009, the threshold will be 25 percent (or the applicable percentage.) We have responded to comments regarding single urban and rural LTCHs elsewhere in these responses. We believe that establishing this policy will result in hospitalized patients who continue to need acute care hospital treatment to not be shifted to another acute care hospital setting before the end of a full episode of care, but rather to complete appropriate treatment at the referring hospital.

Comment: Several commenters contend that the relationship between a referring hospital and a freestanding LTCH should not be subject to the same regulatory standards as should a co-located LTCH and its host hospital. Furthermore, the commenters assert that when we finalized the 25 percent payment threshold for co-located hospitals, we provided a 4-year phase-in to the full 25 percent (or applicable percentage) threshold but in our proposed rule, we have not proposed any such phase-in for those LTCHs who would be affected under the proposed policy at proposed § 412.536. The commenters request that if we finalized the proposed extension of the 25 percent payment adjustment to non-co-located LTCHs and LTCH satellites, that we provide a similar transition period to allow LTCHs the opportunity to adapt to the full impact of the policy. In addition, commenters requested that we also provide for implementation on a site-specific basis, as we had under the existing § 412.534 provision rather than based on admissions to the provider in its entirety. One commenter stated that for purposes of implementation, using a provider number definition on the LTCH side would be simpler to track and control and would be less subject to manipulation.

Response: We have expressed our concerns regarding patient-shifting between host hospitals and co-located LTCHs (HwHs and satellites) since we originally established the separateness and control requirements at § 412.22(e) for FY 2005 (59 FR 45389 through 45393). Upon finalizing the 25 percent (or applicable percentage) threshold policy for co-located LTCHs for FY 2005, we received comments indicating that we should be aware of similar patient shifting patterns between non-co-located LTCHs and their primary referring hospitals (69 FR 49211). Specifically, MedPAC noted that “freestanding LTCHs also have strong relationships with acute care hospitals, and that where on average LTCH HwHs receive 61 percent of their patients from their hosts, freestanding LTCHs receive 42 percent from their a primary referring hospital * * * [that] there are some risks in our proposed 25 percent policy; (a) the 25 percent rule that only applies to LTCH HwHs and not to freestanding LTCHs and may therefore be inequitable; (b) it does not ensure that patients go to the most appropriate post-acute setting; (c) this approach may be circumvented by an increase in the number of freestanding LTCHs instead of LTCH HwH.” As we stated in the FY 2005 IPPS final rule, we believe that “MedPAC shares our concern that the LTCH payment system creates an incentive for unbundling of the IPPS in addition to overpayment for the care provided by LTCHs and that this concern is great, particularly, in the case of a LTCH HwH * * * ” (69 FR 49211). We also provided an in-depth discussion of our growing concerns in the RY 2007 LTCH PPS final rule (71 FR 27874 through 27881). As we have stated, when we evaluate patient discharges from a host or a referring hospital (typically, an acute care hospital) and admission to a LTCH, we are particularly concerned that the acute care hospital has not provided a full episode of care for a patient who continues to need hospitalization, but instead, is discharging this patient to another acute care hospital, one that is paid under the LTCH PPS. Consequently, two Medicare claims are submitted; one from the acute care hospital and the other for payment under the LTCH PPS for what was essentially one episode of care.

In this final rule, while we continue to believe that the expansion of the 25 percent payment threshold policy for at § 412.536 and revised § 412.534 are appropriate, in response to the commenters, we have revisited our original proposal and will provide for a 3-year phase-in of the final payment threshold adjustment at § 412.536 and revised § 412.534. Specifically, in this final rule, we have established a 3-year transition period under § 412.536 for LTCHs that will be governed by the expansion of the 25 percent threshold policy for LTCH discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH and also for those grandfathered co-located LTCHs that we included under this policy at revised § 412.534(h).

Under the policy that we are finalizing for cost reporting periods beginning on or after July 1, 2007 and before July 1, 2008, the threshold will be no less than the lesser of 75 percent or the percentage that the LTCH or LTCH satellite discharged from the referring hospital during its RY 2005 cost reporting period. For cost reporting periods on or after July 1, 2008 and before July 1, 2009, the threshold will be no less than the lesser of 50 percent or the percentage that the LTCH or LTCH satellite discharged from the referring hospital, during its RY 2005 cost reporting period. For cost reporting periods beginning on or after July 1, 2009, all LTCHs and LTCH satellites under § 412.536 and grandfathered LTCHs and LTCH satellites under § 412.534 will be subject to the applicable percentage threshold. (We note that for cost reporting periods beginning on or after October 1, 2007, non-grandfathered co-located subclause (I) LTCHs, under § 412.534, are fully phased-in to the full 25 percent (or applicable percentage threshold) for discharges admitted from their co-located hosts. However, payments for LTCH discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH, are governed under § 412.536.)

Furthermore, under our finalized policy, grandfathered LTCH HwHs and satellites, under § 412.534(h) and § 412.536 will now be subject to the 3-year transition that we are finalizing under this new policy for all their discharges, both admitted from their co-located host and from referring hospitals not co-located with the LTCH or the satellite of a LTCH hospital.

We believe that a 3-year transition is sufficient time for those affected LTCHs to adapt to this payment adjustment. Since the implementation of the existing payment adjustment for co-located LTCHs at § 412.534 for FY 2005, we have clearly articulated our continuing concerns about patient-shifting between non-co-located LTCHs and referring hospitals (69 FR 49213, 71 FR 27878 through 27879). Therefore, we believe Start Printed Page 26929that we have provided ample notice to the LTCH industry of potential impending regulation in this area and that therefore we believe that the industry had time to adjust its behavior. We have also seen articles in trade association newsletters over the past several years indicating that the LTCH industry was well aware of our focus on this issue. However, in response to comments, we have adopted a 3-year transition policy that we believe will provide additional time for LTCHs to adjust to the new regulations.

However, we also want to reiterate, that just as we provided under § 412.534, the payment adjustment specified at § 412.536 will not be applied to discharges (admitted to LTCHs or LTCH satellites from referring hospitals not co-located with the LTCH or the satellite of a LTCH) that reached HCO status at the referring hospital prior to admission to the LTCH or LTCH satellite.

Regarding implementation of the new payment adjustments, we will be implementing the percentage threshold at § 412.536 on the provider as a whole for multi-campus referring sources and also for multi-campus LTCHs or LTCH satellites in contrast to our location-specific implementation of the 25 percent payment adjustment for co-located LTCHs under § 412.534. We agree with the commenter that location-specific implementation was consistent with our policy goals in addressing patient movement between co-located LTCHs and LTCH satellites and their hosts. However, we believe that our goals regarding LTCH discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH are more logically served by basing implementation on the provider as a whole (that is, based on discharge data for the entire provider under its provider number). Discharges from a co-located LTCH or LTCH satellite that were admitted from remote locations of the host hospital not co-located with the LTCH or the satellite of a LTCH would also be held to the expanded 25 percent policy by aggregating the discharges from those locations and determining if they exceeded the applicable threshold. Patients that are admitted from the hospital that is co-located with the LTCH or LTCH satellite facility will continue to be governed by the location-specific implementation of § 412.534.

We have revised our proposed policy regarding transitioning to the full 25 percent threshold adjustment and under our finalized policy, for all subclause (I) co-located HwHs and satellites, including grandfathered subclause (I) LTCH HwHs and LTCH satellites under the extension of the 25 percent (or the applicable percentage) threshold policy that we are finalizing, at revised § 412.534(h) and § 412.536, and we are providing for a 3-year transition period. Accordingly, for cost reporting periods beginning on or after July 1, 2007, and before July 1, 2008, the percentage threshold applied would be no less than the lesser of 75 percent of the total number of Medicare discharges that were admitted from all referring hospitals not co-located with the LTCH or the satellite of a LTCH during that cost reporting period or the percentage of Medicare discharges that had been admitted to the LTCH or LTCH satellite from that referring hospital during the long-term care hospital's or satellite's RY 2005 cost reporting period. Although we proposed to use FY 2005 as the base year for this group of LTCHs in the RY 2008 LTCH PPS proposed rule (72 FR 4815), we will use RY 2005 rather than FY 2005 as the base year since we have revised the transition period under § 412.536 to be effective and applicable for cost reporting periods on a rate year cycle (That is, beginning on or after July 1. We originally chose 2005 because when we published our proposed rule, FY 2005 was our most recent full year of MedPAR data. For cost reporting periods beginning on or after July 1, 2008 and before July 1, 2009, the percentage threshold applied would be no less than the lesser of 50 percent of the total number of Medicare discharges that were admitted from all referring hospitals not co-located with the LTCH or the satellite of a LTCH during that cost reporting period or the percentage of Medicare discharges that had been admitted to the LTCH or LTCH satellite from that referring hospital during the long-term care hospital's or satellite's RY 2005 cost reporting period. For cost reporting periods beginning on or after July 1, 2009, the threshold will be 25 percent (or the applicable percentage.) A 3-year transition period is applicable for all subclause (I) LTCHs and LTCH satellites governed under § 412.536 and to grandfathered LTCHs and LTCH satellites now subject to the threshold under § 412.534. For co-located LTCHs (that is, LTCH HwHs and LTCH satellites) it is important to note that under existing § 412.534(g)(4), for cost reporting periods beginning on or after October 1, 2007, LTCH HwHs and LTCH satellites being phased-in to the full adjustment would enter year 4 and be would be required to meet the 25 percent (or applicable percentage) threshold regarding their percentage of discharges from their co-located hosts. However, these LTCH HwHs or LTCH satellites are governed by § 412.536 regarding discharges that they admitted from any other referral source (that is, other than its co-located host hospital) and would be subject to the 3-year transition beginning with cost reporting periods beginning on or after July 1, 2007.

We also believe that it is important that we note that the 3-year transition to the full 25 percent threshold payment adjustment will coincide with our continuing work on the MedPAC recommendations to attempt to develop facility and patient level criteria for LTCHs. We hope that the LTCH industry will work closely with CMS to pursue this endeavor during the transition period.

Comment: Several commenters maintained that we did not present convincing data-based evidence in the RY 2008 LTCH PPS proposed rule and that in the absence of meaningful data no meaningful comments can be made. Several commenters questioned why we are seeking to expand the 25 percent threshold policy to non-co-located LTCHs when we have not yet evaluated data from the FY 2005 implementation of the same payment adjustment for co-located LTCHs and LTCH satellites. Some commenters included data analyses that they believe refutes the policies that we proposed in the RY 2008 LTCH PPS proposed rule. The commenters urged CMS to review the most current hard data from LTCHs and to base all policy formulations on the conclusions that can reasonably be drawn from such data. Several commenters contended that we proposed policy based on anecdotes rather than on hard data and that we have accused the LTCH industry based on this anecdotal evidence. The commenters requested that we provide data, rather than anecdotal evidence of the purported “gaming” that we believe is occurring between the acute hospitals and LTCHs. The commenters further contended that the research produced by RTI should be the foundation of future CMS rulemaking.

Commenters also maintained that rather than continuing to increase, the absolute number of LTCHs has decreased by one during 2006, and therefore, we should not continue to be concerned about industry growth.

Response: We disagree with the commenters' assertions regarding both our analyses and provision of the best available data evidence for the policies that we proposed and that this lack resulted in LTCH stakeholders being unable to submit “meaningful comments.” In fact, we received 270 comments in response to the RY 2008 Start Printed Page 26930LTCH PPS proposed rule (some of which were very lengthy). We believe that the concerns expressed in these comments, which we present in appropriate sections of this final rule by topic, are indicative that meaningful comments were made. In determining our final policy, we are fully aware of the serious attention that our commenters invested in their policy recommendations, as well as in the challenges that they have articulated presented. Moreover, regarding assertions that we have not provided data that indicates our policy rationale, we note that in December 2006 we posted the RTI report in its entirety on the CMS Web site at http://www.cms.hhs.gov/​LongTermCareHospitalPPS/​02a_​RTIReports.asp#TopOfPage. This report contains detailed data analyses which were the bases of RTI's findings and significantly impacted our decisions to propose specific policies.

With regard to the data analyses that some commenters submitted challenging the correlation that we proffered, between the discharges to LTCHs and fewer high cost outlier cases at referring acute care hospitals we would assert that our data analyses (described below) support this theory.

An analysis of our MedPAR data from acute care hospitals regarding their LOS during CY 2003 to their LOS during CY 2005 in markets where LTCHs opened in CY 2004. Our data analysis focused on acute care hospitals that had been the source of at least 25 percent of the LTCH discharges. (Our data indicated that these communities already had some LTCHs at the time when these additional LTCHs opened.) We compared 304,650 acute care cases in CY 2004 to 316,816 cases in CY 2005. In CY 2003, there were 7,586 outliers and in CY 2005, there were 5,858. The percentage of outliers in the acute care hospitals decreased from 2.5 percent to 1.8 percent and the numbers of patients that were admitted to LTCHs in those communities increased from 2,128 in CY 2003 to 6,597 in CY 2005. Furthermore, the percentage of acute care hospital discharges to LTCHs increased from 0.7 percent in CY 2003 to 2.1 percent in CY 2005. The percentage decline in total outliers between the CY 2003 and CY 2005 was −25.7 percent. The increase in LTCH discharges from CY 2003 to CY 2005 was 198.1 percent.

We would also quote section 3.3 “the RTI report which summarizes its detailed data analyses (which are included in the Report) by noting that LTCH admissions were less likely to have had an outlier payment during the prior acute stay (8 percent compared to 12 percent for non-LTCH admissions). The ALOS in the acute hospital [prior to discharge to the LTCH] tended to be longer for the LTCH admissions, averaging 13.5 days compared to only 11 days for the other acute admissions.” (p. 51) This statement indicates that those patients that were admitted to the LTCH before achieving outlier status at the acute care hospital were “sicker” than other patients in those DRGs, which is logical since they continued to need acute hospital-level treatment. (Elsewhere in these responses, we respond, in greater detail, to comments that we received that challenge our benchmark assumption that reaching outlier status signifies the delivery of a full episode of care. To briefly summarize, it is our belief that a patient at an acute care hospital who still is in need of acute hospital-level care upon discharge from that setting, may not have completed the treatment for which the Medicare is paying) and is using the LTCH as a unit to treat those patients.

In particular, we suggest that commenters revisit Table 3-7 in the RTI Report which indicates that while most patients constituting LTCH admissions were previously hospitalized, only a small proportion of those in the acute hospital generated an outlier payment (less than 20 percent) except for the DRG 452: Complications of Treatment with CC (21.3 percent) and DRG 204: Disorders of the Pancreas Except Malignancy (26.2 percent). About one-fourth of the top 50 LTCH conditions had 15 to 20 percent of their admissions qualifying for an acute outlier payment before being admitted to the LTCH. These included many of the medically complex conditions such as: DRG 475: Ventilator Support 16.9 percent); DRG 316: Renal Failure (19.3 percent); DRG 076: Other Respiratory System OR Procedures with CC (19.2 percent); DRG 188: Other Digestive System (19.5 percent); DRG 483: Tracheostomy (17.8 percent); DRG 461: OR Procedures (17.8 percent); DRG 331: Other Kidney and Urinary Tract Diagnoses with CC (17.1 percent); and DRG 440: Wound Debridements for Injuries (19.4 percent). Still, the majority of LTCH admissions were admitted before reaching outlier status in the acute hospital” (p. 48).

We believe that the above data supports our extension of the 25 percent threshold payment adjustment which distinguishes between patients in need of further acute level care who were admitted to a LTCH or satellite after receiving a full episode of care at the referring acute (that is, they reached outlier status at that hospital) and those needing further acute treatment that were admitted to the LTCH following what appears to be a truncated stay at the acute care hospital.

In response to the comments that suggested that our extension of the 25 percent payment threshold policy was premature since as yet, we had no data on the impact of the 25 percent policy on co-located LTCHs, because the policy is not yet fully phased-in, we reiterate that regulating inappropriate patient shifting to LTCH HwHs and satellites from their co-located hosts does not negate the need to address the same issue between LTCHs and referring hospitals with which they are not co-located. We remain concerned about LTCHs with a pattern of patients who need acute hospital-level care after having received treatment for which Medicare has paid under the IPPS that are immediately admitted for additional hospital-level treatment to other acute care hospitals (LTCHs) for another Medicare payment under the LTCH PPS.

In response to commenters who found fault with our attention to anecdotal information regarding the behavior of some LTCHs, we note that determinations are based on our policy on a variety of factors, including information from our FIs, questions and comments from LTCH consultants and attorneys, LTCH advertisements in both print media and the internet that provided us with irrefutable information about LTCH behavior. We believe that it is our fiduciary responsibility to guard the Medicare Trust Fund from inappropriate and unnecessary expenditures. Therefore, we believe that any and all information regarding the LTCH industry is pertinent to our responsibility to be proactive in the regulatory process. For example, we are aware of a growing trend by some LTCHs to establish “units dedicated to mental health,” identified as a “Mental Health Unit” or “Medical-Behavioral Unit.” Assuming that the LTCH organization is cognizant of the preclusion against the establishment of excluded units (for example, psychiatric or rehabilitation) in a hospital that is excluded from the IPPS (see § 412.25((a)(1)(ii)) establishment of such titular “units” would be reimbursed by Medicare under the LTCH PPS. Clearly patients in any acute care hospital setting (and LTCHs are acute care hospitals) may need psychiatric intervention, but given our regulations governing excluded psychiatric units at § 412.27 and the specific COPs for psychiatric facilities at § 482.62, we are very interested in LTCHs that are advertising mental health care as a primary patient service. Start Printed Page 26931

Regarding the comments that note an absolute decrease in the number of LTCHs that were established in FY 2006, we note that we are well aware of continuing growth in the LTCH industry, which in some part, takes the form of large LTCH companies purchasing existing LTCHs and expanding the facilities, as well as the shifting landscape of the LTCH industry brought about by continuing corporate mergers. (Our information in this regard comes to us from FIs, corporate press releases from LTCHs, newsletters from LTCH trade associations, corporate Web sites, and investment newsletters. For example, one Web newsletter announced, “Private Equity Firms Target Long-Term Acute Care Hospitals.” The article continued, “Two operators of long-term acute care hospitals, or LTACS, agreed to be bought by private equity firms, but for very different reasons. Two notable deals were announced this month targeting companies that manage long-term acute care hospitals, or LTACs. In both cases, leveraged buyout firms initiated transactions to buy out operators of multiple LTACs. The rationale for each, however, is different, reflecting different business plans and different stages in the growth cycles of the two companies.”

With respect to the commenter's suggestion that we have alluded to gaming of the Medicare program by the LTCH industry and that we have provided no substantiation for these beliefs, we would note that we have participated in meetings, conference calls, correspondence, evaluated currently-used patient criteria, arranged site visits with LTCHs (and other providers that treat “long-term care hospital-type” patients), and participated in the Technical Expert Panel (TEP) that was held in January 2007. While we have met and worked with highly skilled physicians and administrators of a number of LTCHs and we are aware that many LTCHs provide high quality services to their patients, we are contemporaneously aware of activity by the LTCHs that appear to be directed towards both evading the intent of Medicare policy and also maximizing Medicare payments.

We are also aware that the dynamic of patient shifting from acute care hospitals to LTCHs are well understood throughout the health care industry. In the February 28, 2000 issue of Critical Care Medicine, an abstract of an article entitled, “The impact of long-term acute-care facilities on the outcome and cost of care for patients undergoing prolonged mechanical ventilation” concluded that “Patients undergoing prolonged ventilation have high hospital and 6-month mortality rates, and 6-month outcomes are not significantly different for those transferred to long-term acute care facilities * * *. Acute care hospitals can reduce the amount of uncompensated care by earlier transfer of appropriate patients to a long-term acute care facility.” (Seneff MG, Wagner D, Thompson D, Honeycutt, C, Silver MR, Department of Anesthesiology and Critical Care Medicine, The George Washington University Medical Center).

Lastly, we note that we believe that the policies that we are finalizing in this final rule are built on solid data analysis, reasonable interpretation of information that has come to our attention from the TEPs and the LTCH industry, and our obligation to propose proactive policy initiatives for the long-term benefit of the Medicare program.

Comment: Several commenters offered data indicating that patients admitted to LTCHs following an acute care hospital stay are generally grouped into a different DRG at the LTCH from the one to which they were grouped in the acute care hospital. The commenter used the example of ventilator dependent patients, who typically fall into a tracheostomy DRG (561/562) upon discharge from the acute care hospital but fall under the respiratory failure DRG (475) upon discharge from the LTCH, suggesting that therefore the two episodes of care are distinct and separate. The commenters also claimed that even those patients with the same DRG in each setting do not constitute a single episode of care because of the nature of the institutions and the differences between them. Therefore, the commenters asserted, there can be no actual claim that there is double payment for the same services for LTCH patients coming from IPPS hospitals. In focusing on the appropriate lengths of stay at acute care hospitals preceding a LTCH admission, many commenters quoted the RTI study that notes that, “Understanding whether acute hospitals are already paid for these services or whether LTCHs are providing specialized services not available in the acute hospitals is poorly understood” (p. 55). The commenters believe that a CMS contractor has contradicted statements that we made. Therefore, the commenters state that the extension of the 25 percent threshold payment adjustment to discharges of patients admitted from referring hospital not co-located with the LTCH or the satellite of a LTCH should not be finalized. Several commenters suggested that if we did finalize this payment adjustment, it should be limited only to those situations where the same DRGs were assigned to both the acute care stay and the LTCH stay.

Response: Our data analysis of the 2005 MedPAR files indicates that, generally, when a patient is admitted to a LTCH immediately upon discharge from an acute care hospital, Medicare is paying for treatment under different DRGs for each submitted claim. However, we disagree with the commenters' assertions that there are clear distinctions between “episodes of care” for a patient who is originally treated at an acute care hospital and eventually admitted to a LTCH, whether or not the same DRG is assigned to each stay. Patients being cared for in both the acute care hospital and LTCH settings are very ill, complicated patients with multiple comorbidities, and typically there is not one clear or distinctive principle diagnosis that is the cause of the patient's failure to get well, but rather a constellation of problems that necessitate further treatment. Nor will one “magic” intervention or procedure necessarily cure the patient's problems. DRG assignment is based on software that attempts to group patients according to individual principal diagnoses and surgical procedures, but the clinical reality is that, especially in the case of complex patients with multiple medical problems, DRG assignment can be a limited way of defining or characterizing the nature of a particular episode of care for a given patient.

The example of respiratory failure that the commenter provides is especially illustrative of this point. A patient who suffers from respiratory failure in the acute care hospital, if it does not resolve, will eventually require a tracheostomy, which will then group the patient to the tracheostomy DRG. The tracheostomy itself is a procedure that is usually done on a semi-elective basis when it becomes apparent that the patient will require prolonged mechanical ventilation. If that patient subsequently is admitted to an LTCH, that discharge will necessarily group to the respiratory failure DRG, because the tracheostomy has already been performed during the acute care hospitalization. However, the clinical characteristics of the patient and the type of care that is required, have not materially changed, and the LTCH stay can hardly be viewed as a separate or unique clinical episode from the immediately preceding acute care hospital stay. From a clinical perspective, in the absence of a sharp line of distinction, or a consistent characterization, of exactly which Start Printed Page 26932patient is appropriate for admission to the LTCH, as well as when that patient should be transferred from the acute care hospital setting to the LTCH setting, we have difficulty understanding when, for example, the patient with respiratory failure stops being appropriately cared for in the acute care hospital and paid for under the IPPS and begins to require care in the LTCH. Recognizing that both settings provide acute hospital level care, and also noting that in areas where LTCHs are not available this level of care is provided exclusively in the acute care hospital until the time of discharge to a nonacute setting, it is therefore appropriate to expand the 25 percent policy to all instances in which a referring hospital is discharging so many patients to the LTCH or satellite that it appears to have created a virtual unit of the referring hospital at the LTCH or LTCH satellite.

To those commenters who quoted a sentence (out of context) from the RTI report, we note that a thorough reading of that page indicates that RTI's purpose does not contradict, but rather reinforces the above stated concerns. RTI's full intent may be best understood from the following paragraphs, which includes the quoted sentence:

“Examining the acute length of stay differences was also useful for understanding the relative role of general acute and LTCHs in treating these severely ill populations. The multivariate work showed that LTCH users have a shorter acute inpatient length of stay. Understanding whether acute hospitals are already paid for these services or whether LTCHs are providing specialized services not available in the acute hospital is poorly understood.

Better measures of acuity are needed to gauge the differences in medical or functional impairments between patients using LTCHs and those using other settings. Additional work in Phase 3 of this project will examine the discharge transitions for acute hospital discharges in areas that lack LTCHs. Using propensity score methods to match patients on diagnosis, severity, and additional factors, as well as control for differences in the availability of services will be important for understanding the potential overlap between acute and LTCH admissions.” (p. 55)

Therefore, we continue to believe that clinical insight offers a significant challenge to the commenters' assertions regarding the alleged existence of some “bright line” which clearly indicates when it is no longer appropriate for a patient to continue treatment in an acute care hospital. Particularly in the case of patients whose conditions fall into the broad category of “medically complex,” clinicians from different provider settings from throughout the country have evaluated existing instruments (that is, Interqual, or MassPRO) and although there appears to be no difficulty in defining a “hospital-level long-term care type patient” there has been considerable difficulty in determining the assignment of such patients to particular provider settings (acute versus LTCH) for purposes of Medicare payment policy.

Accordingly, we are finalizing the extension of the 25 percent (or applicable percentage) threshold policy so that the payment adjustment applies to all subclause (I) LTCHs. We believe it is our responsibility to protect the Medicare Trust Fund from making excessive payments for a single episode of care.

Comment: Many commenters suggested alternatives to specific aspects of the proposed expansion of the proposed 25 percent threshold payment adjustment in the event that we decided to finalize it. A number of commenters suggested that we grandfather existing “freestanding” LTCHs from compliance with the policy because of the significant shift in operation that our policy would mean to their on-going operations. Similarly, these commenters also suggested grandfathering those LTCHs that were already under development (that is, hospitals that were in their 5 of 6 month qualification period for LTCH designation as set forth in § 412.23(e)(3)). Several commenters further suggested that we set a 50 percent threshold for all existing LTCHs and those under development and apply a 25 percent threshold for new LTCHs beginning on July 1, 2007. Other commenters asked us to set the percentage threshold permanently at 50 percent for non-co-located LTCHs in light of our “lesser policy concerns” than we have with LTCH HwHs and satellites. Several commenters urged us to set the threshold for LTCHs in “underserved areas” at 75 percent because of the disparate impact that could be anticipated from implementing this policy. Commenters suggested that we establish a 50 percent threshold for urban LTCHs and a 75 threshold for rural or market dominant LTCHs. We also were requested to apply “temporary, limited” expansion of the threshold while patient and facility level characteristics are being developed and implemented for LTCHs over a 3-year period with the following percentage thresholds: year 1-75 percent; year 2-62.5 percent; year 3-50 percent. According to the commenter, this policy would sunset after year 3 and be replaced by facility and patient criteria.

Response: We appreciate each of the recommendations made by the commenters as to alternatives to extending the 25 percent threshold payment adjustment policy to all subclause (I) LTCHs effective July 1, 2007. We have considered the commenters concerns as we noted earlier, we are finalizing the payment adjustment policy but (as describe elsewhere in these responses), we have provided for a 3-year transition period for all LTCHs and LTCH satellites that will be affected by these changes. Commenters suggested that we exempt currently existing and “under development” LTCHs from the policy because it would require a substantial change in the way that these facilities currently operate. In response to the commenter's question regarding “under development” LTCHs, we are applying the transition to these hospitals as applicable, once they become LTCHs (for example, if a hospital has its first cost reporting period as a LTCH beginning on July 1, 2008, it will be subject to the 50 percent threshold.) We are aware that these new regulations will impact on admission policies at LTCHs (as well as discharge practices at acute care hospitals for patients that continue to need hospital-level care) but such changes are our stated purpose in establishing the original 25 percent threshold payment adjustment policy for co-located LTCHs at § 412.534 and it continues to be our goal for all LTCHs and satellites as we finalize § 412.536. We believe that it is essential that LTCHs reevaluate their existing practices for admittances from referring hospitals. As specified elsewhere in these responses, our data indicates that referring hospitals, primarily acute care hospitals, are discharging patients to LTCHs for continued acute level care when many of these patients could continue to be treated in the acute care hospital. This is particularly true in cases where patient care falls into the broad category of “medically complex.” We believe that Medicare should not be generating two full payments, one under the IPPS and one under the LTCH PPS for what is essentially one episode of care. Although we have had historic concerns with patient-shifting between co-located hospitals, we also believe that it is appropriate to apply the 25 percent (or applicable percentage) threshold payment adjustment to those LTCHs and LTCH satellites that had previously been unaffected by § 412.534, but have similar behavior patterns as co-located HwHs and satellites. (We have responded to concerns about rural, single urban, and Start Printed Page 26933MSA dominant LTCHs elsewhere in these responses.) We would once again remind commenters that the payment adjustment is only applicable for Medicare discharges in excess of the applicable threshold from an individual referring hospital for cases that have not reached outlier status at the referring hospital. We believe that an appropriate and judicious admission policy, on the part each LTCH, could still enable it to admit a specific subset of patients from a referring hospital, prior to the patients' reaching outlier status, and prior to exceeding the applicable threshold. Therefore, even though we continue our work with RTI in Phase 3 of their project to see if we can identify appropriate patient and facility-level criteria for LTCHs, we do not see the development of those criteria and the development of those regulations as contradictory aspects of our fiduciary responsibility for the Medicare program. We further believe that it may be appropriate to establish policies under the LTCH PPS that guard the Medicare Trust Fund from duplicative payments for one episode of patient care even if we are able to develop criteria that identify LTCHs and LTCH-appropriate patients.

Comment: Several commenters expressed concern that the proposed expansion of the 25 percent policy would have a negative impact on Medicare beneficiary access to care, physician choice and authority, and on families of patients who would benefit from LTCH care. Specifically, the commenters noted that LTCHs would be “forced to use a flat 25 percent for each referring hospital, thereby limiting access for Medicare beneficiaries to the level of care deemed most appropriate by their physician.” Another commenter stated that the implementation of the 25 percent rule would force acute care hospitals to keep patients beyond the period for which is medically-appropriate because LTCHs would not be able to accept patients once they met the 25 percent threshold and that overcrowding of acute hospital beds would be the result of the 25 percent policy. Another commenter stated that this policy may result in some patients being transferred to skilled nursing facilities (SNFs) instead of LTCHs, even in cases in which LTCH care would be more appropriate.

Response: We do not believe that the 25 percent policy is unnecessarily “burdensome” or “onerous” to LTCHs for several reasons. The 25 percent policy does not preclude the transfer of any patients from short term acute care hospitals to LTCHs when such transfer is deemed medically necessary and appropriate by the treating physician; rather, it adjusts the payment methodology that is applied to the LTCH for discharges that exceed the applicable threshold. Also, as we noted in the RY 2007 LTCH PPS proposed rule, the payment policy linked to the 25 percent rule helps to remove the perverse incentive that may exist between acute care hospital and LTCH facilities to evade § 412.534 and to prevent both the acute and LTCH from receiving two full Medicare payments for what is essentially one episode of care. Furthermore, this policy also helps to ensure that appropriate transfers from acute to LTCH facilities are occurring based on medical considerations, rather than on the basis of maximizing Medicare payments. We believe that the preexisting relationship between LTCHs and their referring hospitals can be utilized to maximize quality patient care while also making it feasible for LTCHs to comply with the 25 percent policy.

With respect to the commenter's concern that the 25 percent policy would result in transfers to SNFs when LTCH care would be more appropriate, we note that since we are only dealing with patients who require hospital level of care, it would not be appropriate for physicians to transfer these patients to a SNF. However, we do note that it may be appropriate for a subset of LTCH patients, after their condition has stabilized to be transferred to a lower level of care, such as a SNF.

Comment: One commenter noted that Michigan is a “certificate of need” State and that the number of LTCH beds is determined and approved by the State. The commenter further noted that Michigan FIs require that Michigan LTCHs use InterQual admissions standards and recommends that we exempt States who have programs similar to the “certificate of need” because they already adhere to InterQual admissions standards, and therefore, are only treating appropriate “LTCH” patients.

Response: With respect to some LTCHs using InterQual criteria as the standard for admitting a patient, we note that as we stated in the RY 2007 LTCH PPS final rule, InterQual standards focus on the distinction between acute care and sub-acute care, that is, SNF-level of care, and determinations of “medical necessity” or “inappropriate admission” are based only on whether the patient should be hospitalized, rather than on whether the hospitalization should occur at an LTCH or at a general acute care hospital” (71 FR 27869). Furthermore, we recognize and assume that all LTCHs should be using some form of clinical assessment or screening tool to identify appropriate admission candidates; the InterQual is just one model of such a tool that LTCHs may choose to use if they determine that those standards sufficiently identify appropriate patients for their facility. However, we note that the choice of which screening tool an LTCH chooses to use should have no bearing on the percentage of patients being admitted from a particular referring hospital because even under the expansion of the 25 percent policy, it is assumed that all LTCH admissions are hospital-level patients. As explained previously in this section, the expansion of the 25 percent policy is intended to address the situation of an LTCH or satellite that is treating hospital-level patients since it has exceeded the applicable threshold for discharging patients that were admitted from any individual referring hospital and is serving as a unit of the referring hospital. Therefore, we are not exempting LTCHs in “certificate of need” States from the 25 percent policy, but again note that they, along with all other affected LTCH and LTCH satellites will be given a 3-year transition period with respect to implementation of this policy.

Comment: One commenter supported the proposed 25 percent rule and believes that the SSO provision should not apply to subclause II and satellite LTCHs.

Response: We are finalizing our proposal to exempt subclause II and satellite LTCHs from both the 25 percent rule expansion and the SSO policy that we are finalizing in this rule.

Comment: One commenter stated that implementation of the 25 percent rule would result in the following: (1) The loss of local LTCH services in all areas except large metropolitan areas; (2) Patients having to endure long ambulance rides to access LTCH care and possibly being driven past LTCHs with available beds; (3) Families having to drive longer distances to visit their loved ones who may be in LTCHs for extended periods of time; and (4) Some companies, who have already invested in building new LTCHs, possibly being faced with bankruptcy because of the reduced payment associated with the 25 percent rule.

Response: We disagree with the commenter and we do not expect that the 25 percent policy will result in a loss of local LTCH services (in all but large metropolitan areas). Instead, we expect that clinical appropriateness will continue to be used as the standard for LTCH admissions. Since we do not believe that access to LTCH services will be negatively affected by this rule, Start Printed Page 26934we do not believe that beneficiaries will need to endure long ambulance rides to reach an LTCH, nor will families of Medicare beneficiaries have to drive long distances to visit their loved ones. We also remind the commenter that LTCHs will continue to be paid full LTC-DRG payments as long as the 25 percent threshold is not exceeded by any one referral source. In addition, any patients that reach HCO status prior to being transferred to the LTCH would not count towards the 25 percent policy. With regard to the commenter's concern about companies being faced with a financial loss in light of the 25 percent policy expansion, we note that we continue to believe that the LTCH industry can adapt their admission practices to assure that payments will not be reduced, except in rare circumstances. The LTCHs would do this by targeting those patients at referring hospitals that had reached outlier status.

Comment: Some commenters expressed concern that the proposed 25 percent rule would override physician authority and limit physician choice in deciding the most appropriate level of care for his or her patients.

Response: We disagree that this policy overrides physician authority and choice. Rather we believe that this policy appropriately adjusts payments to LTCHs so that the payments reflect the amount of care that is actually provided in the LTCH setting. Furthermore, this policy does not require a change in physician clinical decision-making; rather, it simply seeks to remove any financial incentive that could encourage an LTCH to admit a patient from an acute care hospital prior to that patient receiving a full episode of care at the acute care hospital. Additionally, we would expect that physicians would continue to use their clinical expertise in assessing the level and type of care that is most appropriate for their patients and that the physicians' clinical standards would not be affected by hospital payment policies.

We do not expect that the payment policies implemented in this final rule will deter physicians from making referrals to LTCHs when it is clinically appropriate to do so. We also believe that appropriate clinical care, not payment, should drive physicians' decisions with respect to patients' length of stay and level of care. Additionally, we note that physicians' clinical decisions do not negate the fact that payments should be aligned with the care and resource utilization given in each provider setting.

Comment: Several commenters stated that the payment reductions associated with the proposed 25 percent rule expansion and the proposed “very SSO” policy violate the principles of a PPS in which some cases are expected to cost less than others.

Response: We disagree that these policies violate the principles of averaging found in a PPS. We note that a fundamental premise of the PPS system is that where the costs of some cases may exceed their payment, the opposite is also likely to happen (that is that the costs of some cases will be lower than their payment). As we stated in last year's LTCH PPS final rule, “* * * while some types of cases are always expensive for a hospital to treat, others are, in general, less costly, so it is assumed that hospitals under a DRG-based system, therefore, can typically exercise some influence over their case-mix and their services to achieve fiscal stability” (71 FR 27863). The principles of a PPS begin to break down when there are extreme outliers that are not consistent with the averages calculated, especially when the extreme outliers constitute a disproportionate amount of cases. Additionally, we are attempting to maintain appropriate payment weights for the DRGs by adjusting the LTC-DRG weights for SSO cases. (For a full description of this process, see 71 FR 47978 through 47985). We note that the effect of this adjustment allows the LTC-DRGs to be recalibrated at a weight that is truly representative of average cases instead of at a weight that is skewed towards shorter than average (and presumably, less costly) cases. We also believe that applying the 25 percent (or applicable percentage) threshold payment adjustment to discharges from LTCHs that were admitted from any referring hospital is not a contradiction of the averaging principle intrinsic to PPSs. In fact, one of our rationales for establishing the percentage threshold payment adjustment is to preserve the integrity of the averaging principle under the IPPS because of our concern regarding premature discharges of patients still requiring acute hospital-level care to another acute care provider (and generating another Medicare payment) prior to that case reaching outlier status. Moreover, if LTCHs adjust their procedures so that patients beyond the applicable threshold that are discharged from referring acute care hospitals prior to their LTCH admission have received a full episode of care at the discharging acute (that is, they reach outlier status), Medicare payment for LTCH discharges will be based on the otherwise unadjusted LTCH PPS payment, which has been developed based upon averaging principles.

Comment: Some commenters said that the proposed 25 percent rule would be duplicative of the payment adjustment made under the IPPS post-acute transfer policy. One commenter noted that “* * * 85 percent of DRGs applicable to short-term acute care hospital discharges to LTCHs are subject to [the post-acute transfer] policy.” Another commenter asked CMS to comment on why the IPPS post-acute transfer policy does not appropriately adjust for payment when cases transferred from the acute care hospital ultimately become SSO discharges in the LTCH setting.

Another commenter suggested that we provide policies under the acute IPPS to address inappropriate or early discharges and requested that we use post-acute transfer rules, re-admission rules, and DRGs for acute care hospitals to address the issue of inappropriate transfers instead of penalizing LTCHs.

Response: As we have discussed in the previous LTCH final rules, the IPPS post-acute transfer lessens the incentive for an IPPS hospital to transfer a patient to another hospital early in the patient's stay to minimize its costs while still receiving the full DRG payment from Medicare. Although the post-acute care transfer policy only affects DRGs that meet the criteria specified under § 412.4, we continue to monitor trends in post-acute transfers. In addition, we may make additional DRGs subject to the IPPS post-acute transfer policy if the data demonstrate that it is appropriate to do so. Although we expect the post-acute transfer policy to have an impact on the discharge behavior of acute care hospitals because of the reduced payments that they will receive for qualified discharges, the post-acute transfer policy does not necessarily affect the issues being addressed by the SSO policy change. Both, the IPPS post-acute transfer policy and the proposed RY 2008 SSO policy, help to ensure that Medicare payments are appropriate given the types of treatment provided in each setting.

We believe that the revised payment formula for SSO patients that we are finalizing will appropriately pay LTCHs for delivering services to patients who do not otherwise require the lengths of stay that are characteristic of LTCHs. The SSO policy will address payments to LTCHs for patients discharged from the acute care hospital even after the geometric ALOS.

With respect to the comment about the 25 percent policy being duplicative of the IPPS post-acute transfer provision, we would note that the post Start Printed Page 26935acute transfer policy focuses on a truncated length of stay at an acute care hospital that will be paid for under the IPPS, prior to the case reaching the geometric mean LOS for that DRG as specified in § 412.4(c) and (f). The policy that we are finalizing focuses on determining the appropriate payment to the LTCH, where the patient who has already been treated at the acute care hospital (up to the geometric mean LOS) has been “transferred” to the LTCH care prior to receiving full treatment at the “transferring” hospital. We believe such a stay is a continuation of the patient's original stay at the first hospital, and therefore, that Medicare should pay for such care based on a LTCH PPS payment adjusted to what would otherwise be equivalent to what would have been paid under the IPPS.

Comment: Some commenters wrote in support of extending the comment period from 60 days to 6 months to allow commenters additional time to collaborate for the good of the industry.

Response: We do not believe that a 6-month comment period is warranted or necessary. Consistent with section 1871 of the Act, we provide for a 60-day comment period. This deadline is necessary in order to implement and establish policy changes and payment updates under the LTCH PPS for an effective date of July 1.

We received 270 comments during the comment period and we believe that both the number and the nature of the comments received demonstrate that the comment period was sufficient for commenters to submit relevant and meaningful comments.

Comment: We received many comments that challenged the IPPS-equivalent payment adjustment that we proposed to extend to LTCHs and LTCH satellites for Medicare discharges in excess of the 25 percent (or applicable percentage) threshold that had been admitted from referring hospital not co-located with the LTCH or the satellite of a LTCH.

One commenter maintained that we have determined a payment penalty for freestanding LTCHs for every patient over a 25 percent threshold requiring long term care who is admitted from any single acute care hospital referral source. Another commenter stated that an LTCH could not have more than 25 percent of its patients referred from any one general hospital. Many commenters claimed that our proposal to pay “under the IPPS” for LTCH cases ignores data indicating that LTCHs sustain higher costs than IPPS hospitals in treating Medicare inpatients that are grouped to the same DRG. The commenters stated that costs are higher than they are at acute care hospitals because patients are much sicker than at acute care hospitals. Several commenters included data that indicated that they would sustain substantial financial losses under this policy.

Response: We disagree with the commenters who asserted that under § 412.536 and also the revised § 412.534 we have proposed to pay all LTCHs “under the IPPS” for discharges in excess of 25 percent or the applicable percentage) from an individual referring hospital. As we have noted elsewhere in these responses, if a Medicare beneficiary is treated at an acute care hospital and continues to need further acute hospital-level care, the patient could remain at the acute care hospital. A discharge from the acute care hospital and admission to a LTCH (which is also certified as an acute care hospital) could be appropriately seen as an extension of the stay at the discharging acute care hospital and as such, should not require Medicare to pay for “different resource use”. We further disagree with the commenters who call the extension of the 25 percent threshold a “payment penalty for freestanding LTCHs for every patient over a 25 percent threshold who comes from any single acute care hospital” and the commenter that stated that “an LTCH could not have more than 25 percent of its patients referred from any one general hospital.” As we have noted elsewhere in these responses, the 25 percent threshold is not a patient quota system. By virtue of the fact that more than 25 percent of the LTCH's discharges had been admitted from an individual referring hospital, it is apparent that the LTCH has an ongoing, working relationship with the referring hospital. This policy should lead LTCHs to carefully determine which patients should be admitted from the referring hospital. A patient who is hospitalized in an acute care hospital continues to require acute hospital-level care, generally should not be discharged before the referring hospital has provided the patient with a full episode of care. As discussed elsewhere in these responses, we believe that a patient stay that reaches the HCO threshold at an acute care hospital would be considered to have received a complete episode of care and for such a patient who has received a full episode of care at an acute care hospital, should that patient require further acute level care at a LTCH, Medicare will make an unadjusted additional payment to the LTCH.

Our concern is that many patients that are admitted to LTCHs could have completed this care at the referring hospital to which they were originally admitted. As we have detailed previously in this preamble, in the FY 2005 IPPS final rule (69 FR 48916) we finalized a payment adjustment for co-located LTCHs (that is, HwHs and satellites at § 412.534), which provides that if a LTCH's or satellite's discharges admitted from its host hospital exceed 25 percent (or the applicable percentage) of its discharges for the LTCH HwHs or satellite's cost reporting period, an adjusted payment will be made at the lesser of the otherwise full payment under the LTCH PPS and an adjusted amount under the LTCH PPS that would be equivalent to what Medicare would otherwise pay under the IPPS. In determining whether a hospital meets this percent test, patients transferred from the host hospital that have already qualified for outlier payments at the host would not count as part of the host 25 percent (or the applicable percentage) and the payment for those patients would also not be subject to the adjustment. Those patients would be eligible for an unadjusted payment under the LTCH PPS. (Discharges admitted from the host before the LTCH crosses the 25 percent (or the applicable percentage) threshold would also be paid without the adjustment under the LTCH PPS (69 FR 49213). MedPAC submitted a comment that addressed its concerns with the 25 percent threshold policy for co-located LTCHs in the FY 2005 IPPS final rule.

Specifically, the Commission noted that “freestanding LTCHs also have strong relationships with acute care hospitals, and that where on average LTCH HwHs receive 61 percent of their patients from their hosts, freestanding LTCHs receive 42 percent from their primary referring hospital * * * [that] there are some risks in our proposed 25 percent policy; (a) the 25 percent rule that only applies to LTCH HwHs and not to freestanding LTCHs and may therefore be inequitable; (b) it does not ensure that patients go to the most appropriate post-acute setting; (c) this approach may be circumvented by an increase in the number of freestanding LTCHs instead of LTCH HwH.” As we stated in the FY 2005 IPPS final rule, “MedPAC shares our concern that the LTCH payment system creates an incentive for unbundling of the IPPS in addition to overpayment for the care provided by LTCHs and that this concern is great, particularly, in the case of a LTCH HwH * * *” (69 FR 49211).

In establishing the concept of “functional separateness,” in the FY 1995 IPPS final rule, we were identifying a broader phenomenon than just the relationship between a host Start Printed Page 26936acute care hospital and a LTCH HwH or satellite of a LTCH. We also reviewed MedPAC's comment (discussed previously in this section) on non-co-located LTCH referral patterns and noted that despite the fact that we limited the payment adjustment established in FY 2005 to LTCH HwHs and satellites, “* * * [w]e took considerable note of these comments and the specific information that they included” (59 FR 45391).

We further stated that “* * * [s]ince the October 1, 2004 implementation of the payment adjustment for LTCH HwHs and satellites of LTCHs at § 412.534, through our LTCH PPS monitoring initiative (see section X. of this preamble), we have become aware that the growth in the LTCH universe is now occurring through the development of freestanding LTCHs” and that [r]eviews of public documents posted at the corporate Web site and analysis of the expected consequences of the policy at other investor-oriented sites describe a focus on building freestanding LTCHs, which we believe may imply a response to the payment adjustment for co-located LTCHs established under § 412.534.” At that time, we noted data analyses from FY 2004 and FY 2005 MedPAR files of sole-source (for example, one hospital referring to one LTCH) relationships between acute care hospitals and non-co-located LTCHs and we stated that we believed that the danger of LTCHs functioning as “units” appears to be occurring not only in LTCH HwHs and LTCH satellites but also with freestanding LTCHs (71 FR 27877 through 27879).

We stated that, in many cases, these non-co-located LTCHs and their sole referral source may be functioning in ways that appear to have erased the line of “functional separateness” between these LTCHs and their referring acute care hospitals ((71 FR 27877 through 27879, 59 FR 45391).

Many commenters noted that they would experience considerable financial losses if we implemented the extension of the 25 percent threshold policy. We believe that our finalized policy will result in a behavioral change for LTCHs, and LTCHs will take steps to assure that no more than 25 percent (or the applicable percentage) of the hospital's discharges are patients that had not already reached outlier status at the referring hospital, to assure that all Medicare payments to LTCHs will be made, without adjustment under this policy.

In response to the commenters that asserted LTCH patients are much sicker than acute care patients, we note that it is our understanding from our own data analyses, as well as work done by RTI that costs at LTCHs on a per diem basis are lower than costs for the same DRG at acute care hospitals. For example, RTI performed an analysis of the 2005 MedPAR files and determined the per diem payment for the 20 most common LTC-DRGs treated in LTCHs as outlined in Table 6.

Table 6.—Average Payment per Day for the Top 20 DRGs on LTCH Admissions, LTCH Versus Acute, 2005 MedPAR

  Top 20 LTCH DRGsLTCHAcute
Average paymentAverage length of stayAverage payment per dayAverage paymentAverage length of stayAverage payment per day
475: Respiratory System Diagnosis With Ventilator Support$58,82837.6$1,815$21,69610.4$4,187
271: Skin Ulcers26,65228.81,0095,5256.61,298
087: Pulmonary Edema & Respiratory Failure36,55226.61,4987,2116.31,893
079: Respiratory Infections & Inflammations Age >17 w CC26,54523.71,2358,6548.01,690
088: Chronic Obstructive Pulmonary Disease20,82219.41,1564,4414.81,369
089: Simple Pneumonia & Pleurisy Age >17 w CC22,35620.81,1675,1895.51,355
249: Aftercare, Musculoskeletal System & Connective Tissue21,60125.29143,8163.91,701
416: Septicemia Age >1725,96223.51,1899,3097.42,192
466: Aftercare w/o History of Malignancy as Secondary Diagnosis20,96222.31,0184,6374.71,919
012: Degenerative Nervous System Disorders23,80427.39764,6515.31,298
462: Rehabilitation19,14922.69039,6219.31,125
263: Skin Graft &/or Debrid for Skin Ulcer or Cellulitis w CC41,00642.01,05411,92910.31,930
127: Heart Failure & Shock21,25220.81,0885,4255.01,641
316: Renal Failure25,42023.31,1907,1146.11,936
418: Postoperative & Post-Traumatic Infections25,76625.61,0906,3486.01,633
430: Psychoses15,01927.06513,9557.6869
238: Osteomyelitis27,63930.49737,9347.71,584
277: Cellulitis Age >17 w CC20,00521.79804,4645.31,182
144: Other Circulatory System Diagnoses w CC22,99022.31,1127,2825.72,290
320: Kidney & Urinary Tract Infections Age >17 w CC21,49122.51,0274,3694.91,266
Source: \\rtimas04\hser\Project\08686\006 IPPS\001 LTCH\common\jpotelle\programs\gage030.log.

Furthermore, LTCHs utilize such information regarding their lower costs for treating patients in their advertising. We refer commenters to the following question and answer from the Internet site of a large LTCH chain: The question: “How can a long term acute care hospital be less expensive than a short term acute care hospital?” The answer: “Patients transferred to a long term acute care hospital are medically stable and do not require the critical care resources found in short term acute care hospitals, which are typically the most costly to a patient.”

Comment: Many commenters challenged the basis of the proposed payment adjustment that would result if we finalized our proposed expansion of the 25 percent (or applicable percentage) payment threshold to LTCH and LTCH satellite discharges that were admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH. According to these commenters, in section 123(a)(1) of the BBRA, the Congress specified that the payment policies under the LTCH PPS should “reflect differences in patient resource use and cost.” These commenters asserted that payment adjustments under the LTCH PPS should not be based upon referral sources but rather on the “costs of treatment” and “costs of care” at LTCHs. Start Printed Page 26937

Response: There is considerable precedent regarding our concerns with the financial implications to the Medicare Trust Fund from patient-shifting between acute and post acute settings that could result in two Medicare payments, one to the acute care hospital and another under the LTCH PPS for one episode of care. As noted elsewhere in these responses, this concern was first addressed by the Congress in establishing the post-acute transfer policy at section 1886(d)(5)(J) of the Act, which we subsequently implemented at § 412.4. Furthermore, in the FY 1995 IPPS final rule, we addressed the financial consequences to the Medicare program of the patient-shifting that was occurring between acute care hospitals and co-located LTCHs. At that time, we noted that the “effect of this process is to extend the [LTCH] exclusion to what is for all practical purposes a [LTCH] unit” (59 FR 45389).

We further stated that paying the co-located LTCH as a hospital excluded from the IPPS “may not be appropriate” under these circumstances because “[e]xclusion of long-term care units could inadvertently encourage hospitals to try to abuse the prospective payment systems, by diverting all long-stay cases to the excluded unit, leaving only the shorter, less costly cases to be paid for under the prospective payment systems' (59 FR 45389). Therefore, in accordance with sections 1102 and 1871 of the Act which “confer authority on the Secretary to establish rules and regulations as may be necessary to administer the Medicare program” (59 FR 45390), we established separateness and control criteria at then § 412.23(e)(3)(i) which a co-located LTCH would have to meet to be paid as a hospital excluded from the IPPS. We believed at that time that “the extent to which a facility accepts patients from outside sources can be an important indicator of its status as a separate facility” (59 FR 45392). Therefore, at that time, among other indications of separateness, we adopted a “75 percent referral standard” which required that no more than 25 percent of the LTCHs discharges be admitted from its host to be paid as a hospital excluded from the IPPS. Accordingly, the source of an LTCH's patients as one potential variable since FY 2005 as to whether or not a LTCH receives Medicare payment under the payment system for hospitals excluded from the IPPS, has been a basis for determining whether or not a LTCH was an independent hospital or functioning as a unit of an acute care hospital.

In response to the commenters who maintained that the BBRA mandates that payment under the LTCH PPS is to reflect the “differences in patient resource use and costs” at LTCHs, we note that in general, with respect to the development of the LTCH PPS, section 123(a)(1) of the BBRA requires, among other things, that the Secretary shall develop a PPS and that this PPS shall include an adequate classification system that reflects the difference in resource use and costs. Section 307(b)(1) of the BIPA provides a modification of requirements with respect to the implementation of the PPS. It provides that the Secretary * * * shall examine the feasibility and the impact of basing payments under such a system on the sue of existing (or refined) hospital diagnosis-related groups (DRGs) that have been modified to account for different resource use of long term care hospital patients. The Secretary shall examine and may provide for appropriate adjustments to the long-term care hospital payment system, including adjustments to DRG weights, area wage adjustments, geographic reclassification, outliers, update, and a disproportionate share adjustment * * *. We believe that our payment system fully satisfies these requirements.

If a patient needing additional hospital-level acute care is discharged to another acute care hospital prior to completing a full episode of care at the first hospital, we believe that there is a strong presumption that the second hospital (the LTCH) is behaving like a step-down unit of the first acute care hospital and Medicare will be generating two payments, one under the IPPS and another under the LTCH PPS for one episode of care.

Therefore, we are finalizing our extension of the 25 percent (or applicable percentage) threshold payment adjustment (after the 3-year transition period described elsewhere in this section) for discharges admitted from referring hospital not co-located with the LTCH or the satellite of a LTCH at § 412.536 and grandfathered LTCHs and satellites at § 412.534(h) under the authority of sections 123(a) of the BBRA of 1999 as amended by section 307(b) of the BIPA of 2000 which authorize the Secretary to make adjustments under the LTCH PPS to LTCH hospitals.

In addition, section 123 of the BBRA, as amended by section 307(b)(1) of the BIPA, confers broad discretionary authority on the Secretary to develop and implement a PPS for LTCHs, specifically mandating only “a per discharge prospective payment system” that includes an “adequate payment classification system * * * based on diagnosis-related groups (DRGS) that reflects the differences in patient resource use and costs, and shall maintain budget neutrality.” Section 307 of the BIPA further provides that the Secretary “may provide for appropriate adjustments to the long-term hospital payment system* * *”

As discussed previously, we are finalizing the expansion of the 25 percent (or applicable percentage) payment adjustment (after the 3-year transition period described elsewhere in this section) originally established for co-located LTCHs and satellites with regards to patients admitted to the LTCH from a co-located hospital at § 412.534 to govern the relationship between any referring hospital and an LTCH or LTCH satellite not co-located with that referring hospital. We believe that even in the absence of co-location, the same level of scrutiny must be applied to patient-shifting between acute care hospitals paid for under the IPPS and LTCHs to assure that Medicare is not paying under the IPPS and then generating another unadjusted payment under the LTCH PPS for one episode of care. As discussed elsewhere in these responses, an LTCH is certified as an acute care hospital and we believe that appropriate and responsible payment policy under the Medicare program dictates that if a patient at an acute care hospital paid under the IPPS continues to need treatment at an acute care hospital-level, that patient should remain where he or she is presently being treated until a full episode of care has been delivered prior to being discharged to a LTCH for a different episode of care. We continue to believe that our formulating a payment adjustment for treatment at a second acute care hospital (which is in fact just paid as a LTCH) is both appropriate and necessary for Medicare to be a prudent purchaser of medical care for its beneficiaries. As described above, under this payment adjustment, which we are finalizing at § 412.536 and at revised § 412.534, during a cost reporting period, if an LTCH exceeds the 25 percent threshold of Medicare discharges from any referring hospital (or the applicable adjustment if the referral source is rural, MSA-dominant, or single urban) and the patient did not achieve outlier status at the referring hospital prior to being discharged to the LTCH, Medicare will make a payment adjustment for those discharges under Subpart O for cases beyond the threshold, based upon the lesser of the otherwise unadjusted payment or an adjusted LTCH PPS payment that is Start Printed Page 26938equivalent to the amount that would otherwise be paid under the IPPS.

Comment: Many commenters claimed that the proposed extension of the 25 percent payment threshold is a consequence of our “incorrect assertion” that admission to an LTCH is only legitimate if the patient reaches HCO status at an acute care hospital prior to being discharged for admittance to a LTCH for additional treatment. The commenters believe that under this policy the only way that a patient can receive a full episode of care at an acute is by reaching HCO status. Several commenters quoted data which stated that the percentage of discharges from acute care hospitals which received full Medicare payment is generally close to the percentage of discharges that were admitted to LTCHs that also received a full payment at the acute. The commenters believe that this suggests that a full episode of care is being provided to all of these patients.

Another commenter stated that it is “grossly inappropriate” for CMS to use outlier status as a statistical standard for whether a hospital has furnished a “full “episode of care in a case. Several commenters requested that if we object to two payments for a LTCH patient (that is, one to the referring IPPS hospital and another for payment under the LTCH PPS) we should address the fact that two payments would be generated if the patient was admitted to any post-acute provider such as an IRF or a SNF.

Response: The ultimate goal of our development of payment policy under the LTCH PPS is to assure appropriate and cost-effective payments under the Medicare program for services provided by LTCHs. We have informed the LTCH community in several forums, including notices, that although we were not challenging the high level of care delivered at many LTCHs, it was manifestly unclear how we could identify the point during an acute care hospitalization when a patient would cease to be appropriately placed in that setting such that admission to and further treatment in a LTCH would constitute a reasonable and fiscally responsible standard of care. Our data reveals that approximately 80 percent of LTCH patients are admitted following care at an acute care hospital, where Medicare would have would have paid for their care under the IPPS. We maintain that if a hospitalized patient continues to need acute-level care that such a patient could remain in the acute care hospital for the purpose of receiving this care and not be discharged to another acute care level hospital, like a LTCH until the full episode of treatment has been delivered.

Accordingly, where an LTCH has exceeded the applicable threshold and has thus demonstrated that it is in essence serving as a unit of the referring hospital, it is appropriate to adjust the otherwise payable LTCH PPS payment. We understand that some LTCHs specialize in areas such as ventilator care and weaning or wound care and that some of these facilities are highly respected across all provider settings. However, these same types of patients are being treated by acute care hospitals nationally with similar results. Furthermore, the largest percentage of LTCH patients nationwide would typically fall into the general category of “medically complex.” Nationwide, “medically complex” patients are certainly being successfully treated by acute care hospitals. We have thus far been unable to discover or establish a “bright line” for purposes of demarcating an appropriate discharge from the referring hospital and then admission for appropriate and necessary treatment at an LTCH, paid for under the LTCH PPS. However, since patients who fit the “LTCH profile” are often HCO patients at acute care hospitals (particularly in areas where there is not high LTCH penetration), to determine if a hospital has exceeded its threshold we believe that it is both functional and reasonable to use reaching outlier status at an acute care hospital to determine the delivery of a full episode of care. (RTI report, p. 32-48)

In response to the commenters who noted the comparability of the percentage of all discharges from an acute care hospital that had either reached or not reached outlier status (78 percent) with the percentage of acute care hospital patients who were subsequently admitted to LTCHs following their discharge from the acute care hospital who had either reached or not reached outlier status (also 78 percent), stating that this proved that both had received a “full episode of care,” we do not agree with this conclusion. Furthermore, the commenters data is based on a universe of total discharges from acute care hospitals which is approximately 13 million discharges. The universe of discharges from acute care hospitals to LTCHs is less than 1 percent of those discharges (approximately 112,000). Since the LTCHs are admitting such a small percentage of acute care hospitals' total cases, it is likely that LTCHs are targeting a specific subset of these patients that would have reached outlier status, if not for the presence of the LTCH.

With regard to the comments on patients discharged from acute care hospitals that are admitted to other post-acute providers such as an IRF or a SNF, we would note that there is a distinction in the type of care provided at these settings and at an LTCH. An IRF provides a specialized post-acute service, that is, rehabilitation, for specific medical conditions. A SNF does not even provide hospital-level care. Since an LTCH is certified as an acute care hospital and in fact can provide the same type of care as an acute care hospital that is paid under the IPPS, it is necessary to address the possibility of an LTCH acting as an a unit of an acute care hospital and to differentiate between acute care patients being treated at an (short-term) acute care hospital and those being treated at a LTCH.

We see no correlation between the fact that the commenter has identified a common percentage number and their conclusion that this proves that LTCH patients had received a full episode of care. The fact that nearly 90 percent of LTCH patients had come to the LTCH without achieving outlier status at the acute hospital, which had certainly been providing acute level care to the patient prior to their admission to the LTCH, indicates that for these “medically complex” cases, the acute care hospital may be routinely looking to discharge those patients to the LTCH, prior to their reaching outlier status and thus not receiving a full episode of care at the acute care hospital.

Comment: Several commenters questioned whether the extension of the 25 percent payment threshold would apply to those patients who had been admitted to an LTCH from some other provider setting than an acute care hospital, such as a IRF or a SNF?

Response: The extension of the 25 percent threshold policy to discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH at § 412.536 is based on the policy that we finalized for co-located LTCHs at § 412.534 for FY 2005 in the IPPS final rule (69 FR 48916). As we have stated above, we believe that many of the concerns that we expressed in our analysis of co-located LTCHs, regarding the financially-advantageous but clinically unnecessary shifting of patients from acute care hospitals to LTCHs, is also an issue when the LTCH is not co-located with the referring hospital. Therefore, although the vast majority of host/LTCH HwH or LTCH satellite arrangements are between acute care hospitals and LTCHs, we specified in the FY 2005 final IPPS rule that under § 412.22(e), any inpatient Start Printed Page 26939hospital-level provider could serve as a host to an excluded hospital. Therefore, the policy adjustment that we were finalizing based upon the percentage of patients from one hospital that upon discharge became inpatients at a co-located LTCH, at § 412.534, was also applicable when the host hospital was not an acute care hospital (69 FR 49198).

Furthermore, we stated that applying the option of a discharge payment based upon the lesser of the otherwise unadjusted payment amount under Subpart O or payment under the LTCH PPS based upon an IPPS-equivalent amount was appropriate when the host hospital was an IRF, because “[w]e believe that it is appropriate to pay the LTCH HwH or LTCH satellite that is co-located with an IRF or IPF and exceeds the applicable threshold at the IPPS equivalent rate and not a LTCH PPS rate that would be equivalent to the amount otherwise paid under the IRF or IPF PPS rate, since the HwH and the satellite LTCH are, as we explained earlier in this section, facilities that in many ways are comparable to an acute care hospital” (72 FR 4811; 71 FR 4704 through 4719).

We are finalizing the extension of the 25 percent threshold payment adjustment to discharges from referring hospitals not co-located with the LTCH or the satellite of a LTCH because we believe that our concerns that patient stays are being inappropriately truncated at host hospitals resulting in admissions to LTCH HwHs or satellites also occur between LTCHs and LTCH satellites receiving patients from referring hospitals not on the same campus. As noted elsewhere in this section, we have concentrated on the relationships between referring acute care hospitals and non-co-located LTCHs in this discussion, because approximately 80 percent of Medicare patients in LTCHs are admitted from acute care hospitals. However, we believe that the same concerns, articulated above, would also exist when the patient source is not an acute care hospital. As we noted in the RY 2008 LTCH PPS proposed rule, “[t]here could still be a financial incentive on the part of the referring hospital (for example, an IRF, to prematurely discharge a beneficiary to a LTCH for additional post-acute treatment in order to avoid absorbing high treatment costs under the IRF outlier policy at § 412.624(e)(5)) that would result in two Medicare payments, one to the initial provider and the other for payment under the LTCH PPS for a single episode of beneficiary care” (72 FR 4812). Although we recognize that a patient could experience a medical crisis while an inpatient at an IRF, we would reiterate that typically, the most appropriate setting for such urgent care would be a general acute care hospital, rather than a LTCH. The policy that we are finalizing would not be applicable to a patient admitted to a LTCH from a SNF since a SNF does not deliver hospital-level care and therefore duplication or substitution of services by a LTCH is not a relevant issue.

Comment: One commenter believes that the extension of the 25 percent threshold payment adjustment deprives Medicare beneficiaries of their right to receive medically-necessary services in a LTCH. Therefore, if we finalize the extension of the 25 percent threshold policy, we are violating beneficiary rights and we should provide a notice of non-coverage to beneficiaries regarding this issue. Furthermore, the commenter reminded us that beneficiaries would also be entitled to appeal such a notification to the QIO operating in their State. The commenter stated that the patient whose case would cause the LTCH to exceed the 25 percent threshold referred from a particular referring hospital (that is, the patient who would represent 26 percent) and all those that follow, are entitled to such a notice. The commenter also provides a lengthy discussion of the statutes, regulations, and case law that underlay beneficiary appeal rights.

Response: We would emphasize that we are finalizing a policy in this regulation regarding the payment threshold that Medicare is establishing to avoid generating two payments, one to the initial referring hospital and another under the LTCH PPS, for a single episode of care delivered to a beneficiary. We are not depriving Medicare beneficiaries of their rights to receive treatment at a LTCH, but rather, we have established a payment adjustment for such treatment under particular conditions.

Since the inception of the Medicare program in 1966, policies have been established to determine what the Federal government believes is appropriate payment to hospitals for the delivery of medical services to beneficiaries. Hospitals that elect to participate in the Medicare program are required to comply with the policies established by the program, including the establishment of payment rates and payment adjustments. Therefore, we do not believe that issuing an adjustment that could impact on a hospital's Medicare payments is a radical or unique act. The establishment of a payment policy that may result in payment adjustments for certain admissions is well within the existing regulatory framework. Furthermore, the basis for the policy that we are finalizing at this time, is an extension of a policy that has been in effect since FY 2005, when we established the 25 percent (or applicable percentage) payment threshold policy for co-located LTCHs at § 412.534. At that time, we stated that we were “* * * providing an adjustment to the payment under the LTCH PPS in accordance with the broad authority conferred on the Secretary by the Congress in section 123(a) of the BBRA of 2000 amended by section 307(b) of the BIPA of 2001 to include “appropriate adjustments” in the establishment of a PPS for LTCHs' (69 FR 49204). We continue to believe that there is a clear distinction between medical decision-making and payment policy, particularly * * * when the patient is a Medicare beneficiary and the medically necessary services are covered by Medicare” (69 FR 49204).

LTCHs, for example, are required to meet the greater than 25-day ALOS requirement to retain designation as a LTCH; therefore, LTCHs will factor in that on-going requirement when making specific patient admission decisions during a cost reporting period. The need to comply with various compliance percentage requirements for treating certain conditions in order to qualify for IRF designation, under § 412.23(b), also impacts which patients are admitted to IRFs during a cost reporting period. In these two examples, hospitals currently evaluate admissions during a cost reporting period because a hospital's noncompliance with Medicare requirements regarding LOS and percentage of patients meeting the requirements at § 412.23(b)(2), respectively, could risk its designation as a hospital that is excluded from the IPPS. Therefore, we believe that the circumstance of a LTCH determining which, and under what circumstances, patients should be admitted is an already established feature in the LTCH admission process and should be based on medical criteria and not based on the profitability of treating a specific patient.

Furthermore, the issuance of a Hospital-Issued Notices of Noncoverage (HINNs) by the Medicare program is not applicable to the above described circumstance. Specifically, a LTCH's decision not to admit a specific patient is not a decision by the Medicare program to not cover the service. Rather, it is a determination by the LTCH of the type of service or patient that the facility has a level of expertise in treating. (We specify the conditions under which the Start Printed Page 26940Medicare program is required to issue a HINN on the CMS Web site at http://www.cms.hhs.gov/​BNI/​05_​HINNs.asp#TopOfPage.)

In response to the commenter's belief that a beneficiary who is not admitted to a LTCH because of the payment policy that we are finalizing should appeal the determination to the QIO operating in his or her State, we would state that the decision to admit a patient is made by the hospital. Specifically, section 1802(a) of the Act stipulates that “Any individual entitled to insurance benefits under this title may obtain health services from any institution, agency, or person qualified to participate under this title, if such institution, agency or person undertakes to provide him such services (emphasis added). We emphatically reiterate that we are not preventing the admission of patients to a LTCH; rather, we are establishing a methodology for determining what are fair and reasonable payments based on the type of patient treated by the LTCH. Moreover, it is our expectation that extending the 25 percent (or applicable percentage) payment threshold policy to discharges from referring hospitals not co-located with the LTCH or the satellite of a LTCH will result in LTCHs focusing their mission with respect to referrals from acute care hospitals, and on treating patients that had a complete episode of care at the referring hospital, before being admitted to the LTCH.

Comment: Many commenters stated that there were major differences between the patients treated at LTCHs and at those referred to as “short-term” acute care hospitals. They also listed the significant distinctions between the levels of care delivered by these two types of hospitals. These commenters asserted that acute care hospitals paid under the IPPS are “just not capable” of delivering the level of care required by typical LTCH patients. The commenters noted that MedPAC, RTI, and even CMS have stated that LTCHs effectively treat very sick patients. One commenter stated that there was “evidence that patients who would become subject to the 25 percent rule are different from patients in short term acute care hospitals, and therefore, there is no empirical basis whatsoever for CMS' assumption that LTCHs systematically engage in substitution of service.” According to commenters, LTCHs have specialized care that is not available in acute care hospitals since the treatment model is entirely different. The commenters maintained that acute care hospitals “* * * are diagnosis based where LTCHs provide specialized programs of whole-patients recovery” for patients who require an entire multidisciplinary team. The commenters emphasized that LTCHs use a “* * * team approach towards healing the patient versus stabilizing an acute episode.” They also asserted that LTCHs and acute care hospitals do not treat identical conditions and patients who are forced to remain in an acute care setting could receive “sub-standard care” with the result being poorer health outcomes, longer stays, and even higher costs. The commenter believes that patients who are medically unstable, not progressing, or have failed ventilator-weaning can often benefit from a multidisciplinary program that LTCHs specialize in. In fact, some commenters point to a level of care that is found nowhere else in the medical care continuum but by staff with expertise and experience unique to LTCHs.

Response: In response to the commenters, we would first state the following axiom of hospital policy in the Medicare program: LTCHs, while being unique based on maintaining an average LOS in excess of 25 days, are certified as acute care hospitals and provide hospital-level services to patients. Acute care hospitals paid under the IPPS are throughout the country treating patients requiring hospital-level care often with lengths of stay comparable to those that are typical of LTCHs. We believe the commenters are attempting to establish a clear distinction between the patients that are appropriate for treatment at LTCHs and patients that are appropriately treated at acute care hospitals. Across the United States, the over 3,700 acute care hospitals that discharge approximately 13 million Medicare beneficiaries treat the full range of medical issues including those that the commenters identify as LTCH cases. We do not question that many LTCHs have highly regarded reputations for their success in treating respiratory and ventilator cases (DRG 475), but, as detailed in the RTI report, the 2004 MedPAR files indicate that where LTCHs treated 13,394 cases assigned to DRG 475, acute care hospitals treated 18,727 Medicare patients with an additional 7,072 HCOs, in DRG 475. For DRG 88, chronic obstructive pulmonary disease (COPD), LTCHs treated 4,894 cases where acute care hospitals treated 37,523 cases. Data on other common DRGs treated in LTCHs as compared to the same DRG treated in acute care hospitals reflect a similar pattern, particularly among the DRGs that could fall into the broad category of “medically complex” patients. (Table 3-2, RTI report, p. 35) We understand that MedPAC and RTI noted that many LTCHs deliver a high level of care to very sick Medicare beneficiaries, with fine doctors, exemplary nursing care, and top-notch rehabilitation therapists, but we also know that many acute care hospitals throughout the nation are treating the same types of patients and similarly delivering excellent care. In addition, we are aware that some LTCHs specialize in a particular subset of patients and achieve noteworthy success in their treatment of, for example, ventilator-weaning or wound care; however, similar patients are also receiving care in acute care hospitals with similar results. Therefore, we disagree that acute care hospitals are incapable of competently treating Medicare beneficiaries that happen to fall within the DRGs that LTCHs identify as their specialties and that any patients falling into such categories would receive “substandard” care at an acute care hospital.

Commenters also stated that the Congress established the distinction between acute care hospitals and LTCHs by excluding LTCHs from the IPPS in 1983. In the FY 2003 LTCH PPS final rule (67 FR 55954), which presented the initial payment policies that we established for the LTCH PPS, we briefly reviewed the history of the development of the distinction between hospitals that were to be paid under the IPPS and those that would be excluded, among which were a small group of hospitals that were called LTCHs. In that rule, we stated that “[t]he Congress excluded these hospitals from the acute care hospital inpatient prospective payment system because they typically treated cases that involved stays that were, on average, longer or more costly than would be predicted by the DRG system.” The legislative history of the 1983 Social Security Amendments stated that, “the DRG system was developed for short-term acute care general hospitals and as currently constructed does not adequately take into account special circumstances of diagnoses requiring long stays. (Report of the Committee on Ways and Means, U.S. House of Representatives, to Accompany HR 1900, H.R. Rept. No. 98-25, at 141 (1983)) Therefore, these hospitals could be systemically underpaid if the same DRG system were applied to them (67 FR 55957). Following enactment of the Social Security Amendments of 1983, we implemented the acute care hospital inpatient prospective payment system on October 1, 1983, including the initial publication in the Federal Register of the rules and regulations for the hospital Start Printed Page 26941inpatient prospective payment system—the September 1, 1983 interim final rule” (48 FR 39752, 67 FR 55957).

The 33 LTCHs in existence at the start of the IPPS in 1983 (that were included on the HCFA exclusion list) were described in 1987, in a presentation letter to President George H.W. Bush from then-Secretary Otis R. Bowen, M.D., that preceded a Report to Congress produced by Health Economics Research, Inc. on the “Developing a Prospective Payment System for Excluded Hospitals,” (Department of Health and Human Services, Health Care Financing Administration, Office of Research and Demonstration, HCFA Pub. No. 03262), the Secretary notes that “Long-term Hospitals are a heterogeneous set of institutions located on the Eastern Seaboard, whose mission is the treatment of patients who are seriously or terminally ill with multiple diseases. In other regions of the country, these same patients would be treated in hospitals or skilled nursing facilities * * *”

As discussed in the 1984 Report to Congress, CMS (formerly HCFA) listed 61 hospitals on the “HCFA exclusion list” throughout the United States. (Medicare OSCAR files reveal that 31 of these original facilities are still in existence in 2007.) The Report states that “[t]here were 33 hospitals that both identified themselves as chronic care hospitals * * * [that] are most representative of those primarily providing chronic-disease hospital services. Perhaps of most interest is the very long average LOS of patients in these institutions. With one exception, all average length of stays are over 60 days and, with three exceptions, all are over 100 days. There is probably no clear differentiation between certain types of rehabilitative facilities and LTCHs. The differentiation does seem clearer in the case of psychiatric and children's hospitals, though because these eight psychiatric and three children's hospitals had average lengths of stay greater than 25 days, they were placed under the long-term category of exclusions. The 28 remaining hospitals on the HCFA exclusion list are characterized by a mixture of bed types. Many have a large percentage of psychiatric beds and some a large percentage of rehabilitation beds. Some of those hospitals are institutions with a large number of nursing home beds. For example, one hospital examined houses a small number of acute care beds available for patients routinely cared for in SNF and intermediate care facility (ICF)-level beds. The acute care beds are exempted under PPS. The State licenses beds in this facility as chronic disease hospital beds, though the administrator conceded that these beds are virtually indistinguishable from the SNF and ICF level Medicaid beds * * *” (p. 3-56). The Report identified an additional 25 hospitals that fit the profile of LTCHs, most of which were included in a 1983 AHA Annual Survey. “Lastly, there were 25 hospitals that were not on the exclusion list, but have either self-identified to the HA as chronic care hospitals or have chronic care beds. Seven of these had mostly acute care beds and a short average LOS, such that they would not qualify for the HCFA exclusion. The remaining 18 all had average length of stays greater than 60 days and 11 had average length of stays greater than 100 days. Though several of these were institutions with just chronic care beds, most also had a disproportionate number of nursing home beds. Possibly, those 18 hospitals could qualify for an exclusion at some future point” (p. 3-57). “These hospitals are themselves a diverse, rather anomalous class. As suspected, they have grown up in the interstices of acute, rehabilitation, and nursing home care. Their diversity results from the fact that the role they fill varies with individual State regulatory and financing policies, as well as the surrounding configuration of acute, rehabilitation, and nursing home beds” (p. 3-59).

We quote this report because we believe that it is vital to understand what the Congress was describing when it excluded 33 LTCHs (in the HCFA list) from the IPPS, “* * * because the DRG system was developed for short-term acute care general hospitals and as currently constructed does not adequately take into account special circumstances of diagnoses requiring long stays” and therefore, these hospitals could be systemically underpaid if the same DRG system were applied to them (67 FR 55957). We do not believe that the Congress was identifying the LTCHs in existence in 1983, described above, as facilities expected to deliver care at a level of medical sophistication equivalent to or even surpassing that of a typical acute care hospital.

In 1983, there were 33 LTCHs (plus another 25 from the AHA list); in 1993, there were 105; in 2003, there were 318; and in 2007, there are nearly 400 LTCHs. We do not doubt that the nature and level of the care delivered by most LTCHs has changed markedly since 1983 but we believe that it is both highly inaccurate and misleading to state, as some of our commenters have, that “ ‘short term’ acute care hospitals are “just not capable” of delivering the level of care required by typical LTCH patients; that acute care hospitals “are diagnosis based where LTCHs provide specialized programs of whole-patients recovery;” that acute care hospitals do not treat identical conditions and that patients who are forced to remain in an acute care setting could receive “sub-standard care with the result being poorer health outcomes, longer stays, and even higher costs.” We do not believe that the evidence detailed above indicates that in excluding LTCHs from the IPPS and explaining this act by the above-quoted rationale in 1983, that it was the Congress' intention to declare that henceforth, certain patients could only reasonably be treated in LTCHs and that treatment at an acute care hospitals for such patients would be “sub-standard.” Rather, we believe that the Congress was attempting to describe the provider landscape as it existed at that time and that in so doing, there was a small group of facilities that did not “cleanly” fit into any other category, having “grown up in the interstices of acute, rehabilitation, and nursing home care.” Report to Congress on the “Developing a Prospective Payment System for Excluded Hospitals,” HCFA Pub. No. 03262) (p. 3-59).

Since that time, there have been changes in the LTCH universe, with over 58 percent of the nearly 400 LTCHs being run for-profit (the majority by several large chains); approximately 33 percent run not for profit, and only 8.3 percent now run by a government instrumentality. Accordingly, we believe that the policy we proposed is appropriate to deal with present payment issues that the Medicare program is facing under the LTCH PPS.

Commenters further asserted that acute care hospitals do not and even can not deal with the medical conditions in which LTCH specialize. Even though the LTCH universe has grown to nearly 400, they continue to not be evenly geographically dispersed and therefore, by far, most very sick Medicare inpatients nationwide are treated in acute care hospitals. In FY 2005, there were 130,000 LTCH discharges and 12.7 million discharges from acute care hospitals. A brief review of several major LTCH Web sites contained the following list of conditions in which they specialize:

  • Chronic cardiac disorders;
  • Neuuromuscular/neurovascular diseases
  • Methicillin-resistant staph aureus (MRSA)
  • Complex orthopedic conditions
  • Wound care complications
  • Multi-system organ failure Start Printed Page 26942
  • Immuno-suppressed conditions
  • Respiratory failure
  • Dysphagia management
  • Post-operative complications
  • Multiple intravenous therapies
  • Chemotherapy
  • Pre- and post-operative organ transplant care
  • Chronic nutritional problems
  • Total parenteral nutrition (TPN) issues'
  • Intensive hemodynamic monitoring
  • Renal dialysis
  • Telemetry
  • EKG testing
  • Diagnostic bronchoscopy and endoscopy
  • Speech-language pathology
  • Surgery support
  • Nutritional therapy
  • Radiology services
  • Laboratory services
  • Respiratory therapy
  • Physical therapy
  • Occupational therapy
  • Pharmacy
  • Social services

Furthermore, the list of services noted above, are also hardly unique to the LTCH setting.

Comment: One commenter cited several provisions of Federal and State statutes that generally refer to patient transfers, services furnished to a hospital's patients by others under arrangements made by the hospital with them, or a hospital's responsibility to have services available to meet the needs of patients it accepts for treatment. For example, the commenter cites the provision of the Emergency Medical Treatment and Active Labor Act (EMTALA) (specifically, section 1867(g) of the Act) that requires hospitals with specialized capabilities to accept appropriate transfers of unstabilized individuals protected by EMTALA. The commenter also referred to Florida, Texas, and Illinois legislation authorizing arranged-for services and referral and transfer agreements, and The Joint Commission (formerly JCAHO) guidance directing their surveyors to look closely at transfers. However, no specific comment was made.

Response: We do not believe this discussion in any way calls into question the need for the provisions relating to the policies we have proposed. Though the provisions cited do include references to transfers, they do not spell out conditions under which they are acceptable or otherwise establish specific standards to ensure that transfers and services under arrangements do not jeopardize patient health and safety. More importantly, they do not address the key issue of transfers that may not create clear risks for patients, but nevertheless, increase costs in the health care system because they are undertaken for financial rather than medical reasons. Therefore, even though we reviewed this discussion carefully, we made no changes to our proposals based on it.

Comment: Some commenters highlighted the current medical care situation in New Orleans noting that the city is still trying to recover from Hurricane Katrina. The commenters believed that the proposed changes would result in the closure of LTCHs and this would cause hardships on the limited number of physicians practicing in the area. The commenters requested that affected hospitals should be granted a time limited exemption from these rules for up to 5 years.

Response: We are certainly aware of the current state of medical care in Louisiana in general, and specifically in the New Orleans area. We have worked and continue to work closely with State officials and the hospitals in Louisiana to address issues that are important to helping the State rebuild its medical care infrastructure. As stated previously in response to commenters who claimed that these revisions would cause LTCHs to close, we believe that these changes are necessary to assure that the Medicare program is making appropriate payments to these hospitals in the specific situations addressed by these policies. In the case of the expansion of the 25 percent policy to apply to LTCHs and satellites that exceed the threshold on discharges that were admitted from a referring hospital not co-located with the LTCH or LTCH satellite, since a LTCH is certified as an acute care hospital, we believe it is appropriate to pay the LTCH under the LTCH PPS a rate that is comparable to the rate paid under the IPPS, where it is demonstrating behavior that indicates that it is serving as a “unit” of the referring hospital. Similarly, the revised SSO policy also provides for payments to the LTCH for those SSO cases that have a LOS that is comparable to the LOS of a typical IPPS patient in the same DRG, under the LTCH PPS at an adjusted rate that is comparable to the IPPS rate. We do not believe these policies will cause widespread closure of LTCHs nationally or in Louisiana.

We also note that while in general the threshold under the expansion of the 25 percent policy as finalized in this rule will ultimately be 25 percent, in response to comments requesting that we transition the implementation of this policy, as discussed earlier we are providing for a 3-year transition to allow hospitals additional time to comply with the 25 percent threshold. Therefore, we are establishing a 75 percent threshold for RY 2008 and a 50 percent threshold for RY 2009. The threshold will be reduced to 25 percent beginning with RY 2010. Furthermore, for hospitals in rural areas or those admitting patients from a single hospital MSA effective with RY 2008, the threshold will be 75 percent for RY 2008 and will remain at 50 percent for subsequent rate years. In addition, for LTCHs admitting patients from MSA-dominant hospitals, effective with RY 2009 the threshold will be adjusted based on the referring hospital's percentage of Medicare patients discharged in the MSA, and will be not less than 25 percent and not more than 50 percent.

Comment: Many commenters requested that we clarify how they would be able to comply with the requirements of the 25 percent threshold payment adjustment policy if it was finalized. In the particular situation of a MSA-dominant or urban single hospital, where the threshold depends upon the percentage of referring hospital discharges in that MSA, it was requested that we clarify which year of data would applicable.

Response: In establishing this payment provision, originally for co-located LTCHs for FY 2005, we consulted with Medicare's FIs and we were assured that LTCHs will be able to obtain the information that they need in order to comply with this policy from the referring hospital from which they would be admitting patients.

Further, we understand that typically, acute care hospitals have the GROUPER software which enables them to determine the most likely DRG assignment for their patients and additionally, programs that track the costs being incurred by their patients on a daily basis. Therefore, they are with a high degree of accuracy, able to predict when a particular case crosses the outlier threshold. To facilitate such practices by hospitals, we have provided PRICER software for Medicare PPSs available for download on the CMS Web site. We understand that hospitals, including LTCHs, generally also purchase GROUPER software to track DRG assignments.

Therefore, it is our expectation that LTCHs and their referring hospitals will build on their existing working relationship (since this policy applies to situations where over 25 percent of a LTCH's patients were admitted from an individual hospital) and will find it in their mutual interests to share necessary information. We would also expect LTCHs to monitor their admissions and Start Printed Page 26943discharges from their referring hospitals, a process in which they would typically engage as a component of sound business practice.

In response to the comment questioning the determination of the applicable MSA-dominant or urban-single percentage for purposes of LTCH calculations, we agree that it would be inappropriate for this percentage to be based on data occurring during a cost reporting period. Therefore, we would note that our policy is to base the percentage on the latest available discharge data that is available prior to the beginning of the LTCH's current fiscal year. We are revising proposed § 412.536(d)(2) to reflect this policy. Furthermore, in response to this comment, at this time, we are also revising the regulation text as it applies to co-located LTCHs. Specifically, at § 412.534(e)(2) where we describe the determination of the percentage threshold for MSA-dominant hosts for LTCH HwHs and LTCH satellites, we deleting the phrase, “for the cost reporting period for which the adjustment was made”.

Comment: One commenter stated that implementing the 25 percent threshold payment adjustment policy, under which Medicare payments would be reconciled, would “violate a fundamental rule of PPSs that payments will be prospectively set and known in advance by the providers.” This commenter also stated that the finalizing this regulation would “in a very real sense, would convert the LTCH PPS into a retroactive system of recovery and settlement with related disputes where CMS would be called upon to produce patient records from hospitals that refer cases to LTCHs as well as individual patient coding and referral hospital financial information to support recovery claims.”

Response: In response to these concerns, we would note that the cost report settlement process (governed by Subpart B of Part 413) is a standard feature of all Medicare PPSs. For example, under the IPPS, a hospital DRG payment may be subject to the DSH or IME adjustments. The DSH adjustment is based on the percentage of Medicaid patients discharged by the hospital during the fiscal year, while the IME adjustment is based on the number of residents trained by the hospital during the fiscal year. Both factors are subject to change based on final settlement of the hospital's cost report. The procedures that we have established for this process envision a reconciliation between hospitals and the Medicare program based on claims submission, special interim payments or periodic interim payments and the final amounts due, as determined by the FI. Such reconciliations are both necessary and expected. There are numerous provisions affecting LTCHs that could result in subsequent redetermination of the payment amounts. For example, involvement of a QIO review of a DRG assignment which may result in a change in DRGs as specified in § 412.513(c), as well as any of the reconsiderations and appeals provided for under subparts G, I, J, or R of Part 405. Moreover, since the start of the LTCH PPS, our regulations on special payment provisions for patients who are transferred to onsite providers and readmitted to a LTCH at § 412.532, specified a 5 percent threshold for LTCH readmittances of patients that had been discharged to an onsite acute care hospital. Payments under this policy would be reconciled following cost report settlement. Finally, the 25 percent threshold for co-located LTCHs, which could result in a redetermination of the payment amount if the threshold is exceeded, has been in effect since FY 2005.

Therefore, we do not believe that the principle of PPS issued by the Medicare program is inconsistent with the extension of the 25 percent payment adjustment threshold under the LTCH PPS.

Comment: Several commenters stated that both of our policy proposals, the extension of the 25 percent threshold policy adjustment and the revision of the SSO policy, are effectively establishing “admission criteria” which usurp the exclusive role of QIOs in the Medicare program.

Response: We reiterate that with the finalization of the extension of the 25 percent threshold policy adjustment and the SSO policy, we have not established “admissions criteria” for LTCHs. Rather, in keeping with our fiduciary responsibility to oversee Medicare expenditures, we have established payment policies that provide for appropriate Medicare payments for beneficiary care. We describe each of the policies in detail in this preamble. They are distinct policies but they both focus on our goal of determining payment for Medicare services delivered in LTCHs, under particular circumstances that we believe should not significantly exceed payment for similar services otherwise delivered in acute care hospitals.

Because the comments that we received regarding the QIO's role and the implementation of the expansion of the 25 percent threshold policy were fundamentally the same comments submitted regarding the QIOs role and the SSO policy revision, we responded to comments in the SSOs section of this final rule.

In summary, we are finalizing a new provision at § 412.534(h) that effective with discharges occurring during cost reporting periods beginning on or after July 1, 2007, would apply the policies established under existing § 412.534 to grandfathered subclause (I) LTCH HwHs and LTCH satellites for Medicare discharges that were admitted from their co-located host hospitals. We are also applying those policies for Medicare discharges admitted from referring hospitals not co-located with the LTCH or the satellite of a LTCH to all subclause (I) LTCHs and LTCH satellites at § 412.536, generally tracking § 412.534, where applicable. For example, in determining whether a hospital meets the 25 percent criterion, Medicare discharges that have already qualified for outlier payments at the referring hospital would not be included in the count of Medicare discharges admitted from the referring hospital. (We are entitling § 412.536, Special Payment Provisions for LTCHs and Satellites of LTCHs that Discharged Medicare Patients Admitted From a Hospital Not Located in the Same Building or on the Same Campus as the LTCH or Satellite of the LTCH.)

We are also finalizing adjustments to the 25 percent policy at § 412.536 for specific circumstances consistent with the policy for co-located hospitals under § 412.534. At § 412.536(c) for Medicare discharges from subclause (I) LTCHs or LTCH satellites located in rural areas, Medicare discharges in excess of 50 percent, rather that 25 percent of the LTCH's total Medicare discharges for a cost reporting period from an individual referring hospital not co-located with the LTCH or the satellite of the LTCH would be subject to the payment adjustment specified at § 412.536(c). In addition, in the case of a rural subclause (I) LTCH or LTCH satellite facility, in determining the percentage of Medicare discharges admitted from the referring hospital, any patients that had been Medicare outliers at the referring hospital and then discharged to the LTCH or LTCH satellite are not counted towards the threshold percentage (as described above).

In § 412.536, we are also providing that if the referring hospital not co-located with the LTCH or satellite of the LTCH is the only other hospital in the MSA or is an MSA-dominant hospital as defined at § 412.536(e)(4), we are allowing the subclause (I) LTCH or LTCH satellite facility a threshold percentage equal to the non-co-located referring hospital's percentage of total Start Printed Page 26944Medicare discharges for hospitals in the MSA. Consistent with our policy at existing § 412.534(e), we are applying a floor of 25 percent and a ceiling of 50 percent to this threshold for these hospitals. As with the existing policy for co-located LTCHs, we believe that this adjusted payment threshold responds to “the unique needs of these communities” (69 FR 49207). Similar to the existing provisions at § 412.534, in determining the percentage of Medicare discharges admitted to the LTCH or LTCH satellite facility from the urban single or MSA dominant hospital, any patients that had been Medicare outliers at the referring hospital before being admitted to the LTCH or LTCH satellite would not count towards the applicable threshold, as discussed above.

The payment adjustment at § 412.536 will be phased-in over 3 years for all LTCH discharges affected by the policies that we are finalizing beginning for cost reporting periods beginning on or after July 1, 2007. Under the phase in, the percentage threshold will be the greater of the applicable threshold as specified at 412.536(b),(c), and (d) or the following percentages: For cost reporting periods beginning on or after July 1, 2007 and before July 1, 2008, under the policy that we are finalizing at § 412.536, the percentage of Medicare discharges that may be admitted from a referring hospital not co-located with the LTCH or the satellite of a LTCH with no payment adjustment is the lesser of the percentage of Medicare discharges admitted from the referring hospital during its RY 2005 cost reporting period or 75 percent. For cost reporting periods beginning on or after July 1, 2008 and before July 1, 2009, under the policy that we are finalizing at § 412.536, the percentage of Medicare discharges that may be admitted from the referring hospital not co-located with the LTCH or the satellite of a LTCH, with no payment adjustment, is the lesser of the percentage of Medicare discharges admitted from the referring hospital during its RY 2005 cost reporting period or 50 percent. For cost reporting periods beginning on or after July 1, 2009 (RY 2010), all subclause (I) LTCHs and LTCH satellites will be subject to the 25 percent (or applicable percentage) threshold payment adjustment for discharges during a cost reporting period that were admitted from any referring hospital. In determining the percentage of Medicare discharges admitted from the referring hospital, patients who reached HCO status at the referring hospital before being admitted to the LTCH or LTCH satellite will not count towards the applicable threshold, as discussed above. A similar phase is provided for the expansion at § 412.534 to grandfathered subclause (I) LTCH HwHs and LTCH satellites.

Finally, we believe that these payment adjustments address policy concerns that are consistent with those that we originally expressed when we implemented the payment adjustment for LTCHs discharging patients that were admitted from co-located hospitals.

We also believe that it is important, once again, to note that the 3-year transition to the full 25 percent threshold payment adjustment will coincide with our continuing work on the MedPAC recommendations to attempt to develop facility and patient level criteria for LTCHs. We hope that the LTCH industry will work closely with CMS to pursue this endeavor during the transition period.

VI. Computing the Adjusted Federal Prospective Payments for the 2008 LTCH PPS Rate Year

In accordance with § 412.525 and as discussed in section IV.C. of this final rule, the standard Federal rate is adjusted to account for differences in area wages by multiplying the labor-related share of the standard Federal rate by the appropriate LTCH PPS wage index (as shown in Tables 1 and 2 of the Addendum to this final rule). The standard Federal rate is also adjusted to account for the higher costs of hospitals in Alaska and Hawaii by multiplying the nonlabor-related share of the standard Federal rate by the appropriate cost-of-living factor (shown in Table 3 in section IV.D.2 of this preamble). In the RY 2007 LTCH PPS final rule (71 FR 27827), we established a standard Federal rate of $38,086.04 for the 2007 LTCH PPS rate year. In this final rule, as was proposed, based on the best available data and the policies described in this final rule, the standard Federal rate for the 2008 LTCH PPS rate year will be $38,356.45 as discussed in section IV.C.3. of this preamble. We illustrate the methodology that will be used to adjust the Federal prospective payments for the 2008 LTCH PPS rate year in the following examples:

Example

During the 2008 LTCH PPS rate year, a Medicare patient is in a LTCH located in Chicago, Illinois (CBSA 16974). This LTCH is in the final year of the wage index phase-in, thus, the full (that is, five-fifths) wage index values are applicable. The full LTCH PPS wage index value for CBSA 16974 is 1.0751 (see Table 1 in the Addendum to this final rule). The Medicare patient is classified into LTC-DRG 9 (Spinal Disorders and Injuries), which has a current relative weight of 1.0424 (see Table 3 of the Addendum to this final rule).

To calculate the LTCH's total adjusted Federal prospective payment for this Medicare patient, we compute the wage-adjusted Federal prospective payment amount by multiplying the unadjusted standard Federal rate ($38,356.45) by the labor-related share (75.788 percent) and the wage index value (1.0751). This wage-adjusted amount is then added to the nonlabor-related portion of the unadjusted standard Federal rate (24.212 percent; adjusted for cost of living, if applicable) to determine the adjusted Federal rate, which is then multiplied by the LTC-DRG relative weight (1.0424) to calculate the total adjusted Federal prospective payment for the 2008 LTCH PPS rate year ($42,258.45). (As discussed in section IV.C.5. of this preamble, for the 2008 LTCH PPS rate year, we are no longer applying a transition period BN offset (to account for the costs of the transition methodology) in determining the total adjusted Federal prospective payment.) Table 7 illustrates the components of the calculations in this example.

Table 7

Unadjusted Standard Federal Prospective Payment Rate$38,356.45
Labor-Related Share× 0.75788
Labor-Related Portion of the Federal Rate= $29,069.59
Full Wage Index (CBSA 16974)× 1.0751
Wage-Adjusted Labor Share of Federal Rate= $31,252.71
Nonlabor-Related Portion of the Federal Rate ($38,356.45 × 0.24212)+ $ 9,286.86
Adjusted Federal Rate Amount= $40,539.57
LTC-DRG 9 Relative Weight× 1.0424
Start Printed Page 26945
Total Adjusted Federal Prospective Payment *= $42,258.45
* We are no longer applying a transition period BN offset to account for the costs of the transition methodology in determining the total adjusted Federal prospective payment for RY 2008.)

VII. Transition Period

To provide a stable fiscal base for LTCHs, under § 412.533, we implemented a 5-year transition period whereby a LTCH (except those defined as “new” under § 412.23(e)(4)) received a LTCH PPS payment consisting of a portion based on reasonable cost-based reimbursement principles under the TEFRA system and a portion based on the Federal prospective payment rate (unless the LTCH elected payment based on 100 percent of the Federal rate). As discussed in the August 30, 2002 final rule (67 FR 56038), we believed that a 5-year phase-in provided LTCHs time to adjust their operations and capital financing to the LTCH PPS, which is based on prospectively determined Federal payment rates. Furthermore, we believed that the 5-year phase-in under the LTCH PPS also allowed LTCH personnel to develop proficiency with the LTC-DRG coding system, which will result in improvement in the quality of the data used for generating our annual determination of relative weights and payment rates.

Under § 412.533, the 5-year transition period for all hospitals subject to the LTCH PPS began with the hospital's first cost reporting period beginning on or after October 1, 2002 and extends through the hospital's last cost reporting period beginning before October 1, 2007. During the 5-year transition period, a LTCH's total PPS payment under the LTCH PPS was based on two payment percentages—one based on reasonable cost-based principles and the other based on the standard Federal prospective payment rate. The percentage of the LTCH PPS payment based on the LTCH PPS Federal rate increased by 20 percentage points each year, while the reasonable portion of the LTCH PPS payment based on cost-based principles decreased by 20 percentage points each year, for the next 4 fiscal years. For cost reporting periods beginning on or after October 1, 2006, Medicare payment to LTCHs will be determined entirely under the Federal rate.

In implementing the LTCH PPS, one of our goals was to transition hospitals to prospective payments based on 100 percent of the adjusted Federal prospective payment rate as soon as appropriate. Therefore, under § 412.533(c), we allowed a LTCH (other than new LTCHs defined at § 412.23(e)(4)), which was subject to a blended rate, to elect payment based on 100 percent of the Federal rate at the start of any of its cost reporting periods during the 5-year transition period. Once a LTCH elected to be paid based on 100 percent of the Federal rate, it could not revert back to the transition blend.

VIII. Payments to New LTCHs

Under § 412.23(e)(4), for purposes of Medicare payment under the LTCH PPS, we define a new LTCH as a provider of inpatient hospital services that meets the qualifying criteria for LTCHs, set forth in § 412.23(e)(1) and (e)(2), and under present or previous ownership (or both), has its first cost reporting period as a LTCH beginning on or after October 1, 2002. As we discussed in the August 30, 2002 final rule (67 FR 56040), this definition of new LTCHs should not be confused with those LTCHs first paid under the TEFRA payment system for discharges occurring on or after October 1, 1997, described in section 1886(b)(7)(A) of the Act, as added by section 4416 of the Balanced Budget Act of 1997 (BBA) (Pub. L. 105-33).

Under § 412.533(d), new LTCHs, as defined in § 412.23(e)(4), will be paid based on 100 percent of the standard Federal rate. As we discussed in the August 30, 2002 final rule (67 FR 56040), the transition period was intended to provide existing LTCHs time to adjust to payment under the new system. Since these new LTCHs with their first cost reporting periods as LTCHs beginning on or after October 1, 2002, would not have received payment under reasonable cost-based reimbursement for the delivery of LTCH services prior to the effective date of the LTCH PPS, we did not believe that those new LTCHs required a transition period in order to make adjustments to their operations and capital financing, as will LTCHs that have been paid under the reasonable cost-based methodology.

IX. Method of Payment

Under § 412.513, a Medicare LTCH patient is classified into a LTC-DRG based on the principal diagnosis, up to eight additional (secondary) diagnoses, and up to six procedures performed during the stay, as well as age, sex, and discharge status of the patient. The LTC-DRG is used to determine the Federal prospective payment that the LTCH will receive for the Medicare-covered Part A services the LTCH furnished during the Medicare patient's stay. Under § 412.541(a), the payment is based on the submission of the discharge bill. The discharge bill also provides data to allow for reclassifying the stay from payment at the full LTC-DRG rate to payment for a case as a SSO (under § 412.529) or as an interrupted stay (under § 412.531), or to determine if the case will qualify for a HCO payment (under § 412.525(a)).

Accordingly, the ICD-9-CM codes and other information used to determine if an adjustment to the full LTC-DRG payment is necessary (for example, LOS or interrupted stay status) are recorded by the LTCH on the Medicare patient's discharge bill and submitted to the Medicare FI for processing. The payment represents payment in full, under § 412.521(b), for inpatient operating and capital-related costs, but not for the costs of an approved medical education program, bad debts, blood clotting factors, anesthesia services by hospital-employed nonphysician anesthetists or the costs of photocopying and mailing medical records requested by a Quality Improvement Organization (QIO), which are costs paid outside the LTCH PPS.

As under the previous reasonable cost-based payment system, under § 412.541(b), a LTCH may elect to be paid using the periodic interim payment (PIP) method described in § 413.64(h) and may be eligible to receive accelerated payments as described in § 413.64(g).

For those LTCHs that are being paid under the transition methodology set forth at § 412.533, for cost reporting periods that began on or after October 1, 2002, and before October 1, 2006, the PIP amount is based on the transition blend. For those LTCHs that are paid based on 100 percent of the standard Federal rate, the PIP amount is based on the estimated prospective payment for the year rather than on the estimated reasonable cost-based reimbursement. We exclude HCO payments that are paid upon submission of a discharge bill from the PIP amounts. In addition, Part A costs that are not paid for under the LTCH PPS, including Medicare costs of an approved medical education program, bad debts, blood clotting Start Printed Page 26946factors, anesthesia services by hospital-employed nonphysician anesthetists and the costs of photocopying and mailing medical records requested by a QIO, are subject to the interim payment provisions as specified in § 412.541(c).

Under § 412.541(d), LTCHs with unusually long lengths of stay that are not receiving payment under the PIP method may bill on an interim basis (60 days after an admission and at intervals of at least 60 days after the date of the first interim bill) and this should include any HCO payment determined as of the last day for which the services have been billed.

X. Monitoring

In the August 30, 2002 final rule (67 FR 56014), we described an on-going monitoring component to the new LTCH PPS. Specifically, we discussed on-going analysis of the various policies that we believe would provide equitable payment for stays that reflect less than the full course of treatment and reduce the incentives for inappropriate admissions, transfers, or premature discharges of patients that are present in a discharge-based PPS. As a result of our data analysis, we have revisited a number of our original and even pre-LTCH PPS policies in order to address what we believe are behaviors by certain LTCHs that lead to inappropriate Medicare payments. In recent Federal Register publications, we have proposed and subsequently finalized revisions to the interruption of stay policy in the RY 2005 LTCH PPS final rule (69 FR 25692), and we established a payment adjustment for LTCH HwHs and satellites in the FY 2005 IPPS final rule (69 FR 49191 through 49214). In section V.A.2., we revisited the payment adjustment methodology established for SSOs (71 FR 27845) as a consequence of recent data analysis and are finalizing a policy which revises one of the existing four alternatives under the existing SSO payment methodology for certain SSO cases to an amount under the LTCH PPS that is comparable to an amount that would otherwise be paid under the IPPS.

As we discuss in section X. of this final rule, our monitoring of discharges between acute care hospitals and LTCHs reveals that a significant number of LTCHs that are “freestanding”, that is, not co-located with other hospital-level providers (as defined in § 412.22(e) and § 412.22(h)), admit their patients from one specific acute care hospital. When we established the payment adjustment for LTCH HwHs and satellites of LTCHs at § 412.534, we stated our concern that these on-site LTCHs could be functioning as units of their host (generally, an acute care hospital), a configuration that is not permitted in section 1886(d)(1)(B) of the Act. (The statute specifically allows only for IRF and IPF units in acute care hospitals, but not for LTCH units.) As a result of our data monitoring and analysis, which is detailed in section V.B. of this final rule, we are expanding the existing payment adjustment at § 412.534 and we developed new § 412.536 to apply to certain situations not currently covered by the existing policy for LTCHs co-located with other hospitals.

As we discussed in the RY 2004 LTCH PPS final rule (68 FR 34157), the Medicare Payment Advisory Commission (MedPAC) endorsed our monitoring activity as a primary aspect of the design of the LTCH PPS. Furthermore, the Commission pursued an independent research initiative that led to a section in MedPAC's June 2004 Report to Congress entitled “Defining long-term care hospitals”. This study included recommendations that we develop facility and patient criteria for LTCH admission and treatment and that we require a review by QIOs to evaluate whether LTCH admissions meet criteria for medical necessity once the recommended facility and patient criteria are established (70 FR 24209). In response to the recommendation in MedPAC's June 2004 Report, we awarded a contract to Research Triangle Institute, International (RTI), on September 27, 2004, to conduct a thorough examination of the feasibility of implementing MedPAC's recommendations.

RTI has completed its examination of the feasibility of implementing MedPAC's recommendations in the June 2004 Report to Congress, and as discussed in section XI. of the preamble to this final rule. Both Phases I and II are posted on the CMS Web site (as noted below). We also reproduced the Executive Summary of the report in Addendum B of the RY 2008 LTCH PPS proposed rule (72 FR 4884 through 4886). At that time, we noted, “[t]his material is being reproduced as received from the contractors and does not represent out position or policy” (72 FR 48181).

We are continuing to pursue our on-going program, existing QIO monitoring and studies described in the RY 2006 LTCH PPS final rule (70 FR 24211), and our considerations of expanding the QIO role in the LTCH PPS.

Comment: We received several letters from various Congressional delegations that were critical of the proposed revision to the SSO policy and the extension of the 25 percent threshold payment adjustments. The commenters stated that these policies do not achieve CMS' goal of identifying inappropriate LTCH admissions.

The commenters urged us to establish patient and facility-level criteria for LTCHs to better define the appropriate patient setting and medical conditions required for admission. A number of the commenters further stated that LTCHs admit patients only after applying an objective and rigorous set of admissions screening criteria and Medicare QIOs conduct post-admission reviews of LTCH patients to ensure that admissions are medically-necessary. These commenters further stated that at our direction, QIOs have been reviewing a sample of LTCH cases for admission appropriateness and that these reviews “clearly” show an immaterial number of LTCH claims denied as the result of QIO reviews. Therefore, the commenters maintained that QIO review data does not support our assumption that cases were inappropriately admitted to LTCHs, but rather, QIOs are overwhelmingly finding that LTCH patients have appropriately been admitted and treated in LTCHs.

Response: We reiterate that QIO review of Medicare cases, either based upon the national sample or resulting from specific appeals, presently determine, among other things, whether a patient required hospital-level care. The QIO reviews presently do not distinguish between acute care settings, such as acute care hospitals paid under the IPPS or acute care hospitals paid under the LTCH PPS. Therefore, although the QIO review process, as presently constituted, is a vital component of the Medicare program, the role played by the QIOs does not, at this time, provide a medium through which we can determine appropriate payment policy for acute care hospital patients who are admitted to an LTCH.

However, regarding the commenters' statement that the proposed rule did not target cases that are likely the result of inappropriate admission and that data available to CMS clearly showed an immaterial number of LTCH claims denied as the result of QIO review of a sample of LTCH cases, we would share the results of an LTCH review from FY 2005. In that review, QIOs reviewed a statistically valid, representative national sample of 1,392 LTCH claims annually for the past few years. These samples were utilized for calculation of national payment error rates and the sampling method has been determined to be statistically sound by external audit. While the overall numbers of admission denials is low due to the sample size, statistically-based projections have revealed issues relative Start Printed Page 26947to inappropriate admissions, especially admissions with short length of stays. For discharges occurring during FY 2005, 7.9 percent of the admissions were found to be inappropriate accounting for a projected overpayment of $215,073,309 annually; this admission denial rate is higher than the 4.7 percent found for acute care hospitals paid under the IPPS during the same time period. Of note, 72.7 percent of admission denials for LTCH discharges occurred in claims with a LOS of 25 days or less.

The commenters further asserted that QIO data does not support our assumption that cases were inappropriately admitted to LTCHs as a result of LTCHs acting as extension sites or units of other acute care hospitals or patients receiving less than a full episode of care at the acute care hospital. However, an internal analysis of LOS for FY 2005 LTCH discharges has revealed that over 50 percent of stays were 25 days or less in length and many of those have an LOS comparable to an IPPS LOS for that DRG.

XI. MedPAC Recommendations: The RTI Contract

With the recommendations of MedPAC's June 2004 Report to Congress as a point of departure, RTI evaluated the feasibility of developing patient and facility level characteristics for LTCHs to identify and distinguish the role of these hospitals as a Medicare provider.

RTI completed this project in two phases. In Phase I, RTI prepared a background report summarizing existing information regarding LTCHs' current role in the Medicare system: their history as Medicare participating providers; the types of patients they treat; the criteria QIOs currently use to review appropriateness of care in these settings; and the types of regulations they face as Medicare participating providers. This work reviewed prior analyses of these issues and included discussions with MedPAC, other researchers, CMS, the QIOs, and the hospital associations.

In Phase II, RTI collected additional information on tools currently used by the QIOs and the industry to assess patient appropriateness for admission; analyzed claims to understand differences between hospital patients with outlier stays in non-LTCHs and those treated in LTCHs; and visited different types of hospitals to observe first-hand how LTCH patients differ from those in other settings and how this pattern varies in different parts of the country. RTI worked with different associations, including the National Association of Long Term Hospitals (NALTH), the Acute Long Term Hospital Association (ALTHA), the AHA, and the American Medical Peer Review Association (AMPRA), as well as several of the larger LTCH chains. The final report submitted by RTI summarizes these efforts and makes numerous recommendations to CMS regarding LTCHs.

As noted above, the reports on both Phase I and Phase II of RTI's research have been posted on our Web site at http://www.cms.hhs.gov/​LongTermCareHospitalPPS/​02a_​RTIReports.asp#TopOfPage. Please note that this report does not represent our position or policy. We are currently evaluating RTI's recommendations regarding the feasibility of developing patient and facility level criteria from several standpoints. Most significantly, we have been concerned that several of RTI's recommendations may require statutory changes. Furthermore, even among those recommendations for action that would be accomplished on a regulatory level, there are many significant issues that require further analysis. RTI is proceeding with Phase III of their project and as during Phases I and II, we have consistently encouraged meaningful contact between RTI and industry stakeholders throughout this research phase of the contract.

Comment: We received a comment from MedPAC that urged us to continue working towards the development of patient and facility criteria as the best way to determine appropriate LTCH patients particularly in light of the RTI report which included recommendations similar to those originally suggested by MedPAC in its June 1994 Report to Congress. The Commission noted that approaches other than criteria, such as the 25 percent rule, “may be administratively less complex but are more arbitrary and increase the risk for unintended consequences.” The Commission further suggested that we evaluate patient criteria currently in use by LTCHs and continue to work with LTCH associations that have developed criteria. The commenter also reiterated the Commission's support for severity-rated DRGs for use in the IPPS hospitals and noted that their adoption could reduce necessity for referrals to LTCHs. The Commission also endorsed a larger role for QIOs in the oversight of determinations of medical necessity, as well as in monitoring compliance with patient and facility level criteria.

Response: We thank the Commission for its thoughtful response to our proposed rule. We are mindful of the importance of identifying patient and facility-level criteria for LTCHs and believe that we have contracted with RTI to continue moving in that direction as they begin Phase 3 of their project. The reports on Phase I and Phase II of RTI's work are posted on the CMS Web site. We believe that their analyses of LTCHs and other provider categories that treat LTCH-type patients provide the foundation for any future development of patient level criteria.

We understand MedPAC's preference for patient criteria as opposed to payment adjustments for the purpose of determining appropriate patients for treatment at a LTCH. However, we would note that even with the development of patient criteria, it continues to be our statutory responsibility, under the BBA and BBRA to provide for appropriate adjustments and to establish regulations as may be necessary to effectively administer the Medicare program by way of implementing appropriate payment policies and payment adjustments. Therefore, even though we continue our work with RTI in Phase 3 of their project to see if we can identify appropriate patient and facility-level criteria for LTCHs, we do not see the development of those criteria as contradictory aspects to efforts we have undertaken while performing our fiduciary responsibility for the Medicare program. We further believe that it may be appropriate to continue to maintain such policies under the LTCH PPS that guard the Medicare Trust Fund from duplicative payments for what is one episode of patient care, even if we are able to develop and adopt facility and patient criteria for LTCHs and LTCH patients.

In the following comment and response, we discuss our evaluation of existing patient criteria currently in use by LTCHs, including one that was developed by one of the LTCH associations.

The Commission's support for the adoption of severity-rated DRGs for use in acute care hospitals paid for under the IPPS is discussed in the FY 2008 IPPS proposed rule. As discussed in that proposed rule, we have also proposed adopting the same severity-based DRGs for the LTCH PPS.

Finally, regarding an increasing role for QIOs in the LTCH PPS, we are currently developing the next Quality Improvement Organization Scope of Work. These comments will be considered in that process.

Comment: Many commenters took issue with the payment adjustments that we proposed in the RY 2007 LTCH PPS proposed rule that would revise the existing SSO policy and extend the Start Printed Page 26948scope of the 25 percent threshold payment adjustment. The commenters suggested that rather than issuing further regulations that do not reasonably address our most significant concerns with LTCHs, that we should instead focus on developing LTCH patient criteria as was suggested by MedPAC in 2004 and discussed in the RTI report. Several commenters further contended that we have been “ignoring MedPAC and RTI recommendations.” One commenter stated, “In 3 years, CMS has not implemented MedPAC recommendations.” Many commenters questioned why we have not adopted existing patient criteria instruments that are currently used by LTCHs, such as Interqual or the system developed by MassPRO and the National Association of Long Term Hospitals (NALTH).

Response: In responses to comments in the sections of this final rule that address the SSO policy and the extension of the 25 percent (or applicable percentage) threshold payment adjustment to LTCH and satellite discharges that were admitted from non-co-located hospitals, we specifically address our rationale for issuing both of these provisions. However, aside from objections to our policies, it also appears as if the commenters are combining the production of patient and facility level criteria by RTI with the end of further payment adjustments under the LTCH PPS by CMS. Notwithstanding the future development of appropriate patient and facility level criteria for LTCHs, it will continue to be our statutory responsibility under sections 1102 and 1871 of the Act to establish regulations as may be necessary to adjust LTCH payments appropriately and to effectively administer the Medicare program.

Furthermore, we strongly disagree with statements by the above commenters that we have “ignored” the MedPAC recommendations, as well as those recently resulting from RTI's final report. In awarding contracts, as a Federal Agency, we are required to follow the protocols of the Federal contracting process that are governed by the Office of Federal Procurement Policy (OFPP) and Health and Human Services Acquisition Regulation (HHSAR) (5 U.S.C. 301 and section 205(c) of the Federal Property and Administrative Services Act of 1949 as amended (40 U.S.C. 486(c)) and regulations as follows: The Federal Acquisition Regulation (48 CFR Ch. 1); FAR Supplements (48 CRFR Chs. 2-53); Labor (29 CFR, 41 CFR Ch. 50, Small Business Administration (SBA) 13 CFR, and OMB Circular No. A-130. Even after meeting all of the above requirements, however, we would note that while the MedPAC recommendations were originally published in June 2004, we were able to award the contract to RTI to evaluate MedPAC's recommendations by the start of FY 2005 (October 2004).

We have included an update of RTI's progress in each notice since that time, and we believe that an objective evaluation of the Phase I and II reports presently on the CMS Web site at http://www.cms.hhs.gov/​LongTermCareHospitalPPS/​02a_​RTIReports.asp#TopOfPage indicates steady progress but also demonstrates the thoughtful analysis resulting from RTI's high level of professionalism in pursuit of our goal.

RTI's work over the past 2.5 years has resulted in an extensive and careful analysis of the Medicare populations served by LTCHs, a comparison of these populations with those treated in other acute settings, including IPPS, IRFs, and Inpatient Psychiatric populations, as well as those treated in less intensive settings such as SNFs. This work included analysis of Medicare data to compare patient characteristics and provider costs for certain types of patients; regulatory requirements governing program conditions of participation for these different types of facilities; interviews with private sector developers of level of care determinations; and site visits and interviews with physicians and hospitals treating these typical and frequently overlapping populations.

The results suggested that, while there are distinctive populations with very long acute care needs, there are also many patients whose LOS at the LTCH may trigger a short stay outlier payment, suggesting their LOS was not consistent with an LTCH level of care need as defined by longer term acute level hospital care. While existing patient criteria such as Interqual are useful for distinguishing between the need for hospital-level treatment and a less intensive level, such as SNF care, RTI's analysis has determined that, in fact, the private sector criteria failed to distinguish between patients at LTCHs and patients at acute care hospitals. The criteria proposed by the National Association for Long Term Hospitals (NALTH) also had this shortcoming. While they identified the intensive acute care patient, they failed to identify differences between their admissions' clinical characteristics and those treated in a general acute care hospital step-down unit.

At a recent Technical Expert Panel (TEP) comprised of physicians, nurses, and hospital administrators representing, in addition to LTCHs, acute care hospitals, IRFs, and SNFs, convened by RTI, all participants agreed that LTCHs specialize in treating the types of patients they admit, noting that having a high volume of these patients is one of the reasons for their successful outcomes. However, it was also noted that these services are also provided in general acute care hospitals, particularly in ICU step-down units. So, while LTCHs may specialize in a select group of patients (the more intensively ill), they are not the only providers to successfully provide these treatments. The TEP reached consensus that volume was important for successful treatment of the complicated cases, regardless of site of care. TEP participants continue to be involved in providing feedback to RTI and another TEP is being planned based upon the earlier meeting and participant responses.

We continue to contract with RTI to work on these issues and RTI is presently involved into the next phase (phase III) of their project which will include the refinement of patient specific comparisons of total episode treatment in areas with and without LTCHs. Furthermore, RTI is also participating in the CMS-wide effort to better identify patient-level differences across the various levels of care.

XII. Payment for Direct Graduate Medical Education (GME)

A. GME Background

Section 1886(h) of the Act, as added by section 9202 of the Consolidated Omnibus Budget Reconciliation Act (COBRA) of 1985 (Pub. L. 99-272) and implemented in regulations at existing § 413.75 through § 413.83, establishes a methodology for determining payments to hospitals for the direct costs of approved graduate medical education (GME) programs. Section 1886(h)(2) of the Act, as added by COBRA, sets forth a payment methodology for direct GME costs involving the determination of a hospital-specific, base-period per resident amount (PRA) that is calculated by dividing a hospital's allowable costs of GME for a base period by its number of residents in the base period. The base period is, for most hospitals, the hospital's cost reporting period beginning in FY 1984 (that is, the period beginning October 1, 1983, through September 30, 1984). Generally, for cost reporting periods beginning on or after July 1, 1985, Medicare direct GME payments are calculated by multiplying the hospital's PRA by the weighted number of full-time equivalent (FTE) residents working in all areas of the Start Printed Page 26949hospital (and nonhospital sites, when applicable), and by the hospital's Medicare percentage of total inpatient days. In addition, as specified in section 1886(h)(2)(D)(ii) of the Act, for cost reporting periods beginning between October 1, 1993, through September 30, 1995, each hospital-specific PRA for the previous cost reporting period is not updated for inflation for any FTE residents who are not either a primary care or an obstetrics and gynecology resident. As a result, hospitals that trained primary care, and obstetrics and gynecology residents, as well as nonprimary care residents in FY 1994 or FY 1995 have two separate PRAs: one for primary care, and obstetrics and gynecology residents; and one for nonprimary care residents.

The Medicare, Medicaid, and SCHIP [State Children's Health Insurance Program] Balanced Budget Refinement Act of 1999 (Pub. L. 106-113) (BBRA) amended section 1886(h)(2) of the Act to establish a methodology for the use of a national average PRA in computing direct GME payments for cost reporting periods beginning on or after October 1, 2000, and on or before September 30, 2005. The BBRA established a “floor” for hospital-specific PRAs that is equal to 70 percent of the locality-adjusted national average PRA. In addition, the BBRA established a “ceiling” that limited the annual inflation update to a hospital-specific PRA if the hospital's PRA exceeded 140 percent of the locality-adjusted national average PRA. Section 511 of the Benefits Improvement and Protection Act of 2000 (Pub. L. 106-554) (BIPA) increased the floor established by the BBRA to equal 85 percent of the locality-adjusted national average PRA. For purposes of calculating direct GME payments, each hospital-specific PRA is compared to the floor and the ceiling to determine whether a hospital-specific PRA should be revised.

Section 1886(h)(4)(F) of the Act established limits on the number of allopathic and osteopathic residents that a hospital may count for purposes of calculating direct GME payments. For most hospitals, the limits are the number of allopathic and osteopathic FTE residents training in the hospital's most recent cost reporting period ending on or before December 31, 1996.

B. Residents Training in Nonhospital Settings

1. Background

For purposes of direct GME payments, since July 1, 1987, the statute allows hospitals to count the time residents spend training in sites that are not part of the hospital (referred to as “nonprovider” or “nonhospital sites”) under certain conditions. Section 1886(h)(4)(E) of the Act requires that the Secretary's rules concerning computation of FTE residents for purposes of direct GME payments “provide that only time spent in activities relating to patient care shall be counted and that all the time so spent by a resident under an approved medical residency training program shall be counted towards the determination of full-time equivalency, without regard to the setting in which the activities are performed, if the hospital incurs all, or substantially all, of the costs for the training program in that setting.” (Section 1886(h)(4)(E) of the Act, as added by section 9314 of the Omnibus Budget Reconciliation Act of 1986 (Pub. L. 99-509) (OBRA 86).) Regulations regarding the treatment of time spent by residents training in nonhospital sites for purposes of direct GME payments were first implemented in the September 29, 1989 final rule (54 FR 40286). In regulations adopted in that same rule at § 413.86(f)(3) (now § 413.78(c)), we stated that a hospital may count the time residents spend in nonprovider settings for purposes of direct GME payment if the residents spend their time in patient care activities and there is a written agreement between the hospital and the nonprovider entity stating that the hospital will incur all or substantially all of the costs of the program. The regulations at that time defined “all or substantially all” of the costs to include the residents' compensation for the time spent at the nonprovider setting. Before October 1, 1997, for IME payment purposes, hospitals were not permitted to count the time residents spent training in nonhospital settings. Section 4621(b)(2) of the BBA revised section 1886(d)(5)(B) of the Act to allow providers to count time residents spend training in nonprovider sites for IME purposes, effective for discharges occurring on or after October 1, 1997. Specifically, section 1886(d)(5)(B)(iv) of the Act was amended to provide that “all the time spent by an intern or resident in patient care activities under an approved medical residency program at an entity in a nonhospital setting shall be counted towards the determination of full-time equivalency if the hospital incurs all, or substantially all, of the costs for the training program in that setting.” In the July 31, 1998 final rule (63 FR 41004 through 41005) at § 412.105(f)(1)(ii)(C) and § 413.78(d) (formerly designated § 413.86(f)(4)), we specified the requirements a hospital must meet to include the time spent by residents training in a nonhospital site in its FTE count for portions of cost reporting periods occurring on or after January 1, 1999 for purposes of both direct GME and IME payments. Section 413.75(b) redefined “all or substantially all of the costs for the training program in the nonhospital setting” as the residents' salaries and fringe benefits (including travel and lodging where applicable), and the portion of the cost of teaching physicians' salaries and fringe benefits attributable to direct GME. Section 413.78(e) provides that, in order for a hospital to be permitted to count FTE residents training in a nonhospital setting, a written agreement must be in place between the hospital and the nonhospital site providing that the hospital will incur the costs of the resident's salary and fringe benefits while the resident is training in the nonhospital site. The hospital must also provide reasonable compensation to the nonhospital site for supervisory teaching activities, and the written agreement must specify that compensation amount.

2. Moratorium on Disallowances of Allopathic or Osteopathic Family Practice Residents Training Time in Nonhospital Settings, and Questions and Answers (Qs&As) on CMS Web Site (Section 713 of the MMA and § 413.78)

In order for the hospital to incur “all or substantially all” of the costs in accordance with the regulations, the actual cost of the time spent by teaching physicians in supervising residents in the nonhospital setting must be compensated by the hospital. The amount of supervisory GME costs is dependent upon the teaching physician's salary and the percentage of time that he or she devotes to activities related to the residency program at the nonhospital site. (We note that the teaching physician's involvement in the provision of patient care is not considered attributable to direct GME.) As long as there are supervisory GME costs associated with the nonhospital training, the hospital must reimburse the nonhospital setting for those costs to count FTE resident time spent in the nonhospital site for purposes of IME and direct GME payments.

Many hospitals have entered into written agreements with nonhospital sites that state that the teaching physician is “volunteering” his or her time in the nonhospital site, and, therefore, the hospital is not providing any compensation to the teaching physician. Other hospitals have paid only a nominal amount of compensation for the supervisory teaching physicians' Start Printed Page 26950time in the nonhospital setting. Because § 413.78(d) requires that the hospital must incur “all or substantially all” of the direct GME costs, including those costs associated with the teaching physician, regardless of whether the written agreement states that the teaching physician is “volunteering,” we have required that the hospital pay these costs to count FTE residents training in the nonhospital site, as long as these teaching physician costs exist.

Section 713 of the MMA imposed a 1-year moratorium relating to certain nonhospital site teaching physician costs for the period from January 1, 2004, through December 31, 2004. During this 1-year period, we were required to allow hospitals to count FTE allopathic or osteopathic family practice residents training in nonhospital settings for IME and direct GME payment purposes without regard to the financial arrangement between the hospital and the teaching physician practicing in the nonhospital setting to which the resident was assigned.

We instructed our contractors (formerly called “fiscal intermediaries” or “FIs”) regarding the effect of section 713 of the MMA in the One-Time Notification (OTN), “Changes to the FY 2004 Graduate Medical Education (GME) Payments as Required by the Medicare Modernization Act of 2003 (MMA)” (CR 3071, Transmittal 61, issued on March 12, 2004). Generally, we stated in the OTN that, when settling prior year cost reports during this 1-year period, or for family practice residents actually training in nonhospital settings during this 1-year period, contractors should allow hospitals to count allopathic and osteopathic family practice residents training in a nonhospital setting for direct GME and IME payment purposes without regard to the financial arrangement between the hospital and the nonhospital site pertaining to the teaching physicians' costs associated with the residency program. For further information on this provision and for a summary of comments and responses related to this provision, please refer to the FY 2005 IPPS final rule (69 FR 49176).

Furthermore, in response to questions and concerns raised by the industry and Medicare contractors as to how to determine the costs associated with residency training at the nonhospital setting, as well as how and when to pay the nonhospital setting for these costs, we posted Qs&As on the CMS Web site on April 8, 2005 at http://www.cms.hhs.gov/​AcuteInpatientPPS/​Downloads/​nonhospQA.pdf. In the Qs&As, in response to the question of whether there are situations where it is acceptable for the teaching physician to “volunteer” his or her time supervising residents at the nonhospital site, we stated that “* * *'the relevant question is not whether volunteerism is permissible, but whether there is a cost to the nonhospital site for supervising the resident training. If there is a cost, the hospital must reimburse the nonhospital site for those costs.” We further stated that we believe in situations where the teaching physician receives a predetermined compensation amount for his or her time at the nonhospital site that does not vary with the number of patients he or she treats, there is a cost for the teaching physician time spent in nonpatient care direct GME activities. In contrast, if the physician's compensation at the nonhospital site is based solely on his or her billings, there is no cost for teaching physician time spent in nonpatient care direct GME activities. Accordingly, the statute continues to require that a hospital must pay “all or substantially all” the costs of training residents at the nonhospital site to count FTE residents training at that site, including teaching physician costs, as long as those costs exist.

3. Requirements for Written Agreements for Residency Training in Nonhospital Settings (§ 413.78(e))

In implementing section 1886(h)(4)(E) of the Act, to assist contractors in determining whether a hospital incurred “all or substantially all” of the costs of the program in the nonhospital setting, we required in § 413.78(c) and (d) (formerly § 413.86(f)(3) and (4)) that there must be a written agreement between the hospital and the nonhospital site stating that the hospital will incur “all or substantially all” of the costs of training in the nonhospital setting. We later specified at § 413.78(d)(2) that the written agreement must indicate the amount of compensation provided by the hospital to the nonhospital site for supervisory teaching activities.

In an effort to respond to concerns expressed by hospitals about the administrative burden associated with meeting the written agreement requirements, in the FY 2005 IPPS final rule (69 FR 49179), at § 413.78(e), we revised our regulations to allow hospitals to choose to either enter into a written agreement with the nonhospital site before the hospital may begin to count residents training at the nonhospital site, or to pay concurrently for the cost of training at the nonhospital setting. That is, in the absence of a written agreement, hospitals are required to pay “all or substantially all” of the costs of the training program in the nonhospital setting by the end of the third month following the month in which the training occurs.

4. Modification of the Definition of “All or Substantially All of the Costs for the Training Program in the Nonhospital Setting”

We have met numerous times with industry representatives with the goal of developing a proposal which would respond to the concerns expressed by the teaching hospital community about the administrative burden associated with determining and documenting that hospitals are paying for “all or substantially all” of the costs for the training in the nonhospital setting. Some industry representatives recently suggested that we could ease administrative burdens by modifying the requirements hospitals must satisfy to meet the statutory requirement to incur “all or substantially all” of the costs by allowing a teaching physician to attest that at least 90 percent of the teaching physician's GME time is spent in patient care activities. However, we explained in response that the statutory test is tied to whether the hospital has incurred “all or substantially all” of the costs of the training at that site, not to how the teaching physician's GME time is spent. Therefore, we do not believe the attestation proposed by the industry adequately addresses the statutory requirement that the hospital incur “all or substantially all” of the costs of the training program at that site. We continue to believe that any Medicare policy approach to allowing hospitals to count FTE residents training in nonhospital settings for IME and direct GME payment purposes must be consistent with the statutory requirement that hospitals incur “all, or substantially all” of the costs of a training program in a nonhospital setting. The statute is clearly concerned about the cost to the nonhospital site, and we believe the statute has set a priority to move resources, in terms of both residents and funding, out into community settings. Therefore, where there is a cost to the nonhospital setting for training residents, we believe that the Medicare program is obligated to ensure that the nonhospital settings receive the funding they are entitled to receive from hospitals under the statute.

Accordingly, we continue to believe that our current definition of “all or substantially all” of the costs, which is based on the costs of the training program at the nonhospital site, is true to the intent of the statute. However, to address the industry's concerns related Start Printed Page 26951to burdensome documentation requirements, we are establishing an alternative methodology that hospitals may choose to use in determining and paying for the teaching physician costs attributable to direct GME in the nonhospital sites. As we explain below in this section, we are revising the current definition of “all or substantially all” of the costs to require hospitals to incur a percentage of the costs of the training program at the nonhospital site. This revision also generally incorporates the industry representatives' concept of a 90 percent threshold, but does not specifically relate it to the percentage of time spent by the teaching physician on nonpatient care direct GME activities, as suggested by industry representatives. Furthermore, as explained in more detail below in this section, in determining whether a hospital has met the 90 percent cost threshold, we are allowing hospitals to use certain shortcuts or proxies in the place of actual cost data specific to each teaching physician at each nonhospital site. However, hospitals would always still have the option of calculating the actual teaching physician costs and the 90 percent threshold using actual cost data specific to all, or some of their applicable teaching physicians. That is, even if a hospital chooses to calculate the direct GME costs of a program using actual teaching physician time and cost data (as under existing regulations) rather than using the proxies, under this revision, a hospital will only be required to pay at least 90 percent of the total of the residents' salaries and fringe benefits (including travel and lodging where applicable) and the portion of the teaching physicians' costs attributable to direct GME for a program at the nonhospital site. That is, a hospital would no longer be required to pay 100 percent of the residents' salaries and fringe benefits (including travel and lodging where applicable), plus the portion of the teaching physicians' costs attributable to direct GME at the nonhospital site. Instead, a hospital will be required to pay for 90 percent of the GME costs of a training program in a nonhospital site, and will have a choice between two approaches for calculating teaching physician's costs.

Currently, “all or substantially all of the costs for the training program in the nonhospital setting” is defined at § 413.75(b) as the residents' salaries and fringe benefits (including travel and lodging where applicable) and the portion of the cost of teaching physicians' salaries and fringe benefits attributable to direct GME. We are defining “all or substantially all of the costs for the training program in the nonhospital setting” under § 413.75(b) (prospectively for cost reporting periods beginning on or after July 1, 2007) to mean at least 90 percent of the total of the costs of the residents' salaries and fringe benefits (including travel and lodging where applicable) and the portion of the cost of teaching physicians' salaries attributable to direct GME. We believe this standard is consistent with the statute, in that hospitals would still be required to incur “all or substantially all” of the costs of training programs in nonhospital settings, and we would expect this standard to further encourage hospitals to shift training to nonhospital settings as intended by the statute. Under this revised definition of “all or substantially all” of the costs for the training program in the nonhospital setting, we will create a 90 percent threshold that hospitals must meet to count FTE resident time spent training at the nonhospital setting for IME and direct GME payment purposes. Additionally, under the new definition, hospitals will only have to incur a minimum of 90 percent of the costs of the program at a nonhospital site to count FTE resident time spent training at the site. Furthermore, as is the case with the current definition of “all or substantially all,” the new definition will not include overhead costs.

We solicited comments on our proposed effective date for purposes of both direct GME and IME as to whether our proposal should be effective immediately for portions of cost reporting periods occurring on or after July 1, 2007, or alternatively, for cost reporting periods beginning on or after July 1, 2007. Although an effective date of “portions of cost reporting periods occurring on or after July 1, 2007,” provides a more immediate response to concerns raised by teaching hospitals, we had concerns that establishing new policies in the middle of hospitals' cost reporting periods may present some logistical challenges, both from an implementation and an audit perspective. Therefore, we proposed that the new definition of “all or substantially all” of the costs would be effective for both direct GME and IME for cost reporting periods beginning on or after July 1, 2007.

As we explained, rather than adopt the industry's suggested standard of 90 percent of the teaching physicians' time spent in patient care activities, which we do not believe would be sufficiently true to the requirements of the statute, as a compromise, we would accept that hospitals have incurred “all or substantially all” of the costs of the program at the nonhospital site (and are therefore permitted to count the FTE residents training at the nonhospital site for IME and direct GME Medicare payment purposes) if the hospital incurs at least 90 percent of the costs of training at that site. Under this revised policy, a hospital would not have to demonstrate that it has incurred the costs of the teaching physician's time if it has otherwise incurred at least 90 percent of the nonhospital site training costs by paying the residents' salaries and fringe benefits (including travel and lodging where applicable) during the time spent training at the site. However, if the residents' salaries and fringe benefits (including travel and lodging where applicable) account for less than 90 percent of the costs of training at the nonhospital site, the hospital would have to compensate the nonhospital site for its teaching physician costs so that the hospital is incurring at least 90 percent of the training program costs at the nonhospital site. If the hospital does not meet the 90 percent threshold by only paying for the cost of the residents' salaries and fringe benefits (including travel and lodging where applicable), the hospital would have to meet the threshold by incurring some portion of the teaching physicians' salaries that is attributable to direct GME.

As previously stated in the Qs&As on the CMS Web site on April 8, 2005 at http://www.cms.hhs.gov/​AcuteInpatientPPS/​Downloads/​nonhospQA.pdf (Answer #4), we believe there are typically no costs for teaching physician time if the physician's compensation at the nonhospital site is based solely and directly on the number of patients treated and for which he or she bills, which is the case with a solo practitioner. When the solo practitioner is not treating patients, he or she is not receiving payment for any other duties at the nonhospital site. Therefore, in this instance, there is no cost to the nonhospital site for the teaching physician's time. Thus the hospital has to incur only 90 percent of intern and resident salaries to meet the new regulatory requirements. However, in the case of a group practice or clinic setting, the physician often receives a predetermined payment amount, such as a salary, for his or her work at the nonhospital site. This predetermined payment amount reflects all of his or her responsibilities at the nonhospital site, including treating patients, training residents, and other administrative activities (as applicable), and he or she may receive that predetermined Start Printed Page 26952payment from the nonhospital site regardless of how many patients he or she actually treats. The predetermined amount implicitly also compensates the physician for supervising residents. A portion of this implicit compensation is the cost attributable to teaching activities. Under current regulations, in order to count the residents training at that site, the hospital must pay the nonhospital site this amount. However, there may be instances in a group practice, where a teaching physician is not receiving a form of predetermined compensation for his or her work at the nonhospital site. For example, several physicians may work in the same office and share overhead expenses such as electricity and rent, but there is no sharing of revenues from patient care activities. Rather, the physicians operate as solo practitioners and are not compensated according to some predetermined arrangement. In cases such as these, we assume that the teaching physician is functioning as a solo practitioner and that teaching physician costs for GME training at the nonhospital site are zero. Accordingly, the revised policy being adopted in this final rule would more likely be applicable to members of group practices (or physicians in other arrangements) where the teaching physician receives a salary or other form of predetermined compensation for his or her work at the nonhospital site. However, we note that under the revised policy, in the case of solo practitioners, hospitals must continue to pay for at least 90 percent of the total cost of the residents' salaries and fringe benefits, including travel and lodging where applicable.

Comment: We received several comments noting the commenters' appreciation of the efforts CMS has devoted towards the issue of residency training at nonhospital sites and the belief that the proposed rule is a good first step in further improving the regulations regarding residency training at nonhospital sites. The commenters believe that by not requiring hospitals to pay for 100 percent of the costs of training at the nonhospital site and by allowing the use of proxies, the proposed rule may provide for considerable administrative relief.

Response: We appreciate the commenters' support of the proposed rule. We agree with the commenters and believe that the final rule will provide significant administrative relief and support the training of residents at nonhospital sites.

Comment: Several commenters maintained that the FY 1998 IPPS final rule (63 FR 40986 July 31, 1998), as well as a program transmittal A-98-44 from December 1998 stated that whatever reasonable amount was agreed upon by the nonhospital site and the hospital, that amount would be accepted as reflecting the costs of the nonhospital site.

Response: Although some may have read our previous guidance to suggest that the amount of payment for teaching physician costs in the nonhospital setting could be decided based solely upon negotiations between the hospital and nonhospital site that has not been our policy. As we indicated in the Qs&As posted on the CMS Web site on April 8, 2005 at http://www.cms.hhs.gov/​AcuteInpatientPPS/​Downloads/​nonhospQA.pdf, to the extent that there is a cost associated with teaching physicians for the residency training program at the nonhospital site, according to statute and regulations, the hospital must pay “all or substantially all” of the cost.

Comment: Several commenters requested a return to the definition of “all or substantially all” that was in place prior to 1999, which did not include costs associated with teaching physicians in the nonhospital site. One commenter specifically stated that reversing the unintended consequences of the previous definition change was difficult and, likewise, “Once in place, the costs of reversing this new rule and definition would be similarly difficult.”

Response: As explained earlier, we believe that our current definition of “all or substantially all of the costs for the training program in the nonhospital setting,” which includes the GME portion of the teaching physicians' salary, is most consistent with the statutory language and legislative intent. Therefore, we are not returning to the pre-1999 definition of that term.

Comment: We received many comments regarding the effective date for our proposed policy revision. Some commenters believe that the policy revision should be effective for portions of cost reporting periods occurring on or after July 1, 2007 while others believe that the policy revision should be effective for cost reporting periods beginning on or after July 1, 2007. One commenter asked that hospitals be able to apply the new method to any years where residents were disallowed. Other commenters requested that the proposed policy revision be effective retroactively to previous cost reporting periods.

Response: We solicited comments concerning the effective date of the proposed policy revisions. After carefully considering these comments, we have decided to finalize this policy revision to be effective for cost reporting periods beginning on or after July 1, 2007. As we stated in the proposed rule, we are concerned that establishing new policies in the middle of hospitals' cost reporting periods would present burdensome technical and administrative difficulties, both from an implementation and an audit perspective. In addition, we do not believe that we have the authority to follow the commenters' suggestions to implement this provision retroactively. Section 1871(e)(1)(A) of the Act generally prohibits the Secretary from making retroactive substantive changes in policy unless retroactive application of the change is necessary to comply with statutory requirements, or failure to apply the change retroactively would be contrary to the public interest. Only in very rare cases do we apply a rule retroactively (for example, in the wake of Hurricanes Katrina and Rita in 2005 where a retroactive change was clearly in the public interest). In those instances, we believed that the failure to apply regulatory changes retroactively would be contrary to the public interest because hospitals affected by the hurricanes could otherwise face dramatic financial hardship, which would threaten the stability of GME programs in the emergency area. In contrast, we do not believe that there is a compelling argument that demonstrates a degree of public interest that would justify applying this proposed policy revision retroactively.

Comment: Several commenters stated that they do not believe the proposed policy revision actually addresses the real concern that the hospital industry has with our current policy. These commenters believe the central issue is supervisory physician volunteerism in nonhospital settings. The commenters stated that volunteerism is historically endemic to physician education, and therefore, hospitals should not need to pay the costs of the supervisory physician when a physician is willing to volunteer as a supervisor. One commenter stated, “We urge CMS in the final rule to issue a clear policy statement that volunteer status of faculty will be determined by the hospital and nonhospital site and that even physicians in group practices who are compensated a predetermined amount not based on patient billings may still be volunteering their teaching services.” The commenter further stated that there is no cost for supervising residents in group practices since the physicians are making the same amount per year regardless of whether or not the teaching physicians are supervising residents. Some commenters believe Start Printed Page 26953that since physicians are “exempt” from wage and hourly rules under labor law, there is no reason why the physician and the physician's employer could not agree that the physician's teaching responsibilities are undertaken voluntarily by the physician, do not lessen the physician's duties to the employer, and involve time besides the time that is necessary for the physician to meet fully his or her responsibilities to the employer. The commenters noted that the rules applicable to Federal government employers recognize that volunteer time, even in the course of usual business hours, is not compensated by the Federal government (http://www.opm.gov/​oca/​leave/​html/​volunteer2.asp).

Response: According to the statute, a hospital is required to incur “all or substantially all” of the costs for a training program at the nonhospital setting in order to count the FTE residents training in the nonhospital setting for GME payment. There is no reference in statute to other labor laws that might apply to physicians. Accordingly, our proposal only addresses the issue of determining costs of training programs in nonhospital settings. With regards to supervisory physician time, we address the issue of the costs to the nonhospital site for supervising the resident training. Our policy has been that if there is a cost, the hospital must reimburse the nonhospital site for those costs. If there are no costs, then no payment for supervisory physician time is required. Typically, there is a cost for teaching physician time. For example, there is a cost to the nonhospital site when the physician receives a predetermined compensation amount for his or her time at the nonhospital site that does not vary with the number of patients he or she treats. In contrast, there is typically no cost for teaching physician time if the physician's compensation at the nonhospital site is based solely and directly on the number of patients treated and for which he or she bills. The most obvious example of this situation would be a solo practitioner that serves at a nonhospital site. We note that the hospital is required to compensate the nonhospital site for the costs of the teaching physicians' time spent in activities in connection with an approved residency training program other than the supervision of residents while furnishing billable patient care services. That is, only the costs associated with teaching time spent in activities within the scope of the GME program, but not in billable patient care activities, would be considered direct GME costs that would need to be incurred by the hospital.

Comment: Generally, commenters were pleased that CMS is moving away from the requirement that hospitals need to pay 100 percent of the costs of training at nonhospital sites in order to comply with the statutory mandate of incurring “all or substantially all” of the costs. However, many commenters feel that the threshold for “all or substantially all” should be further reduced beyond 90 percent. Commenters stated that the threshold should be reduced to 75 percent in accordance with our interpretation of “substantially all” under the “Stark” provisions. One commenter stated that in addressing the “Stark” provisions, “CMS requires ‘substantially all of the patient care services of the physicians who are members of a group (that is, at least 75 percent of the total patient care services of the group practice members) must be furnished through the group * * *’ ” In reference to whether these provisions conflict with the requirements under Stark, one commenter asked CMS to “Please confirm in your commentary that a reasonable attempt to comply with the requirements to pay for the costs at nonhospital sites, whether it be under the written agreement standard or under the concurrent payment standard, using proxies or real costs, is considered by CMS to be in compliance with Stark law.” The commenter further stated that if the action taken in the aforementioned sentence is not in full compliance with Stark law, CMS should make an exception under Stark for payments to nonhospital sites where the payments are made to referring physicians. Another commenter stated that “* * * none of the key organizations involved in this issue have recommended such a [90 percent] standard. To be fair, the community did raise the question of preceptors attesting to 90 percent of their time being spent with residents in patient care * * * but we are unaware of any stakeholder group that has recommended ‘substantially all’ be defined as 90 percent of costs in the nonhospital setting.” Other commenters requested that the threshold should be reduced to 75 percent because, as one commenter stated, “Courts have also defined ‘substantially all’ as being 75 percent or greater in the context of corporate and securities law.” Another commenter requested that the threshold be reduced to 60 or 70 percent because such a number would provide for increased flexibility at the local level, while another commenter believed that a threshold of 70 percent was more appropriate because it was more reflective of the reimbursement amounts hospitals receive from the government. A request was also made that the threshold be reduced to 80 or 85 percent.

Response: The statute requires hospitals to pay for “all or substantially all,” not just “substantially all,” of the cost of the training program in the nonhospital setting. We believe that in using the term “all or substantially all,” Congress’ intention was that hospitals pay close to 100 percent of the nonhospital site GME training program costs (otherwise the “all” would add no meaning). As we described in the proposed rule, prior to proposing to revise the definition of “all or substantially all” to mean at least 90 percent of the total of the costs of the residents' salaries and fringe benefits (and travel and lodging if applicable) and supervisory teaching costs associated with direct GME, we had received a suggestion from industry representatives that hospitals should be considered by CMS to meet the statutory mandate to pay “all or substantially all” of the costs if the teaching physician can attest that he or she is spending at least 90 percent of his or her GME time in nonpatient care direct GME activities at the nonhospital site. Since the issue is the cost associated with that teaching time, we did not agree with this suggestion. However, we continue to believe that a standard of 90 percent of the total costs is an appropriate interpretation of “all or substantially all.” In response to whether a reasonable attempt to comply with the regulations for residency training at nonhospital sites is considered to be in compliance with the Stark law, we believe that provided that the rate paid to the supervising physician is fair market value for the supervisory duties, the arrangement should not be inconsistent with the Stark law. Since both the use of proxies and actual data would be consistent with fair market value, we believe that this final policy conforms with the Stark law.

Comment: One commenter believes that we clearly stated in the proposed regulations at § 413.75(b)(2), § 413.78(f)(2) and § 413.78(f)(3)(ii) that a hospital only has to incur 90 percent of teaching costs. The commenter also believes that, although not restated in proposed regulations, the 90 percent threshold also applies to the requirements in § 413.78(f)(3)(i).

Response: We agree with the commenter that the 90 percent threshold also applies to § 413.78(f)(3)(i). Start Printed Page 26954

Comment: Many commenters stated that members of a group practice should be able to attest that they are volunteering and be viewed in the same manner as CMS views solo practitioners. Commenters also stated that it is more common for residents to train in group practice settings than with solo practitioners. One commenter stated business agreements vary among group practices and that, “Compensation is based on patient volume and, in effect, each physician is a solo practitioner.” Another commenter stated that for its specific nonhospital site, there is no additional payment made to a physician who teaches, nor is salary removed from a physician who does not teach. One commenter stated that although the commenter believes the proposed rule should not apply to solo practitioners, the commenter also believes that our logic is incorrect in determining why there are typically no teaching physician costs associated with solo practitioners and group practitioners that function as solo practitioners. The commenter stated, “The fact that the physicians’ compensation is derived solely from patient care revenues is not definitive in and of itself. Rather it demonstrates that the physician received no compensation for supervisory activities.” The commenter further noted that, “At a minimum, group practices should be permitted to rebut the ‘implicit’ compensation presumption by demonstrating that no portion of physicians' salaries is linked to resident supervision.” Another commenter stated that teaching hospitals and nonhospital sites are in the best position to determine if there are any costs for training residents at the nonhospital site, and if so, how the costs should be compensated. The commenter stated that residents gain clinical experience while training at nonhospital sites. Therefore, the costs associated with their training are de minimus and if the group practice decides collectively that it is volunteering as a practice, it should be able to do so.

Response: As we have previously stated in the April 8, 2005 Qs&As and in the RY 2008 LTCH PPS proposed rule “* * * the relevant question is not whether volunteerism is permissible, but whether there is a cost to the nonhospital site for supervising the resident training. If there is a cost, the hospital must reimburse the nonhospital site for those costs.” Therefore, if a teaching physician in a group practice is receiving a predetermined salary for his or her activities, and included in his or her activities are supervisory GME activities at a nonhospital site, then there is a cost associated with those activities. If teaching physicians that are members of a group practice can document that their circumstances are similar to solo practitioners in that they receive no predetermined salary and receive income solely from the patients they treat and the services for which they bill, the hospital may supply this documentation to the Medicare contractor during audit.

5. Implementation of a 90 Percent Cost Threshold

In revising the definition of “all or substantially all” of the costs of the program at a nonhospital site, and in establishing a 90 percent threshold, there are several variables that are important in the methodology for determining the minimum amount of training program costs that a hospital must pay in order to count FTE residents training in a nonhospital site. These variables are: teaching physicians' salaries, residents' salaries and fringe benefits (including travel and lodging where applicable), the number of hours per week that the teaching physician spends in direct GME (not billable patient care) activities in the nonhospital site, and the number of hours that a nonhospital site is open each week. To provide the reader with a context for the new methodology, we will first explain the methodology briefly, provide two examples, and then proceed to an in-depth discussion of each variable (see section XII.B.5.b. of the preamble of this final rule).

a. Methodology

One of the primary complaints voiced by the hospital industry over the past several years is that our policy requiring hospitals to determine the portion of the teaching physician cost attributable to direct GME in the nonhospital site results in an untenable documentation burden since many physicians are reluctant to disclose their salary information to the hospitals. One solution to this problem suggested by the hospital industry is to use national average physician salary information as a proxy for teaching physician-specific salaries in the determination of the total cost of the program at a nonhospital site. In addition, since the cost of the teaching physician time that the hospital must incur is based on the amount of time the teaching physician spends in nonpatient care direct GME activities, the hospital industry has been concerned that determining this GME time could require burdensome time studies. Therefore, we are adopting an alternative methodology that hospitals may choose to use, instead of actual costs, to calculate teaching physician costs in nonhospital sites. Using this alternative methodology, to facilitate a less burdensome way for a hospital to calculate the teaching physician costs associated with GME training at the nonhospital site, we are allowing hospitals to use 3 hours per week as a presumptive standard number of hours that a teaching physician spends in nonpatient care direct GME activities at a particular nonhospital site. To determine the percentage of the average salary associated with the 3 hours the teaching physician is presumed to spend in nonpatient care direct GME activities, a hospital would divide 3 hours by the number of hours the nonhospital site is open each week. Next, the hospital would multiply this percentage of time spent in nonpatient care direct GME activities by the national average salary of that teaching physician's specialty to calculate the cost of the teaching physician's direct GME time. The cost of the teaching physician's direct GME time would then be added to the costs of the salaries and fringe benefits (including travel and lodging expenses, where applicable) of the FTE resident(s) rotating in that program to that nonhospital site to determine the GME costs for that program at that site. (If FTE resident(s) are not rotating to a particular nonhospital site throughout a whole year, then the national average salary of the teaching physician would be prorated accordingly. The cost of the residents' salaries and fringe benefits (including travel and lodging where applicable) would already be reflective of an FTE count). The hospital must pay at least 90 percent of these total GME costs for the program at that nonhospital site to count the resident(s) training there for direct GME and IME purposes. If the hospital is already paying all, or even a portion of the residents' salaries and fringe benefits (including travel and lodging where applicable), and if the amount that the hospital is paying for the residents' salaries and fringe benefits (including travel and lodging where applicable) is equal to at least 90 percent of the GME costs at the nonhospital site (that is, the 90 percent threshold), then the hospital would be considered to be incurring “all or substantially all” of the costs, and need not incur an additional amount for teaching physician compensation to be permitted to include the FTE residents training in the nonhospital site in its FTE count for purposes of direct GME and IME payments. However, if the costs of the residents' salaries and fringe Start Printed Page 26955benefits (including travel and lodging where applicable) does not equal at least 90 percent of the GME costs of the training program at the nonhospital site, then the hospital must incur an additional amount for teaching physician costs based on the national average salary information until it is incurring at least 90 percent of the GME costs for that nonhospital site program. That is, under the alternative definition of “all or substantially all” of the costs, a hospital is required to incur at least 90 percent of the total GME costs for a particular program at a particular nonhospital site. The GME costs of a particular program at a particular nonhospital site consist of FTE residents' salaries and fringe benefits (including travel and lodging costs where applicable), and the portion of teaching physician compensation (which may be based on national average survey data) attributable to direct GME. As will be explained in more detail below in this section, the hospital always has the option of documenting the actual teaching physician's cost using actual time or salary information to pay at least 90 percent of the total of the costs of the program at the nonhospital site. In summary, the formula for determining the 90 percent threshold, or the minimum amount that a hospital must pay for the GME costs of a particular program at a particular nonhospital site is:

0.90 × [(sum of each FTE resident's salary + fringe benefits (including travel and lodging where applicable)) plus the portion of the teaching physician's compensation attributable to nonpatient care direct GME activities.]

The portion of the teaching physician's compensation attributable to nonpatient care direct GME activities may be calculated as follows:

(3/number of hours nonhospital site is open per week) × (national average salary for each teaching physician*)

* The number of teaching physicians included in this formula is subject to a 1:1 resident to teaching physician limit, as explained below in this section.

The following are two examples of the alternative methodology:

End Further Info

Example 1:

Assume one teaching physician is supervising one FTE resident in a nonhospital site for one residency year. The national average published salary amount for that teaching physician's specialty is $120,000, and he works in a clinic that is open 60 hours per week. Using the standard of 3 hours spent in nonpatient care direct GME activities per week, the teaching physician spends 5 percent of his time in GME activities (that is, 3/60 = 0.05 or 5 percent). To determine the cost of the teaching physician's time, the hospital may make the following calculation: $120,000 × 0.05 = $6,000. This teaching physician's cost is added to the resident's salary and fringe benefits to calculate the cost of the training at the nonhospital site in the following manner: $6,000 [cost of one teaching physician] + $60,000 [actual cost of the FTE residents' salary & fringe benefits] = $66,000. To meet the new definition of “all or substantially all,” the hospital would be required to pay at least 90 percent of the costs of the training program at the nonhospital site, which in this example equals $59,400 (that is, 0.90 × $66,000). Since in this case the cost of one FTE resident's salary and fringe benefits is $60,000, the hospital could reach the 90 percent cost threshold by simply incurring the resident's salary and fringe benefits during training at the nonhospital site.

Example 2:

Assume one teaching physician is supervising one FTE resident in a nonhospital site for an entire residency year. The national average published salary amount for that teaching physician's specialty is $200,000, and she works in a clinic that is open 40 hours per week. Using the standard of 3 hours spent in nonpatient care direct GME activities per week, the teaching physician spends 7.5 percent of her time in GME activities (that is, 3/40 = 0.075 or 7.5 percent). To determine the cost of the teaching physician's time, the hospital may make the following calculation: $200,000 × 0.075 = $15,000. This teaching physician's cost is added to the resident's salary and fringe benefits to calculate the cost of the training at the nonhospital site in the following manner: $15,000 [cost of one teaching physician] + $60,000 [actual cost of the FTE residents' salary and fringe benefits] = $75,000. To meet the new definition of “all or substantially all,” the hospital would be required to incur at least 90 percent of the costs of the training at the nonhospital site, which in this example equals $67,500 (that is, 0.90 × $75,000). Since in this case the cost of one FTE resident's salary and fringe benefits is $60,000, the hospital has not met the 90 percent threshold by only incurring the resident's salary and fringe benefits. The hospital would have to incur at least an additional $7,500 of the cost (that is, $67,500 − $60,000) to reach the 90 percent threshold to be permitted to count the FTE resident for IME and direct GME purposes. Alternatively, the hospital could document the actual teaching physician cost using time or salary information specific to that teaching physician at that site, and use that amount to calculate 90 percent of the actual training program costs.

b. Explanation of Variables

In the following section, we discuss each variable in the methodology for determining the cost that a hospital must incur to count FTE residents training in nonhospital sites, and explain our rationale for employing each of these variables. As stated previously, the variables are: teaching physicians' salaries; residents' salaries and fringe benefits (including travel and lodging where applicable); the number of hours per week that the teaching physician spends in nonpatient care GME activities in a nonhospital site; and the number of hours that a nonhospital site is open each week.

(1) National Average Physician Salary Data by Specialty

One of the foremost objections voiced by the hospital industry to our current policy is the documentation burden associated with requesting salary information from individual teaching physicians in nonhospital sites. Hospitals believe that many teaching physicians in nonhospital sites are reluctant to disclose their personal salary information, yet this disclosure is necessary to enable the hospital to determine and pay the nonhospital site for the actual costs of the GME program in accordance with our current regulations. One suggestion mentioned by the hospital industry as an alternative to obtaining individual teaching physician-specific salary information is to allow hospitals to use national average salary survey data by specialty. We understand that there are a number of organizations that conduct annual national surveys on physician compensation. We proposed to allow hospitals to use physician compensation survey data as a proxy to determine the teaching physician costs associated with GME in a program at a particular nonhospital site. For example, one such national organization that collects data on physician compensation that we are considering using is the American Medical Group Association (AMGA). AMGA's 2006 Medical Group Compensation and Financial Survey was performed under contract by RSM McGladrey. Founded in 1950, AMGA (formerly the American Association of Medical Clinics) is a trade association which dedicates itself to making the “* * * multi-specialty medical group model the preferred delivery system for patient-centered, affordable, quality medical care in America,” and represents 283 medical groups that include an average of 272 physicians. AMGA's use of the term “medical group” is based on the American Medical Association's definition of “group practice,” which is defined as a group that “includes the provision of health care services by three or more physicians who are formally organized as a legal entity governed by physicians in which business, clinical, and administrative facilities, records and personnel are shared and the practice goals, objectives, and values are commonly defined. Income from medical services provided by the group Start Printed Page 26956is treated as receipts of the group and is distributed according to some prearranged plan.” AMGA has been performing surveys like the 2006 Medical Group Compensation and Financial Survey since 1986. The 2006 survey was sent to over 2,600 medical groups, including medical groups that are not members of AMGA. To give readers an idea of the average compensation amounts in the survey, we have randomly selected 10 specialties included in the 2006 survey and listed their compensation information in Table 8. If we adopt the AMGA survey for use to determine the cost of teaching physicians' time attributable to GME, we would make the salary information for all specialties accessible to hospitals on our Web site and would provide it in a manner similar to Table 8.

Table 8.—Physician Salary Information

*SpecialtyMean salary (in dollars)Median salary (in dollars)
Cardiology411,916363,081
Dermatology336,531306,935
Family Medicine187,891178,366
Gynecology and Obstetrics286,418271,273
Internal Medicine192,264183,840
Ophthalmology307,044281,112
Pediatrics & Adolescent: General191,122182,186
Physical Medicine and Rehabilitation208,442207,004
Diagnostic Radiology: Non-Interventional415,521400,000
General Surgery331,970310,736
* This information was obtained from the 2006 Medical Group Compensation and Financial Survey published by the American Medical Group Association® (AMGA). For further information, visit AMGA's Web site at http://www.amga.org/​.

We solicited comments as to whether we should use the mean or median compensation amounts for purposes of determining the teaching physicians' cost. In addition, although we recognize that there are generally geographic variations in salary amounts within each specialty (and, although not included in Table 8, AMGA does provide some detail of salaries by geographic area), we proposed to use the single national average or median salary amount for each specialty, rather than consider geographic variations, because we want to simplify and streamline the methodology for determining the GME costs in nonhospital sites as much as possible. We also solicited comments about whether AMGA's salary information should be used, and if not, which other physician compensation survey (or possible mix of surveys) would be more appropriate for this purpose, and whether we should consider additional factors such as geographic variation in physician salaries within each specialty. We noted that we believe it is important for the organization providing specialty-specific physician compensation information for this purpose to be one that is nationally recognized as an authoritative source. Additionally, we believe the data should contain compensation amounts for the fullest range possible of specialties and subspecialties, and should be issued annually so that hospitals will always have the most current data to use in determining the teaching physician costs in nonhospital sites. In addition, we would prefer a survey that is available to the public at no cost. (We understand that a number of these surveys are proprietary.) In addition, we solicited comments as to how to make the survey data available in the most efficient possible manner.

Regardless of the survey source that we ultimately use, we proposed that hospitals would use the most recent survey data available as of the beginning of the hospital's particular cost reporting year. For example—

  • If residents are rotating to a particular nonhospital site to receive training in family practice in a hospital's cost reporting year beginning January 1, 2008, then the hospital would use the family practice average salary from the most recently issued survey (in the case of AMGA, 2007) as the salary cost of that teaching physician, even though that teaching physician may in fact earn more or less than that national average salary amount.
  • If the teaching physician is a neurologist providing residents with neurology training in a nonhospital site in a hospital's cost reporting year beginning July 1, 2007, then the hospital would use the neurology average salary from the most recently issued survey (in the case of AMGA, 2006, since AMGA's surveys are typically released in August) as the salary cost of that teaching physician.

Comment: Numerous commenters suggested that in determining the proxy amount for teaching physician supervisory costs, hospitals should be able to use CMS's reasonable compensation equivalents (RCEs). One commenter, specifically stated “The RCEs have been relied upon by CMS and its predecessor, the Health Care Financing Administration, for nearly 24 years as its measure of the reasonableness of physician compensation and, thus, those amounts should be used in this regulation as well.” Furthermore, many commenters stated that if we choose to use AMGA data as its teaching physician salary proxy source, we would be requiring the use of data with values that “substantially exceed” what it considers to be reasonable under the RCEs. Some commenters view use of AMGA data, which produces physician salary amounts which are higher than RCEs as being “arbitrary and capricious.” Several commenters stated that if we choose not to use RCEs, we should use data from the AAMC's Faculty Salary Survey, which has an excellent response rate, can be made accessible to the public, and includes a “broad range of specialties” and as reported by one commenter, the AAMC's 2005-2006 survey report “* * * includes data provided by all 125 accredited allopathic medical schools in the United States.”

In addressing whether hospitals should be able to use mean or median physician salary amounts in determining the proxy for teaching physician supervisory costs, several commenters requested that median salaries be used since medians are not affected by outlier data. Another commenter stated that since the salary amounts in AMGA's survey are not adjusted by the geographic area wage index, median physician salary amounts Start Printed Page 26957should be used. One commenter stated that mean salary amounts should be used because using the mean salary would account for both range and frequency, while using the median would only account for frequency. Another commenter stated that for situations in which there is no salary information available for a certain subspecialty, we should consult with the AMA or AOA and encourage national data survey groups to start tracking data for these subspecialties.

Some commenters suggested that when available, hospitals should be able to use physician salary data that accounts for geographic variations including variations between rural and urban areas, while other commenters were opposed to using data that accounted for geographic adjustments because of the potential for added complexity. One commenter stated that hospitals should be allowed “* * * to use a comprehensive source of locality adjusted physician compensation information as a proxy for actual compensation in determining non-hospital training costs.” Another commenter stated that if we do not allow hospitals to account for geographic variations, we would be requiring that hospitals rely on national salary data which is inaccurate and make it necessary for hospitals to collect their own hospital-specific data. One commenter stated that since the goal of proxies was to simplify the process, there should not be more than one national salary amount for each specialty. Another commenter stated, that within specialties, the commenter “* * * has not identified significant regional variations, and any large variation that might exist would be accounted for by simply using the median.” Lastly, a commenter stated that in states such as Utah, using a national salary proxy amount would not account for the fact that physicians' wages are lower than in other parts of the country and, therefore, if Utah used the national salary proxy it would be paying more than 90 percent of the total costs of training residents at the nonhospital site.

Response: In the RY 2008 LTCH PPS proposed rule, we solicited comments on what specific survey should be used as a proxy source in determining supervisory teaching physician costs. We also requested comments on whether we should consider geographic adjustments and whether we should use a mean or median salary amount. We appreciate the commenters' suggestions regarding what survey data should be used and whether we should use data adjusted for geographic variations, or use the mean or median salary point as the proxy for physician salary amounts.

In response to the commenters' suggestions that the proxy not be based on the AMGA data but rather be based on salary data used to establish Medicare's reasonable compensation equivalent (RCE) limits, we disagree with the commenters that the RCE limits would be an appropriate measure in the context of nonhospital site GME training programs. Although RCEs are appropriate as they are currently used in conjunction with other Medicare payment policies, we do not believe they are appropriate for use in determining a proxy for supervisory teaching physician costs in nonhospital sites. Currently, RCEs are only applied in the determination of reasonable costs of physician compensation in the few remaining types of facilities paid on a reasonable cost basis, the vast majority of which are not teaching hospitals. RCEs are not applied to the costs of any physician compensation in teaching hospitals that are paid under the IPPS. Thus, we do not believe RCE limits would represent an appropriate proxy to account for supervisory GME teaching physician costs in nonhospital settings. In addition, we note that under the RCE limits, exceptions are made for providers, such as small or rural hospitals, that may have difficulty recruiting or retaining physicians at the prescribed RCE level. As stated in the August 1, 2003 Federal Register (68 FR 45459) “* * * if a provider is able to demonstrate to the intermediary its inability to recruit or maintain physicians at a compensation level allowable under the RCE limits * * * the intermediary may grant an exception to the RCE limits established under these rules.” Since it may be difficult to recruit and retain physicians in rural nonhospital sites, we believe the use of RCEs as a proxy for the cost of teaching physician time in rural nonhospital sites could underestimate those costs since they are generally lower than market levels, or the AMGA salary amounts.

The updated RCEs published in the August 1, 2003 Federal Register (68 FR 45459), only include nine specialties. We do not believe the RCEs would provide the best representation of specialties for purposes of establishing proxies for supervisory teaching physician costs in nonhospital settings. In the August 1, 2003 Federal Register, we also stated, “If no specialty category is appropriate (for example, in determining the reasonable cost for an emergency room physician), the intermediary will use the reasonable compensation equivalent level for the ‘Total’ category, which is based on income data for all physicians” (68 FR 45459). The goal in using the physician salary proxy to determine supervisory teaching physician costs, for purposes of determining whether a hospital has met the statutory requirement to pay “all or substantially all” of the costs of the training at the nonhospital site, is to allow the hospital to use a figure that reflects the physician's actual salary without having the administrative burden of determining the physician's actual salary. Since the RCEs only exist for nine physician specialties, it would be frequently necessary to use the “Total” category when salary information for a specific specialty is not available. This would be contrary to our goal of using a proxy which reflects the actual amount. For the reasons cited above in this section, we do not believe RCEs are the most appropriate source of physician salary data to use in the context of policies regarding supervisory teaching physician salaries in nonhospital settings; and therefore, we will not use them as proxies for supervisory teaching physician costs.

In response to the request that we use the AAMC's Faculty Salary Survey to establish proxies for supervisory teaching physician costs, we question the appropriateness of using the AAMC's data in the determination of a proxy since we note that several salary amounts in the AAMC data are close in value to that of the RCE amounts which, as we explained earlier, may not fully reflect total physician compensation amounts. As we explained above, we believe AMGA's survey data are extremely comprehensive and by making the necessary information available on our Web site, AMGA data would be easily accessible to the public. Therefore, we are finalizing our policy to use survey data published by AMGA as a proxy for physician compensation in nonhospital settings, and thus, in determining supervisory teaching physician costs. However, we will continue to monitor the various survey options and consider whether other data sources are appropriate for this purpose.

Since some members of the teaching hospital community have claimed that collection of actual data is burdensome, we are seeking, through the use of proxies, to make the calculation of supervisory teaching physician costs for GME training at the nonhospital site as straightforward as possible. Therefore, we believe that for each available specialty, only one national physician salary amount should be used. Further, we agree with many commenters that this physician salary amount should not be adjusted for geographic variation because doing so would add an Start Printed Page 26958additional layer of complexity. In cases where no subspecialty salary amount is available in the AMGA data, hospitals should use the physician salary amount for the closest less-specialized form of that specialty. For example, as we proposed in the RY 2008 LTCH PPS proposed rule (72 FR 4824), “* * * if residents are receiving training from a forensic pathologist, and the national average salary for the subspecialty of forensic pathology is not included in the physician compensation survey, then the hospital should instead use the national average salary for the specialty of pathology to determine the cost of that teaching physician.” We also agree with the commenters' suggestion that median salary amounts should be used as the proxy physician salary amount since median salary amounts would not be influenced by outlier data. Therefore, we are finalizing the policy to require hospitals that choose to use the proxy method to calculate supervisory teaching physician costs to use AMGA's median physician salary amount for the required specialty.

Comment: One commenter stated that CMS should use average compensation figures for dental faculty based on specialty and regional variation. The commenter stated that the commenter would be happy to work with CMS to develop compensation figures for dental programs.

Response: While we appreciate the point raised by the commenter that the AMGA data does not apply to dental faculty, at this point we are unaware of a comparable data source for dental faculty salaries. We will work with the commenter to determine whether we can develop proxy salary amounts for supervisory dentists.

Comment: One commenter suggested that for added administrative simplicity in determining proxies, hospitals should be able to use “* * * two ‘blended’ supervising physician salary amounts—‘one for primary care and one for non-primary care * * *.” These “blended” salary amounts would be determined using the published data source. The commenter stated that to determine which salaries should be included in the blends, a periodic survey could be taken to determine the composition of teaching physicians at each nonhospital site. Another commenter stated, “We would also like to recommend that the CMS maintain as part of the final rule, the provision that allows providers to use actual teaching physician salaries for the calculation of the recommended cost threshold instead of the national average physician salary data by specialty.”

Response: We appreciate the commenter's innovative suggestion to use “blended” salary amounts in determining a proxy for supervisory teaching physician costs. However, in choosing a proxy for national physician salaries, in order to determine the teaching physician cost at the nonhospital site, we believe the proxy should be as close to the actual salary amount as possible. Therefore, we believe it is most appropriate for hospitals to use the published AMGA specialty salary amounts in determining the supervisory teaching physician costs at the nonhospital site. In response to the commenter's request that we maintain the option for hospitals to use actual physician salary information, we note that the proposal was to add a proxy calculation as an alternative to hospitals documenting that they have paid the actual teaching physician costs at the nonhospital site. Hospitals always have the option of using actual data instead of any of the proxies. We also note that under our revised policy, hospitals that use actual data are required to only pay 90 percent of the total of the costs of the residents' salaries and fringe benefits (including travel and lodging where applicable) and the portion of the cost of the teaching physicians' salaries attributable to nonpatient care direct GME activities.

Comment: Several commenters questioned the potential availability of AMGA's survey data and requested that it be made available on our Web site. One commenter stated that AMGA charges a fee to access its data and if we are requiring hospitals to use AMGA data, the data, as well as information on AMGA's methodology should be made available without cost to the public on CMS’ Web site. The commenter stated “* * * because the AMGA survey and its methods are not freely available, providers may not easily be able to analyze and concur with AMGA's methodology or the amounts set forth in Table 8 * * *” One commenter noted that since there is a fee to access AMGA data, using that data or other similar data (which requires a fee) would be inappropriate because we would be imposing additional costs on GME. The commenter further noted (referring to AMGA's data), “It is not clear how representative of all practicing physicians these respondents are.”

Response: We will make available any physician specialty salary survey data that is needed to compute teaching physician supervisory costs available free of charge on our Web site. Additionally, we will consider posting information on the AMGA's survey methodology. By posting the AMGA data on our Web site, we are not imposing any additional cost on GME training that occurs at nonhospital sites. Since AMGA's survey data will be posted free of charge, we do not believe there will be any costs associated with accessing the necessary data.

We disagree with the commenter regarding the level of physician salary representation in AMGA's survey. AMGA's survey includes a range of physician specialty salaries. In fact, because of the broad range of specialties included in the survey we believe AMGA's survey data are particularly appropriate for use to establish a proxy for teaching physician salaries and well-suited to meet our goal to use salary information that reflects physicians' actual salaries.

Comment: One commenter asked whether a provider could use an alternative survey similar to AMGA if it can demonstrate that the survey was compiled in a similar manner. Another commenter stated that in determining the proxy salary amounts to be used, we should “* * * consider the approach used by the Department of Veterans Affairs in setting salaries for its physicians, notably by employing multiple surveys of physician compensation.”

Response: In response to the commenters' question of whether a survey similar to AMGA's could be used as a proxy source or a combination of surveys, in establishing the proxy, we are allowing a hospital to base its determination on either AMGA survey data or actual physician salary amounts. However, as previously mentioned, we will continue to consider the appropriateness of using other options for sources of physician salary data.

Determining Teaching Physicians' Cost

In determining the teaching physicians' cost, the specialty of the teaching physician is the relevant criterion, not the specialty of the residents that the teaching physician is training in the nonhospital site. Generally, we believe the specialty of the teaching physician will be self-evident, and the hospital can easily locate the national average salary information for that teaching physician's specialty on the survey (for example, if family practice residents are rotating to a dermatology practice to receive training in dermatology, then the national average salary for dermatologists would be used from the survey). However, it is possible that the teaching physician is highly specialized and the average compensation for his or her subspecialty is not listed in the survey we decide to use. In such a case, Start Printed Page 26959we proposed that the hospital should use the immediately less-specialized form of that specialty applicable to that teaching physician (or the hospital may use the physician's actual salary information). For example, if residents are receiving training from a forensic pathologist, and the national average salary for the subspecialty of forensic pathology is not included in the physician compensation survey, then we proposed that the hospital should instead use the national average salary for the specialty of pathology to determine the cost of that teaching physician. We believe this is the simplest method of assigning a national average physician compensation amount in the instance where the teaching physician's actual subspecialty is not included in the survey. However, we solicited comments as to whether it is possible or appropriate to use survey data from other sources in the event that data is not available from the particular survey source.

In addition, although it may not be a common occurrence, it is possible that residents could be receiving training in a nonhospital site from a teaching physician that is board certified in more than one specialty, but the residents are only receiving training in one of the specialties in which the physician is board certified. In this case, we proposed that the national average salary that should be used to determine the teaching physician's cost should be the one for the specialty in which the teaching physician is training the residents. For example, if residents are being supervised by a cardiologist who is board certified in internal medicine and cardiology, but the residents are training with him or her specifically to learn internal medicine, then we proposed that the hospital should use the national average salary for internal medicine, and not cardiology, to determine the teaching cost of that physician. That is, in instances where the residents are receiving training at a nonhospital site from a teaching physician that is board certified in more than one specialty, and it is unclear which specialty to use for purposes of assigning a national average salary to that physician, we proposed that the question for the hospital to ask is, why are the residents training with that physician? If the answer is, “to receive training in Specialty X,” then the national average salary amount for Specialty X should be used to determine the teaching physician's cost. If the answer is, “to receive training in Specialty Y,” then the national average salary amount for Specialty Y should be used to determine the teaching physician's cost, regardless of the specific board certification that the teaching physician has actually received. In general, the hospital, with assistance from the GME Program Director as necessary, should be able to document for the Medicare contractor the specialty in which the residents are receiving training at the nonhospital site, and the national average physician compensation amount for that specialty used in paying “all or substantially all” of the costs, as defined in this final rule.

Comment: A commenter stated that the specialty of the resident and not of the teaching physician should be used in determining the specific salary proxy. The commenter provided the example that a cardiologist will teach an internal medicine resident what he or she is required to know regarding heart disease and the cardiovascular system as an internist and not a cardiologist. The commenter further requested that we “* * * clearly state that proxy salaries for subspecialty physicians originally trained in the specialty of the residents they are teaching be set to the salary of specialists in the residents' field regardless of the certification status of the faculty person.”

Response: In response to the commenter's request that the specialty of the resident be used in determining the supervisory teaching physician cost, we stated in the proposed rule * * * that the national average salary that should be used to determine the teaching physician's cost should be the one for the specialty in which the teaching physician is training the residents.” For example, if a resident happens to be supervised by a physician who is board certified in internal medicine and cardiology, but the resident is training with him or her specifically to learn general internal medicine, then we proposed that the hospital should use the national average salary for internal medicine, and not cardiology, to determine the teaching cost of that physician. However, if the internal medicine resident is at the nonhospital site to receive cardiology training as part of his or her 3-year internal medicine program, the salary for cardiologists should be used. In instances where the residents are receiving training at a nonhospital site from a teaching physician that is board certified in more than one specialty, and it is unclear which specialty to use for purposes of assigning a national average salary to that physician, we proposed that the question for the hospital to ask is, why are the residents training with that physician? If the answer is, “to receive training in X,” then the national average salary amount for Specialty X should be used to determine the teaching physician's cost. If the answer is, “to receive training in Y,” then the national average salary amount for Specialty Y should be used to determine the teaching physician's cost, regardless of the specific board certification that the teaching physician has actually received. We believe the teaching physician supervisory cost should reflect the value of the training received as it relates to the training the resident is receiving. Therefore, we are not adopting the commenter's suggestion to use the physician salary of the specialty program of the resident regardless of the specifics of the training received.

Multiple Teaching Physicians and Residents: 1:1 Resident to Teaching Physician Ratio

We understand that it is not unusual for several residents in the same program to rotate to a particular nonhospital site at the same time, and be supervised by one teaching physician, or for residents to be supervised by several teaching physicians during their time at that nonhospital site. In determining the total costs of the training program at the nonhospital site, it is necessary to consider all of the residents' salaries and fringe benefits (including travel and lodging where applicable), and the teaching physicians' national average salaries. However, to maintain administrative simplicity, we are allowing hospitals to apply a maximum of a 1:1 resident-to-teaching physician ratio “limit” in determining the total GME costs applicable to a program at a nonhospital site. For example, if at the nonhospital site there are two teaching physicians and one FTE resident, the hospital may determine 90 percent of the total costs of the program using a 1:1 resident-to-teaching physician ratio, not a 1:2 resident-to-teaching physician ratio. The 90 percent threshold would be based on the total cost of the one FTE resident (salary and fringe benefits, and travel and lodging where applicable) and one teaching physician (national average salary for the specialty multiplied by the percentage of time spent in nonpatient care direct GME activities). Similarly, if a hospital rotated 3 FTE residents in the same program to a particular nonhospital site with 7 physicians, unless the hospital documents otherwise, we would assume that all 7 physicians supervise the residents at some point during the training, but, for purposes of determining the 90 percent threshold, we assume that there are only 3 FTE residents being supervised by 3 teaching Start Printed Page 26960physicians. Accordingly, the 90 percent threshold would be based on the total cost of the 3 FTE residents' salaries and fringe benefits (including travel and lodging where applicable) and 3 teaching physicians (national average salaries for the specialties multiplied by the percentage of time spent in nonpatient care direct GME activities). (In addition, we note that the 1:1 limit may be applied to FTE fractions, as well. That is, if in the preceding example, 3.5 FTE residents were being supervised by 7 physicians, the 90 percent threshold would be determined based on the costs associated with a resident-to-teaching physician ratio of 3.5:3.5.)

In the case of multiple teaching physicians, we must also consider that a particular nonhospital site may be staffed by physicians in different specialties. For example, an orthopedics practice may include orthopedists and radiologists. In this case, we would still maintain the 1:1 resident-to-teaching physician limit, even if the teaching physicians are in different specialties, unless the hospital can document that the number of physicians actually teaching the residents is less than the number of FTE residents training at that nonhospital site. Once the number of teaching physicians is established, the hospital would determine the national average salary for each of those teaching physicians from the national survey data, and then calculate the average national salary of the mix of physician specialties in the practice to be used in computing the 90 percent threshold. For example, assume that 3 FTE residents are rotating to an orthopedic surgery practice staffed by a total of 7 physicians; 4 are orthopedic surgeons, and 3 are diagnostic radiologists. Again, unless the hospital documents otherwise, we would assume that all 7 physicians supervise the residents at some point during their rotation to this practice. First, the hospital would access the national average salary for orthopedic surgeons (assume $400,000), and the national average salaries for diagnostic radiologists (assume $412,000). Then, the hospital would calculate the average salary for these physicians as follows: [($400,000 × 4) + ($412,000 × 3)]/7 = $405,143. Next, the 1:1 resident-to-teaching physician ratio would be applied, such that for purposes of determining the 90 percent threshold, there would be 3 FTE residents and 3 teaching physicians. Since the 3 teaching physicians are not in the same specialty, the hospital would multiply the average salary cost of $405,143 by 3 to get the total teaching physician salaries for the training program at that site ($405,143 × 3 = $1,215,429). The hospital would then multiply $1,215,429 by the percentage of time spent by the teaching physicians in nonpatient care direct GME activities (that percentage is 3 hours divided by the number of hours the practice is open during a week) to determine the teaching physician GME cost for the training program at that site. This teaching physician cost is then added to the salaries and fringe benefits (including travel and lodging where applicable) of the 3 FTE residents to determine the GME cost of the program at that practice, and the hospital must ensure that it incurs at least 90 percent of that GME cost to count the 3 FTE residents training at the nonhospital site.

We note that, as we indicated above in this section, if there are several physicians in a nonhospital site, we would assume that they all supervise the residents at some point during the residents' training. However, it may be that in fact only some of the physicians actually supervise the residents, while other physicians are not involved in the training program at all. The hospital may wish to document that only certain physicians are involved in the training program (to more accurately represent the structure and costs of the training program in a particular nonhospital site). Such documentation would increase the number of residents relative to teaching physicians that is used to calculate the teaching physician costs. That is, using the example above where the resident-to-teaching physician limit was presumed to be 3:3, since there were actually 3 FTE residents and 7 physicians, if the hospital can document that only 2 physicians supervised the residents (and the other 5 physicians were not involved in the GME program at all), then the resident-to-teaching physician ratio would be 3:2. As a result, the hospital might be required to incur less teaching physician costs, if any, to meet the 90 percent threshold.

Comment: One commenter stated that in using a 1:1 ratio in determining the 90 percent threshold, it is unlikely that a hospital will meet the 90 percent threshold because physician salaries are quite a bit higher than resident salaries and fringe benefits particularly among specialties. Commenters also asked what documentation we are requiring to show that only certain teaching physicians at nonhospital sites are supervising residents. One commenter asked that we confirm that this information should be provided after the resident rotation to the nonhospital site has occurred.

Response: We proposed to adopt the 1:1 ratio so that there would be an upper limit on the number of physicians that are supervising residents in the nonhospital site. We believe that use of a 1:1 ratio greatly reduces the cost a hospital would have to pay when there is actually a higher teaching physician to resident ratio. For example, if two teaching physicians were supervising one resident, in the absence of the 1:1 ratio, the costs for both of those teaching physicians would be included for purposes of making the “all or substantially all” calculation. Thus, hospitals could be required to pay significantly more of the physician salaries if the teaching physician to resident ratio is not capped at 1:1. The 1:1 cap does not apply to the number of residents (and thus the resident salary and fringe benefit calculation). Therefore, where there is one teaching physician training three residents, the hospital would calculate teaching physician costs using one teaching physician salary and all three of the residents' salary and fringe benefit data. In response to the commenters' request that we advise what type of documentation hospitals need to submit to show that only certain teaching physicians are supervising residents, the hospital should have the teaching physicians that were not involved in the training submit documentation at the end of the rotation or by the end of the applicable academic year (June 30) to indicate that they were not involved, either directly, or indirectly, with the education of residents in their practice. Alternatively, those physicians involved in the training can be identified in the written agreement, or the hospital may submit contemporaneous documentation from the GME program director specifying which physicians were involved in supervising the residents.

(2) Residents' Salaries and Fringe Benefits

The second variable in our methodology for determining the costs of a program at a nonhospital site is the salaries and fringe benefits (including travel and lodging where applicable) of the FTE residents that are rotating to a particular nonhospital site. We understand that since the salaries and fringe benefits (including travel and lodging where applicable) of most residents are already paid by hospitals (either directly, or by reimbursing another entity such as a medical school), the portion of the actual cost of the residents attributable to training in the nonhospital setting can be easily identified and documented by a Start Printed Page 26961hospital. Therefore, as under existing regulations, in determining the 90 percent threshold for a particular program at a specific nonhospital site, the hospital must use the actual cost of each FTE resident's salary and fringe benefits (including travel and lodging where applicable). In addition, the cost of the residents will vary by specialty and by program year. Furthermore, as with current policy, the total residents' costs will be based on the FTE number rotating to a particular nonhospital site in a cost reporting period, not the number of individuals actually training in a nonhospital site.

Comment: Several commenters requested that we specify what is included in resident salaries and fringe benefits. Several commenters also requested that we specify that resident malpractice insurance is included in resident fringe benefits.

Response: It is not our intent to cause hospitals to modify their human resources policies regarding residents' salaries and fringe benefits. Hospitals should maintain their definition of residents' salaries and fringe benefits that was in place prior to the RY 2008 LTCH PPS proposed rule. Hospitals should not include resident malpractice insurance or other costs in residents' fringe benefits solely for the purpose of increasing the total cost of residents' salaries and fringe benefits and minimizing the portion of teaching physician costs they have to pay. Furthermore, we note that historically, malpractice costs were not to be included in the intern and resident cost center on the cost report. Accordingly, malpractice costs should not be included as a fringe benefit in the calculation of the 90 percent threshold.

Comment: One commenter was concerned about our requirement that a hospital must use the actual costs of each FTE resident's salary and fringe benefits as one of the variables under the proposed methodology for determining the minimum amount that a hospital must pay to count FTE residents training in a nonhospital site. The commenter stated that under our current policy, a hospital only needs to know in general that it incurred the costs of residents' salaries and fringe benefits, but need not know the actual amounts paid; whereas under the proposed methodology, a hospital would have the significant administrative burden knowing the precise program year and corresponding salary and fringe benefits amount for each resident that trains in the nonhospital setting. The commenter suggested that we allow hospitals the option of using an average salary plus fringe benefit amount as a means of simplifying the proposed methodology and to provide administrative relief for hospitals.

Response: In the RY 2008 LTCH PPS proposed rule, we stated that we would allow a hospital to use physician compensation survey data as a proxy to determine the teaching physician costs associated with a program at a particular nonhospital site. We proposed to allow the hospital to use a proxy amount because hospitals stated that the existing regulation was administratively burdensome since many teaching physicians in nonhospital sites are reluctant to disclose their personal salary information. We proposed this policy because teaching physicians in a nonhospital site may not be employed or paid by the hospital, and hospitals indicated they had great difficulty establishing the teaching physicians' salaries and the portion of the cost attributable to the nonpatient care direct GME activities of the teaching physicians.

In contrast, we believe resident salary and fringe benefits amounts are more readily available to hospitals since they ordinarily pay these costs directly. Because hospitals have ready access to this data, we believe it is appropriate that hospitals use the actual costs of resident salaries and fringe benefits for the calculation of the 90 percent threshold, rather than some sort of proxy.

The commenter is correct that to calculate the actual resident salary and fringe benefits amounts, hospitals will have to take into account the actual salary and fringe benefits for each FTE resident that trains in the nonhospital site, which may vary by resident.

Comment: Several commenters inquired about which travel and lodging expenses should be considered as applicable to direct GME in the nonhospital site.

Response: Residents' fringe benefits (including travel and lodging where applicable) are considered a part of “all or substantially all of the costs for the training program in the nonhospital setting.” The only travel and lodging costs that are applicable are the additional travel and lodging costs that a hospital incurs due to the fact that a resident is training at a nonhospital site. For example, if a resident needs to travel long distance to another part of the state, and is staying in a hotel for the duration of the nonhospital site training, the costs of the traveling and accommodations would be costs that the hospital must incur and include in the determination of the 90 percent threshold. However, expenses that are normally incurred when the resident trains at or nearby the hospital, such as commuting and living expenses, would not be applicable.

(3) The Number of Hours Spent in Nonpatient Care Direct GME Activities in a Week and the Number of Hours That the Nonhospital Site is Open in a Week

The third variable used in the determination of the costs of a training program at a nonhospital site is the amount of time that the teaching physician(s) spends on direct GME (nonpatient care) activities in a week. As we first explained in the July 31, 1998 Federal Register (63 FR 40987), and more recently in the August 8, 2005 Qs&As posted on the CMS Web site at http://www.cms.hhs.gov/​AcuteInpatientPPS/​Downloads/​nonhospQA.pdf, determination of the teaching physician costs to the nonhospital site is dependent upon the teaching physician's salary and the percentage of time he or she devotes to activities related to non-billable direct GME activities at the nonhospital site (such as conferences, practice management, lectures, and administrative activities like resident evaluations). Hospitals and teaching physicians have protested that documenting the percentage of time that teaching physicians spend on activities relating to nonpatient care direct GME activities at the nonhospital site is an onerous and impractical task. In an effort to eliminate the documentation burden on physicians of keeping track of the amount of time they spend in nonpatient care direct GME activities in the nonhospital site, rather than require teaching physicians to estimate the number of hours per week that they spend in such activities with or on behalf of the residents, we proposed an alternative option that hospitals may choose to use to determine the percentage of the teaching physician's time that is spent in nonpatient care direct GME activities. This option is an administrative shortcut or a proxy, rather than continuing to require in all cases that the hospital must document and pay for the actual costs of a training program at a nonhospital site. However, a hospital always has the option of documenting and paying for at least 90 percent of the costs of a program at a nonhospital site using the teaching physician's actual salary and information on the time spent in nonpatient care direct GME activities.

Under the proxy methodology, we would apply a presumed standard number of hours spent by teaching physicians in nonpatient care direct Start Printed Page 26962GME activities in every nonhospital site. Specifically, we proposed to use a standard of 3 hours per week spent in nonpatient care direct GME activities by teaching physicians. The 3 hour standard would be used in all cases in the formula for determining the teaching physician costs at all nonhospital sites, regardless of the specialty of the residents or the number of teaching physicians or residents training at that nonhospital site. Although some hospital industry representatives have stated that the amount of time spent by teaching physicians in nonpatient care direct GME activities in nonhospital sites is “de minimus,” and, therefore, there is typically little if any teaching cost to the nonhospital site, we believe there is also evidence indicating that in many cases the teaching physician is spending a significant amount of time with or on behalf of the residents in nonpatient care direct GME activities. We believe the standard of 3 hours of nonpatient care direct GME activities per week is a reasonable proxy based on data collected from surveys conducted by the Association of American Medical Colleges (AAMC), the American Osteopathic Association (AOA), and the Academic Family Medicine Advocacy Alliance (AFMAA), in addition to information compiled from our own informal surveys of teaching physicians.

In September 2005, in response to a request by CMS, the AFMAA, AOA, and AAMC conducted informal surveys to determine the amount of time spent in nonpatient care direct GME activities by teaching physicians in nonhospital sites. In the survey results shared with CMS by these associations, we received a range of hours for the amount of teaching physician time spent per week in nonpatient care direct GME activities at the nonhospital site. Such nonpatient care GME time included time spent by the teaching physician in training activities when the patient was not present and time spent in administrative activities related to the GME program. The surveys showed means ranging from 1.1 to 4.0 hours per week and medians of 1.5 to 4.0 hours per week for time spent on residency training when patients were not present. The surveys also showed means ranging from 1.6 to 4.7 hours per week and medians of 0 to 2 hours per week for time spent on administrative activities related to residency training at the nonhospital site. Given the range of survey results, we believe that 3 hours per week serves as a reasonable number to use as a shortcut or a proxy for determining teaching physician time spent in nonpatient care direct GME activities at the nonhospital site. As previously stated, hospitals always still have the option of calculating teaching physician costs and the 90 percent cost threshold using actual data (as under current regulations) specific to the number of hours the teaching physician spends per week on nonpatient care direct GME activities at the nonhospital site. For example, if a hospital can document that a teaching physician actually spends 1.5 hours per week on nonpatient care direct GME activities at the nonhospital site, then the hospital may use 1.5 hours per week in calculating the teaching physician cost and the 90 percent cost threshold.

We proposed to use the standard of 3 hours of nonpatient care direct GME activities per week as the proxy regardless of the number of FTE residents the teaching physician is supervising because we believe that when the number of FTE residents at a nonhospital site increases, the teaching physician time associated with those FTE residents in many instances will increase by only a small multiple. For example, a teaching physician would provide a lecture to the residents together, rather than separately lecturing each FTE resident who is training at the nonhospital site. Accordingly, the time spent by the teaching physician in nonpatient care direct GME activities may increase only slightly with each additional FTE resident being supervised.

While we proposed to use the standard number of hours spent by teaching physician(s) in nonpatient care direct GME activities across all training occurring at all nonhospital sites (that is, 3 hours per week), we are introducing a fourth variable in the determination of the cost of a training program in a nonhospital site that will vary depending on the specific nonhospital site. This fourth variable is the number of hours that a nonhospital site is open each week. Since only a percentage of the teaching physician's salary is attributable to direct GME activities, and that percentage is based on time he or she devotes to activities related to non-billable direct GME activities at the nonhospital site, we are determining this percentage by dividing the standard number of hours spent in nonpatient care direct GME activities by the number of hours the specific nonhospital site is open each week. We proposed that the numerator will always be 3 hours, and the denominator will vary depending on the nonhospital site. For example, if FTE residents rotate throughout the year to a nonhospital site that is open 40 hours per week, then the percentage of time spent by the teaching physician(s) in nonpatient care direct GME activities throughout the year at that site is 3/40 = 0.075 or 7.5 percent. (If FTE residents rotate to that nonhospital site for only a portion of a year, then the ratio of 3/40 would be further multiplied by the percentage of the year that the FTE residents train there. For example, if the FTE residents only rotate to this nonhospital site for 3 months of the year, then the percentage of time that the teaching physician(s) spends on nonpatient care direct GME activities at that site equals (3/40 × 0.25 = 0.019 or 1.9 percent). Similarly, if FTE residents rotate throughout the year to a nonhospital site that is open 50 hours per week, then the percentage of time spent by the teaching physician(s) in nonpatient care direct GME activities throughout the year is 3/50 = 0.06 or 6 percent. We recognize that the teaching physician(s) may not spend 100 percent of his or her time in that nonhospital site. In fact, many teaching physicians spend some of their week working in a hospital or other facilities. However, we believe that deriving the true amount of time spent by each teaching physician in each nonhospital site in nonpatient care GME direct GME activities would involve the imposition of another form of the documentation burden that the hospital industry and teaching physicians have found onerous up to this point. This methodology eliminates the need for any time studies and it is easy to gather the information needed.

We also acknowledge that the proposal to use the number of hours that a particular nonhospital site is open as a proxy in the denominator for determining the percentage of time spent by the teaching physician(s) in nonpatient care direct GME activities could, in some extreme instances, result in an unusually high percentage of teaching time, which, in turn, would result in a determination of unusually high teaching costs. This is so because, since 3 hours is a constant in the numerator, the fewer the number of hours the clinic is open (the denominator), the greater the calculated percentage of time spent by the teaching physician in nonpatient care direct GME activities. To use an extreme example, if a clinic is only open 10 hours a week, then 3/10, or 30 percent of the national average salary for the teaching physician's specialty would represent the teaching physician's cost that would be used to determine 90 percent of the costs of the program at the clinic. However, we believe that, for most nonhospital training situations, this revision to use the 3 hour standard and the number of hours the nonhospital Start Printed Page 26963site is open per week is a reasonable alternative to the current procedures for determining the actual teaching physician's cost because these proxies are easily obtainable, discrete numbers that do not necessitate any time studies. Nevertheless, we solicited comments on alternative proxies that might be appropriate to use in the place of the ratio of 3 hours to the number of hours a nonhospital site is open per week. We also note that in the event that this methodology for calculating teaching physician costs in a particular nonhospital site results in an unrealistic amount, we reiterate that a hospital always has the option of determining and paying at least 90 percent of the GME costs using actual physician salary and teaching time information, for all, or some of its training programs occurring in nonhospital settings. In fact, a hospital may choose to use a combination of actual information and proxy information for determining the teaching physician cost. For example, a hospital may choose to use actual physician salary information instead of the national average survey data, but use the 3 hour standard and the number of hours the nonhospital site is open per week to determine the percentage of time spent on teaching activities, or vice versa. Furthermore, we reiterate that under the new definition of “all or substantially all,” even if a hospital chooses to document the teaching physician cost using actual teaching physician-specific information, the hospital need only incur 90 percent of the residents' salaries and fringe benefits (including travel and lodging where applicable), and the portion of the teaching physicians' salaries attributable to direct GME, and not 100 percent of those costs.

Under our revised policy, 90 percent of the GME costs for a particular program at a particular nonhospital site would be the minimum amount that a hospital must pay to count the FTE resident(s) training at that site for direct GME and IME purposes. If the hospital is already paying the resident's salaries and fringe benefits (including travel and lodging where applicable), and if the costs of the resident's salaries and fringe benefits are equal to at least 90 percent of the total GME costs at the nonhospital site (that is, the 90 percent threshold), then the hospital is paying “all or substantially all” of the costs in accordance with our definition, and need not pay an additional amount for teaching physician compensation to count the FTE residents. However, if the hospital is paying less than 90 percent of the costs of the training program at the nonhospital site, then the hospital must pay an additional amount toward the teaching physician costs until it is paying at least 90 percent of the GME costs for that program. We believe our revised policy is relatively simple, easy to administer, and eliminates the documentation burdens cited by the industry as being associated with the current policy. However, we note again that even under our revised policy, a hospital is not precluded from choosing to calculate and pay 90 percent of the teaching costs of a program in a nonhospital site in accordance with the existing policy requirements. That is, the hospital may still choose to document the actual teaching physician cost using actual time and salary information from the teaching physician(s) to determine what the true direct GME costs are at that nonhospital site. Once the hospital calculates the actual direct GME costs, it would only be required to pay at least 90 percent of the actual direct GME costs, consistent with our definition of “all or substantially all of the costs for the training program in the nonhospital setting.”

The following is an additional example of the application of the methodology:

Example: For the July 2008 through June 2009 academic year, a hospital with a family practice program sends 3 FTE residents (in different program years) to train at the Family Medicine Center (FMC), a nonhospital site. The hospital's cost reporting period began on January 1, 2008. The FMC is staffed by 5 physicians, all of whom supervise the residents at some point during the year. Four of the physicians are family practitioners, and 1 physician is a psychiatrist. The FMC is open for 50 hours per week. To determine the cost of the teaching physicians, the hospital refers to the most recent national average salary amounts on the national survey published prior to January 1, 2008, which is the 2007 survey. Assume that the national average published salary amount for family practice is $180,000, and the national average published salary amount for psychiatry is $187,000. Since there are multiple physicians in different specialties (absent specific documentation provided by the hospital), the average salary of one FMC physician is calculated as follows: [($180,000 × 4 family practice physicians) + ($187,000 × 1 psychiatrist)]/5 = $181,400. Since the residents are on the payroll of the hospital, the hospital knows that the total actual cost of the 3 FTE residents' salaries and fringe benefits (including travel and lodging, if applicable) is $182,000. After applying the 1:1 resident-to-teaching physician limit, there are 3 FTE residents to 3 teaching physicians (again, absent specific documentation provided by the hospital). Thus, the GME cost of the 3 teaching physicians is calculated as follows: ($181,400 × 3) × (3 hours/50 hours) = $32,652. This teaching physicians' cost of $32,652 is added to the residents' cost of $182,000 to arrive at the total cost of the training program at the nonhospital site of $214,652. To meet the definition of “all or substantially all,” the hospital would be required to pay at least 90 percent of the costs of the training program at the nonhospital site, which in this example equals $193,187 (that is, 0.90 × $214,652). Since in this case the cost of the 3 FTE residents' salaries and fringe benefits is $182,000, the hospital would not reach the 90 percent cost threshold by simply incurring the costs associated with the residents. The hospital must pay at least an additional $11,187 (that is, $193,187−$182,000) to meet the 90 percent threshold and satisfy the requirement to pay “all or substantially all” of the costs of the family practice program at the FMC.

Comment: One commenter, the Association of American Medical Colleges (AAMC), noted that in the proposed rule, we stated that “the standard of 3 hours of nonpatient care GME activities per week is a reasonable proxy based on data collected from surveys conducted by the Association of American Medical Colleges (AAMC), the American Osteopathic Association (AOA), and the Academic Family Medicine Advocacy Alliance (AFMAA), in addition to our own informal surveys of teaching physicians” (72 FR 4826). The AAMC commented that they would “like to clarify that the AAMC did not provide CMS with survey data.” The AAMC indicated that we may have been confused on this issue because the surveys were presented to CMS in a meeting in which representatives of the AAMC were in attendance, and they noted that AAMC staff provided some input to the survey questions. A commenter said that we were correct to describe the surveys as “informal” (72 FR 4826), since these surveys were developed and conducted by AOA and AFMAA policy staff who, due to time constraints, did not consult with persons who have expertise in survey development. Another commenter stated that any data collected by CMS informally and used as the basis for a regulation should be available to the public. A commenter referred to the limitations to the data that the AFMAA Start Printed Page 26964noted when it submitted its survey data to CMS, and questioned why CMS would use such “extremely flawed” data, when anecdotal evidence suggests that any time greater than one hour per week spent in didactic training is “way out of line with actual circumstances.” Commenters enlisted a professor from the Department of Economics at Hunter College in New York, to analyze the survey data and opine as to whether the survey responses provide a valid source for establishing a national proxy. The professor expressed concerns about the data provided to CMS, stated that the data are extremely limited and questionable and should not form the basis of public policy, and suggested that CMS conduct its own rigorous study to identify the best proxy. The professor's analysis also recommended that in the meantime, if CMS wishes to make a decision based on the AOA and AFMAA survey, a proxy that is better supported by the current survey is 2 hours.

Some commenters also asked that CMS consider that the surveys were conducted prior to the issuance of the FY 2007 IPPS final rule in which CMS clarified that time spent in nonpatient care activities in nonhospital sites cannot be counted by a hospital for direct GME and IME purposes. Because of this clarification, hospitals may now be conducting as much of their didactic activities as possible in the hospital complex. Lastly, the commenters noted that to the extent that a resident may spend only a half a day at a nonhospital site per week, “the idea that [the] 2 or 3 hours of that time is spent in nonpatient care activities defies conventional logic.”

Several commenters suggested that the 3 hour proxy should be reduced to either 1 or 2 hours. One commenter stated that according to the commenter's survey of 54 physicians, the average hours per week spent on nonpatient care direct GME activities was 1.45, with a range of 0 to 6 hours. Another commenter stated that teaching physicians spend 1.2 to 1.5 hours a week in nonpatient care direct GME activities, while one commenter mentioned that for family practice, a teaching expectation of 20 minutes per half day would work best. Several commenters stated that CMS should adjust the proxy according to a resident's program year. For example, one commenter suggested that the number of hours spent in nonpatient care direct GME activities per week should be 1 hour for third year residents, 2 hours for second year residents, and 3 hours for first year residents.

Response: We regret that we inadvertently misattributed the surveys in part to the AAMC. The AAMC is correct that we believed they did have a role in conducting the surveys, but based on their comments, we understand that their role was limited to providing some input into the survey questions. We acknowledged that the surveys conducted by CMS, the AFMAA, and the AOA respectively were informal, and we understood that persons with expertise in survey development were not necessarily consulted due to time constraints. In light of these considerations, we carefully reviewed the analysis of the surveys provided by the professor from Hunter College. We agree that it is inappropriate to apply a proxy of 3 hours to one nonhospital site if the residents only rotate to that nonhospital site for a portion of the week. As we explain further below in response to the comments we received about prorating the teaching physician's cost, in this final rule, we are allowing hospitals to prorate the teaching physician's costs to reflect the FTE time spent by the residents in a program at each nonhospital site. Since we have heard from the teaching hospital industry that it is unlikely that a resident will spend an entire week at the same nonhospital site, in those cases, the hospital would be applying a prorated proxy, which would be less than 3 hours, and may even be less than the 2 hours which the professor from Hunter College indicated could be supported by the survey data. The suggestion from the professor at Hunter College that we conduct a rigorous study is sensible, and we will consider it.

In response to the commenters who request that the 3 hour proxy be adjusted according to a resident's program year, we believe that requiring a hospital to adjust the proxy for each of its residents who are training at a nonhospital site would add unnecessary complexity. Therefore, we are finalizing our policy to use 3 hours in the numerator of the teaching physician cost ratio. We note that if a hospital believes that 3 hours is greater than the actual amount of time spent in nonpatient care direct GME activities in a particular nonhospital site, the hospital always has the option to work with the teaching physician to provide an actual amount of teaching time for use in calculating the 90 percent cost threshold.

In response to the comment requesting that we consider that the amount of time currently spent in nonpatient care direct GME activities in the nonhospital site could be less than the amount shown in the surveys (since the surveys were conducted prior to the issuance of our clarification regarding didactic activities), we believe this might be true. We acknowledge that the availability of Medicare GME funding is certainly an important factor in a hospital's decision to rotate (or not rotate) residents to nonhospital settings. However, we also recognize there are other significant factors that hospitals must consider in making residency rotation decisions, such as the requirements of accrediting organizations (like the ACGME or the AOA), and local health “outreach” initiatives. Thus, we are skeptical that hospitals’ longstanding rotational models would shift so dramatically and in such a short period of time due to clarification of the agency's policy regarding the time that residents spend in didactic activities. Further, the commenter is raising a point that can be made about any survey which captures data as of a certain period of time, and cannot necessarily be used to predict future scenarios. However, we may re-evaluate the use of the 3-hour per week standard, possibly in conjunction with a new survey, in the future if appropriate.

Comment: Commenters suggested that since the goal of the proposed rule was to reduce administrative burdens, instead of requiring that hospitals determine the number of hours each nonhospital site is open, we should consider using a national average proxy for total physician work hours per week. A commenter mentioned that there are limited, but still apparently reasonable, data that exist on national average physician work hours. For example, in its 2006 physician workforce report, the Health Resources and Services Administration (HRSA) used the American Medical Association's (AMA) Socioeconomic Monitoring System (SMS) from 1998 to estimate work hours by specialty. (The commenter noted that this survey has been discontinued due to response rates that were often too low for individual specialties and practice settings.) The direct patient care hours reported by HRSA ranged from 47 to 58 hours per week. Another study conducted in 2005 by the AAMC's Center for Workforce Studies of physicians over age 50 showed an average of 55 hours worked per week based on over 9,000 respondents, with work hours varying by specialty. For instance, pathologists worked an average of 50 hours weekly on the lower range, while cardiologists worked an average of 63 hours a week. Similarly, data from the Center for Tracking Health System Change reported an average of Start Printed Page 2696553 hours worked per week based on interviews with about 6,600 physicians in all specialties. The commenter asked that we adopt 55 hours as the proxy to use, but suggested that it might be best to use specialty-specific proxies, since there is a range of work hours across specialties. Another commenter suggested that physician work hours as published in JAMA, 2003 be used in the denominator. Alternatively, if we decide to adopt our proposal regarding the clinic hours of operation, then the commenters requested that we confirm that this means the “posted” hours, and not the actual hours (for example, the hospital need not account for the closure of the site due to a holiday). Another commenter asked that CMS include a definition of “hours open” in the final rule, and specify what documentation would be required.

Other commenters suggested that instead of the clinic hours of operation, the denominator of the ratio used to calculate the teaching physician cost proxy should be the number of hours the teaching physician is working since the physician's salary is relative to the number of hours worked. One commenter requested that we allow adjustments as appropriate when the teaching physician spends only a portion of his or her time at the nonhospital site. Yet another commenter stated that the denominator should be 51 hours, which is derived from the CMS data that is the basis for the RCEs that are currently in use. This commenter noted that if a proxy is being used for both the numerator and denominator, then there is no need to use hours at all. Instead, the formula can be simplified by using a single percentage proxy of the time the physician spends teaching. The commenter thought the formula should be:

Physician compensation proxy using RCEs

× Percentage of business days in year when resident is at site

× Percentage of presumed training time [number of proxy hours/51 hours based on RCEs]

= Physician compensation attributable to training.

Response: We appreciate the commenters' proposals for alternatives to use in the denominator of the ratio that represents the percentage of time the teaching physician spends in nonpatient care direct GME activities. The suggestion to use national average proxies for total physician work hours per week is an interesting idea that we will explore more fully and consider for future rulemaking. In particular, we would like to evaluate thoroughly the alternative data sources that are available, and the ramifications of using specialty-specific proxy data. We expect to investigate this issue, and if appropriate, may propose to use specialty-specific data for physician work hours in the future. We are also not adopting the commenter's suggestion to make adjustments to recognize the number of hours the specific teaching physician works each week as the denominator in the ratio. We believe the relevant figure for this purpose is the time the teaching physician spends in the specific nonhospital site, not the time the physician works elsewhere. Furthermore, if we were to allow for adjustments when the teaching physician spends only a portion of his or her time at the nonhospital site as the commenter recommended, the result might be a physician salary percentage that is much higher than the percentage that would result from use of the number of hours the nonhospital site is open in the denominator. For example, if a teaching physician works a total of 60 hours per week, spending 30 hours in the hospital and 30 hours in the nonhospital site, but the nonhospital site is open 40 hours a week, then the teaching physician cost ratio (to be applied to the survey-based physician salary proxy) would be 3/30, or 10 percent under the commenter's suggestion, and 3/40, or 7.5 percent under our proposal. Accordingly, as we stated in the proposed rule, we believe that deriving the true amount of time spent by each teaching physician in each nonhospital site in nonpatient care direct GME activities would involve the imposition of another form of the documentation burden that the hospital industry and teaching physicians have found onerous up to this point. Therefore, we are finalizing our proposal to use the number of hours a nonhospital site is open each week as the denominator in the ratio for calculating the teaching physician cost ratio.

We are also confirming that in determining the number of hours a clinic is open per week, we do not mean the actual hours the nonhospital site is open per week, but instead, we mean “posted” or advertised hours. Therefore, the fact that a nonhospital site might be closed several days in a year on legal holidays, for example, would not affect the denominator. That is, if a nonhospital site's posted hours are 9 a.m. to 5 p.m. from Monday through Friday, then the denominator would be 40 hours, even if that site was closed for a day(s) for a holiday or some other reason. The hospital may obtain the nonhospital site's posted or advertised hours of operation as documentation to support the number of hours used in the denominator of the teaching time proxy.

Comment: Commenters stated that a reasonable and easy way to administer the supervisory teaching physician cost ratio would be to use 2 hours in the numerator, as supported by the conclusions generated by the professor from Hunter College, and 55 hours in the denominator, which would result in a “maximum fixed ratio” of 3.6 percent. Alternatively, if we reject that suggestion, the commenters urged CMS to adopt a ratio “cap”. The commenters noted that we solicited comments on how to address situations in which that ratio “could, in some extreme instances, result in a determination of unusually high teaching costs” in instances where the nonhospital site is open very few hours per week (72 FR 4827). One commenter suggested that this ratio “cap” should be 5 percent, and would prevent any extreme or atypical results in determining the portion of teaching physicians' salaries attributable to direct GME. Another commenter recommended that the proxy for determining teaching costs be capped at 3 percent, which would be the result of using 2 hours in the numerator (as suggested by the professor from Hunter College's analysis), and 60 hours in the denominator, since 60 hours is the amount of time a typical teaching physician works (in total, in all settings) per week.

Response: As we explained in response to other comments, we believe it is appropriate at this time to finalize our proposals to use 3 hours in the numerator and the number of hours the nonhospital site is open each week in the denominator. However, the commenter is correct that we solicited comments on how to address situations in which that ratio “could, in some extreme instances, result in a determination of unusually high teaching costs” in instances where the nonhospital site is open very few hours per week (72 FR 4827). We believe that in light of these extreme circumstances, the commenters' suggestion to establish a “cap” on the ratio is reasonable. We are not adopting the commenters' suggested cap of 3 percent or 5 percent, since both of these caps are based on using 2 hours in the numerator. Since we are finalizing our proposal to use 3 hours in the numerator, we believe an appropriate cap would be 7.5 percent, which would result from using 3 hours in the numerator and 40 hours in the denominator. We believe it is Start Printed Page 26966appropriate to use 40 hours in the denominator because 40 hours is an established, universally recognized, typical work week. However, we may reevaluate this cap in the context of other possible changes we may consider making to the teaching physician cost ratio. Thus, in this final rule, we are instituting a cap of 7.5 percent on the teaching physician cost ratio, such that a hospital need not employ more than 7.5 percent of the teaching physician cost in calculating the amount of payment necessary to meet the 90 percent threshold. However, in adopting this policy, we note that application of the 7.5 percent cap must always be after a hospital prorates the teaching physician cost to reflect the amount of FTE time that the residents are in the particular nonhospital site per year. Since half-day rotations appear to be a common model of nonhospital training, which would already reduce the ratio well below 7.5 percent, we anticipate that the cap will only be applicable in the extreme circumstances we mentioned when soliciting comments, and which were of concern to the commenters.

Comment: One commenter referred to a letter received from CMS in which CMS stated that the cost of training a resident in a non-hospital setting is based on the “percentage of time” the teaching physician spends in GME activities. Therefore, the commenter asserted, if the hospital is paying for all of the costs of the resident, and the physician can attest that the percentage of time spent in nonpatient care direct GME activities is only 10 percent or less (that is, the remainder of the costs of the program), then the test of a hospital incurring “all or substantially all” of the costs of training the resident should be met.

Response: CMS's policy for determining the costs of nonpatient care direct GME activities of the teaching physician is, indeed, based on the “percentage of time” that the teaching physician spends in such activities. We most recently explained this policy explicitly in the April 2005 Qs&As. In response to Question 5, we stated “Determination of the teaching physician costs to the nonhospital site is dependent upon the teaching physician's salary and the percentage of time he/she devotes to activities related to non-billable direct GME activities at the nonhospital site.” [see http://www.cms.hhs.gov/​AcuteInpatientPPS/​Downloads/​nonhospQA.pdf] As we have stated in those Qs&As, and in this rule, the statutory test is tied to whether the hospital has paid “all or substantially all” of the costs of the training program, and not to how much time the teaching physician spends in nonpatient care direct GME activities, although that time percentage is certainly necessary for determining the amount of the cost that the hospital must pay. Accordingly, the revised policy is consistent with the previous policy in that the hospital must establish the percentage of time spent by the teaching physician in nonpatient care direct GME activities in order to determine the cost of the teaching physician's GME time. However, the revised policy allows for the use of proxies in order to make those calculations. That is, the ratio of 3 hours of nonpatient care direct GME time per week to the number of hours that the nonhospital site is open also represents the percentage of time the teaching physician spends in nonpatient care direct GME activities, and when applied to the physician's salary (as established using survey data), will result in a proxy for the teaching physician cost. As mentioned in the preceding summary of comments, commenters requested that CMS place a cap on the percentage of the teaching physician's time spent in nonpatient care direct GME activities, as determined using the ratio. As explained above, in this final rule, we are instituting a cap of 7.5 percent on this teaching physician cost ratio, which is less than the 10 percent to which this commenter requested that physicians be allowed to attest. Furthermore, we do not believe it is appropriate to say that a hospital has met the test of incurring “all or substantially all” of the costs based simply on a physician's attestation that 10 percent or less of his or her time is spent on nonpatient care direct GME activities. Again, it is the cost that is important, not the amount of the teaching physicians' time.

Comment: We received several comments relating to the method for computing the teaching physician cost in instances where the residents rotate to multiple nonhospital sites for varying periods of time, and whether prorating is applicable. The commenters explained that typically, nonhospital rotations consist of partial day rotations, which can be either partial days or partial weeks, to 3 or 4 different nonhospital sites per week. The commenters mentioned that continuity clinics, which are required for internal medicine residents, are generally rotations of one half-day per week to a specific nonhospital site over the 3-year internal medicine program. The residents may also rotate to other nonhospital sites during each week. The commenters asserted that if hospitals were to assume 3 hours of supervisory teaching physician time for each clinic during a week, the estimate of teaching physician costs would be “severely inflated,” and the hospital would be “paying several times over for training costs incurred during the same time period.”

One commenter noted that we mention the issue of prorating in instances where the residents are not rotating to the nonhospital site for a whole year. Specifically, the preamble states, “If FTE residents are not rotating to a particular nonhospital site throughout a whole year, then the national average salary of the teaching physician would be prorated accordingly. The cost of the residents' salaries and fringe benefits (including travel and lodging where applicable) would already be reflective of an FTE count (72 FR 4822).” In addition, the preamble stated in the context of the teaching physician cost ratio, “For example, if FTE residents rotate throughout the year to a nonhospital site that is open 40 hours per week, then the percentage of time spent by the teaching physician(s) in nonpatient care direct GME activities throughout the year at that site is 3/40 = 0.075 or 7.5 percent. If FTE residents rotate to that nonhospital site for only a portion of a year, then the ratio of 3/40 would be further multiplied by the percentage of the year that the FTE residents train there. For example, if the FTE residents only rotate to this nonhospital site for 3 months of the year, then the percentage of time that the teaching physician(s) spends on nonpatient care direct GME activities at that site equals (3/40 × 0.25 = 0.019 or 1.9 percent)” (72 FR 4827). The commenter continued that although the concept of prorating is supported by the preamble, in discussions with CMS staff, it seems that we intended to allow prorating “selectively.” The commenter stated that their understanding of our position is that if a resident rotates to a nonhospital site for several days each week over a period of time, the resident's salary and fringe benefits would be prorated, but not the physician's salary. The physician's salary would only be prorated if the rotation occurred in a block situation, such as 3 months (in the proposed rule example mentioned above).

The commenter included an addendum which contained examples to illustrate what they believe to be the “flaws” in our position. In the first example, a resident rotates to a nonhospital site for 6 consecutive months, and then spends the rest of the year in a hospital. In the second example, the resident spends 2.5 days a Start Printed Page 26967week at a nonhospital site throughout the entire year (an aggregate time of 6 months), with the remaining time in a hospital setting. In the first example, the commenter understands that we would prorate by 0.5 the resident's stipends and benefits, as well as the physician's salary. In the second example, the commenter understands that we would only prorate the resident's stipends and fringe benefits. The commenter stated that the result is that even though “in the aggregate” the resident spends the same amount of time in the nonhospital site, if he or she rotates in increments of less than a week, the hospital will incur more in supervisory costs. Another commenter believed that there is no basis for distinguishing between these “half-time” rotations, and teaching hospitals should not have to incur any additional costs if the sum of the assignments for the resident on an FTE basis is the same in either case. The former commenter concluded that as long as both the resident and physician salaries are prorated to match the length of time of the rotation, the supervisory cost amount will not be overstated. Alternatively, the commenter noted that the three hour presumption could be prorated, rather than the physician salary, as the result would be the same either way.

Response: In responding to these comments on the issue of prorating, it is important to first understand the context in which we made the decision to propose that 3 hours be used as the proxy for the amount of time a teaching physician spends per week in nonpatient care direct GME activities. As we explained in the proposed rule, we derived the 3 hour figure from informal surveys conducted by the AFMAA and the AOA, which essentially showed ranges of 0 hours to 4.7 hours for the time that physicians spend on nonpatient care direct GME activities (72 FR 4826). Although we acknowledge that the surveys were not rigorous, we believed (and still believe) the survey data warrant the use of 3 hours, and not a lower number, as a proxy in determining the costs hospitals must pay in accordance with the statute. This is especially so since, as explained above, the 3 hour figure is subject to prorating based upon the proportion of time residents are present in the nonhospital site. If “half day” rotations to nonhospital sites are a very common training model as the commen