[Federal Register: July 31, 1998 (Volume 63, Number 147)] [Rules and Regulations] [Page 40953-41002] From the Federal Register Online via GPO Access [wais.access.gpo.gov] [DOCID:fr31jy98-25] [[Page 40953]] _______________________________________________________________________ Part II Department of Health and Human Services _______________________________________________________________________ Health Care Financing Administration _______________________________________________________________________ 42 CFR Parts 405, 412, and 413 Medicare Program: Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 1999 Rates; Final Rule [[Page 40954]] DEPARTMENT OF HEALTH AND HUMAN SERVICES Health Care Financing Administration 42 CFR Parts 405, 412, and 413 [HCFA-1003-F] RIN 0938-AI22 Medicare Program; Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 1999 Rates AGENCY: Health Care Financing Administration (HCFA), HHS. ACTION: Final rule. ----------------------------------------------------------------------- SUMMARY: We are revising the Medicare hospital inpatient prospective payment systems for operating costs and capital-related costs to implement applicable statutory requirements, including section 4407 of the Balanced Budget Act of 1997 (BBA), as well as changes arising from our continuing experience with the systems. In addition, in the addendum to this final rule, we describe changes in the amounts and factors necessary to determine rates for Medicare hospital inpatient services for operating costs and capital-related costs. These changes are applicable to discharges occurring on or after October 1, 1998. We also set forth rate-of-increase limits as well as changes for hospitals and hospital units excluded from the prospective payment systems. Finally, we are implementing the provisions of section 4625 of the BBA concerning payment for the direct costs of graduate medical education. DATES: The provisions of this final rule are effective October 1, 1998. This rule is a major rule as defined in Title 5, United States Code, section 804(2). Pursuant to 5 U.S.C. section 801(a)(1)(A), we are submitting a report to the Congress on this rule on July 31, 1998. ADDRESSES: Copies: To order copies of the Federal Register containing this document, send your request to: New Orders, Superintendent of Documents, P.O. Box 371954, Pittsburgh, PA 15250-7954. Specify the date of the issue requested and enclose a check or money order payable to the Superintendent of Documents, or enclose your Visa or Master Card number and expiration date. Credit card orders can also be placed by calling the order desk at (202) 512-1800 or by faxing to (202) 512- 2250. The cost for each copy is $8.00. As an alternative, you can view and photocopy the Federal Register document at most libraries designated as Federal Depository Libraries and at many other public and academic libraries throughout the country that receive the Federal Register. This Federal Register document is also available from the Federal Register online database through GPO Access, a service of the U.S. Government Printing Office. Free public access is available on a Wide Area Information Server (WAIS) through the Internet and via asynchronous dial-in. Internet users can access the database by using the World Wide Web; the Superintendent of Documents home page address is http://www.access.gpo.gov/su__docs/, by using local WAIS client software, or by telnet to swais.access.gpo.gov, then login as guest (no password required). Dial-in users should use communications software and modem to call (202) 512-1661; type swais, then login as guest (no password required). FOR FURTHER INFORMATION CONTACT: Nancy Edwards, (410) 786-4531, Operating Prospective Payment, DRG, and Wage Index Issues. Tzvi Hefter, (410) 786-4487, Capital Prospective Payment, Excluded Hospitals, and Graduate Medical Education Issues. SUPPLEMENTARY INFORMATION: I. Background A. Summary Sections 1886(d) and (g) of the Social Security Act (the Act) set forth a system of payment for the operating and capital costs of acute care hospital inpatient stays under Medicare Part A (Hospital Insurance) based on prospectively-set rates. Under these prospective payment systems (PPS), Medicare payment for hospital inpatient operating and capital-related costs is made at predetermined, specific rates for each hospital discharge. Discharges are classified according to a list of diagnosis-related groups (DRGs). Certain specialty hospitals are excluded from the prospective payment systems. Under section 1886(d)(1)(B) of the Act, the following hospitals and units are excluded from PPS: psychiatric hospitals or units, rehabilitation hospitals or units, children's hospitals, long term care hospitals, and cancer hospitals. For these hospitals and units, Medicare payment for operating costs is based on reasonable costs subject to certain limits. Under section 1886(a)(4) of the Act, costs incurred in connection with approved graduate medical education (GME) programs are excluded from the operating costs of inpatient hospital services. Hospitals with approved GME programs are paid for the direct costs of GME in accordance with section 1886(h) of the Act; the amount of payment for direct GME costs for a cost reporting period is based on the number of the hospital's residents in that period and the hospital's costs per resident in a base year. The regulations governing the hospital inpatient prospective payment system are located in 42 CFR part 412. The regulations governing excluded hospitals are located in both parts 412 and 413, and the graduate medical education regulations are found in part 413. B. Summary of the Provisions of the May 8, 1998 Proposed Rule On May 8, 1998, we published a proposed rule in the Federal Register (63 FR 25576) setting forth proposed changes to the Medicare hospital inpatient prospective payment systems for both operating costs and capital-related costs, which would be effective for discharges occurring on or after October 1, 1998. We also proposed changes in payments for excluded hospitals and payments for graduate medical education costs. The following is a summary of the major issues addressed and changes we proposed to make: We proposed changes to the FY 1999 DRG classifications and relative weights, as required by section 1886(d)(4)(C) of the Act. We proposed to update the hospital wage data for FY 1999. We also proposed changes to the data categories included in the wage index and revisions to the wage index based on hospital redesignations. We discussed several provisions of the regulations in 42 CFR parts 412 and 413 and set forth certain proposed changes concerning definition of transfer cases, rural referral centers, disproportionate share adjustment, bad debts, and direct graduate medical education programs. We discussed several provisions of the regulations in 42 CFR Part 412 and set forth certain proposed changes and clarifications concerning capital indirect medical education payments and payments to new hospitals. We discussed the criteria governing excluded hospitals including caps on the target amounts for FY 1999 and exceptions. In the addendum to the proposed rule, we set forth proposed changes to the amounts and factors for determining the FY 1999 prospective payment rates for operating costs and capital-related costs. We also proposed update factors for determining the rate-of- increase limits for cost reporting periods [[Page 40955]] beginning in FY 1999 for hospitals and hospital units excluded from the prospective payment system. In Appendix A of the proposed rule, we set forth an analysis of the impact that the proposed changes would have on affected entities. In Appendix B of the proposed rule, we set forth the technical appendix on the proposed FY 1999 capital cost model. In Appendix C, as required by section 1886(e)(3)(B) of the Act, we set forth a report to Congress on our initial estimate of a recommended update factor for FY 1999 for both hospitals included in and hospitals excluded from the prospective payment systems. In Appendix D of the proposed rule, we set forth our recommendation of the appropriate percentage change for FY 1999 for the large urban area and other area average standardized amounts (and hospital-specific rates applicable to sole community and Medicare- dependent, small rural hospitals) for hospital inpatient services paid for under the prospective payment system for operating costs. In Appendix D of the proposed rule, we also set forth our recommendation of the appropriate percentage change for FY 1999 for target rate-of-increase limits to the allowable operating costs of hospital inpatient services furnished by hospitals and hospital units excluded from the prospective payment system. In the proposed rule, we discussed in detail the March 1, 1998 recommendations concerning hospital inpatient policies made by the Medicare Payment Advisory Commission (MedPAC) as well as our responses to those recommendations. Under section 1805(b) of the Act, MedPAC is required to submit a report to Congress, not later than March 1 of each year, that reviews and makes recommendations on Medicare payment policies. C. Public Comments Received in Response to the Proposed Rule A total of 214 items of correspondence containing comments on the proposed rule were received timely. The main areas of concern addressed by the commenters were the change in the definition of transfer cases and the revisions to the wage index. We also received a number of comments on the proposal to pay qualified nonhospital providers for the direct costs of graduate medical education. Summaries of the public comments received and our responses to those comments are set forth below under the appropriate section. II. Changes to DRG Classifications and Relative Weights A. Background Under the prospective payment system, we pay for inpatient hospital services on the basis of a rate per discharge that varies by the DRG to which a beneficiary's stay is assigned. The formula used to calculate payment for a specific case takes an individual hospital's payment rate per case and multiplies it by the weight of the DRG to which the case is assigned. Each DRG weight represents the average resources required to care for cases in that particular DRG relative to the average resources used to treat cases in all DRGs. Congress recognized that it would be necessary to recalculate the DRG relative weights periodically to account for changes in resource consumption. Accordingly, section 1886(d)(4)(C) of the Act requires that the Secretary adjust the DRG classifications and relative weights annually. These adjustments are made to reflect changes in treatment patterns, technology, and any other factors that may change the relative use of hospital resources. The changes to the DRG classification system and the recalibration of the DRG weights for discharges occurring on or after October 1, 1998 are discussed below. B. DRG Reclassification 1. General Cases are classified into DRGs for payment under the prospective payment system based on the principal diagnosis, up to eight additional diagnoses, and up to six procedures performed during the stay, as well as age, sex, and discharge status of the patient. The diagnosis and procedure information is reported by the hospital using codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). The Medicare fiscal intermediary enters the information into its claims system and subjects it to a series of automated screens called the Medicare Code Editor (MCE). These screens are designed to identify cases that require further review before classification into a DRG can be accomplished. After screening through the MCE and any further development of the claims, cases are classified by the GROUPER software program into the appropriate DRG. The GROUPER program was developed as a means of classifying each case into a DRG on the basis of the diagnosis and procedure codes and demographic information (that is, sex, age, and discharge status). It is used both to classify past cases in order 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 Medicare Provider Analysis and Review (MedPAR) file. The data in this file are used to evaluate possible DRG classification changes and to recalibrate the DRG weights. Currently, cases are assigned to one of 496 DRGs in 25 major diagnostic categories (MDCs). Most MDCs are based on a particular organ system of the body (for example, MDC 6, Diseases and Disorders of the Digestive System); however, some MDCs are not constructed on this basis since they involve multiple organ systems (for example, MDC 22, Burns). In general, cases are assigned to an MDC based on the principal diagnosis, before assignment to a DRG. However, there are five DRGs to which cases are directly assigned on the basis of procedure codes. These are the DRGs for liver, bone marrow, and lung transplant (DRGs 480, 481, and 495, respectively) and the two DRGs for tracheostomies (DRGs 482 and 483). Cases are assigned to these DRGs before classification to an MDC. Within most MDCs, cases are then divided into surgical DRGs (based on a surgical hierarchy that orders individual procedures or groups of procedures by resource intensity) and medical DRGs. Medical DRGs generally are differentiated on the basis of diagnosis and age. Some surgical and medical DRGs are further differentiated based on the presence or absence of complications or comorbidities (hereafter CC). Generally, GROUPER does not consider other procedures; that is, nonsurgical procedures or minor surgical procedures generally not performed in an operating room are not listed as operating room (OR) procedures in the GROUPER decision tables. However, there are a few non-OR procedures that do affect DRG assignment for certain principal diagnoses, such as extracorporeal shock wave lithotripsy for patients with a principal diagnosis of urinary stones. We proposed several changes to the DRG classification system for FY 1999. The proposed changes, the comments we received concerning them, our responses to those comments, and the final DRG changes are set forth below. Unless otherwise noted, our DRG analysis is based on the full (100 percent) FY 1997 MedPAR file based on [[Page 40956]] bills received through September 30, 1997. 2. MDC 5 (Diseases and Disorders of the Circulatory System) In the August 29, 1997 hospital inpatient final rule with comment period (62 FR 45974), we noted that, because of the many recent changes in heart surgery, we were considering conducting a comprehensive review of the MDC 5 surgical DRGs. We have begun that review, and based upon our analysis thus far, we proposed the following DRG changes. a. Coronary Bypass. There are two DRGs that capture coronary bypass procedures: DRG 106 (Coronary Bypass with Cardiac Catheterization) and DRG 107 (Coronary Bypass without Cardiac Catheterization). The procedures that allow a coronary bypass case to be assigned to DRG 106 include percutaneous valvuloplasty, percutaneous transluminal coronary angioplasty (PTCA), cardiac catheterization, coronary angiography, and arteriography. In analyzing the FY 1997 MedPAR file, we noted that, of cases assigned to DRG 106, the average standardized charges for coronary bypass cases with PTCA were significantly higher than those cases without PTCA. There were approximately 4,400 cases in DRG 106 where PTCA is performed as a secondary procedure. These cases had an average standardized charge of approximately $69,000. The average charge of the approximately 95,000 cases in DRG 106 without PTCA was approximately $52,000. Based on this analysis, we proposed to create a new DRG for coronary bypass cases with PTCA. The cases currently in DRG 106 without PTCA would be assigned to another DRG and the cases currently assigned to DRG 107 would be unmodified. Because we would replace two DRGs with three new DRGs, we proposed to revise the DRG numbers and titles accordingly. The new DRGs and their titles are set forth below: DRG 106 Coronary Bypass with PTCA DRG 107 Coronary Bypass with Cardiac Catheterization DRG 109 Coronary Bypass without Cardiac Catheterization We note that DRG 109 has been an empty DRG for the last several years. We received several comments regarding this proposal. Comment: While the commenters supported the creation of a new DRG to capture coronary bypass surgeries with PTCA, some of the commenters were concerned about the renumbering of the current DRGs 106 and 107. They believe splitting the cases currently assigned to DRG 106 into new DRGs 106 and 107 and reassigning the cases currently assigned to DRG 107 to DRG 109 will make it difficult to conduct DRG trend analyses. The commenters suggested that DRGs 106 and 107 should not be modified and that DRG 109 be used to capture coronary bypass with PTCA. Two commenters stated that a DRG that has been invalidated (109) should not be reintroduced. Response: Although we understand the commenters' concern, we also believe that the sequencing of surgical DRGs in hierarchy order is appropriate. In this case, our alternative to the proposed revision would have been to delete DRGs 106 and 107 and create three new DRGs that would have been placed at the end of the DRG table, that is, after current DRG 503. Because we did have an empty surgical DRG in MDC 5 and it was numerically close to DRGs 106 and 107, we believed our proposed retitling was the best alternative. We note that the surgical DRGs in MDC 5 have been renumbered and retitled several times since they were first introduced in 1983. As stated above, we are currently conducting a comprehensive review of the MDC 5 surgical DRGs. If that review results in the reclassification of procedures among the current DRGs, we will probably renumber and retitle those DRGs. Comment: We received one comment requesting clarification of the DRG assignment for PTCA and cardiac catheterization procedures when performed in conjunction with coronary bypass. The commenter suggested that we add the phrase ``without PTCA'' to the titles of DRGs 107 and 109 to more aptly describe the cases assigned to those DRGs. Response: Coronary bypass performed in conjunction with single or multiple PTCA or percutaneous valvuloplasty will be assigned to DRG 106. The procedure codes for PTCA and percutaneous valvuloplasty are as follows: 35.96, 36.01, 36.02, and 36.05. Procedures assigned to DRG 107 would include any coronary bypass with cardiac catheterization, coronary angiography, or coronary arteriography, and DRG 109 is for cases with the coronary bypass procedure only. We believe that the proposed titles accurately describe the cases assigned to each of the DRGs and that adding the phrase ``without PTCA'' to the titles of DRGs 107 and 109 is unnecessary. We are incorporating our proposed DRG changes and DRG numbers and titles in the final DRG classifications. b. Implantable heart assist system and annuloplasty. In the August 29, 1997 final rule with comment period, we moved implant of an implantable, pulsatile heart assist system (procedure code 37.66) from DRGs 110 and 111 (Major Cardiovascular Procedures) 1 to DRG 108 (Other Cardiothoracic Procedures). Although this move improved payment for these procedures, they were still much more expensive than the other cases in DRG 108 ($96,000 for heart assist versus an average of $54,000 for all other cases in the FY 1996 MedPAR file). We stated that we would continue to review the MDC 5 surgical DRGs in an attempt to find a DRG placement for these cases that would be more similar in terms of resource use. --------------------------------------------------------------------------- \1\ A single title combined with two DRG numbers is used to signify pairs. Generally, the first DRG is for cases with CC and the second DRG is for cases without CC. If a third number is included, it represents cases with patients who are age 0-17. Occasionally, a pair of DRGs is split between age >17 and age 0-17. --------------------------------------------------------------------------- As discussed in the proposed rule, in reviewing the FY 1997 MedPAR file, we noted that heart assist system implant continues to be the most expensive procedure in DRG 108. In fact, other than heart transplant, heart assist system implant is the most expensive procedure in MDC 5. The average FY 1997 charge for these cases, when assigned to DRG 108, is over $150,000 compared to about $53,000 for all cases in DRG 108. Obviously, the charges for heart assist implant are increasing at a much greater rate than the average charges for DRG 108. In addition, the length of stay for cases coded with 37.66 is approximately 32 days compared to about 11 days for all other DRG 108 cases. One possibility for improving payment for these cases is to move them to DRGs 104 and 105 (Cardiac Valve Procedures). Those DRGs, which split on the basis of the performance of cardiac catheterization, have average charges of approximately $66,000 and $51,000, respectively. While heart assist implant cases are still more expensive than the average case in these DRGs, payment would be improved. Clinically, placement of heart assist implant in DRGs 104 and 105 is not without precedent. Effective with FY 1988, we placed implant of a total automatic implantable cardioverter defibrillator (AICD) in these DRGs. In addition, the vast majority of procedures assigned to DRG 108 involve surgically splitting open the sternum to perform the procedure. However, implant of the heart assist device does not require this approach. While reviewing the DRG 108 cases, we also noted that procedure code 35.33 [[Page 40957]] (annuloplasty) is assigned to this DRG. Annuloplasty is a valve procedure and is clinically more similar to the cases assigned to DRGs 104 and 105 than it is to the cases assigned to DRG 108. In addition, the average standardized charge for annuloplasty cases assigned to DRG 108 is about $67,000, well above the overall average charge of approximately $53,000 for cases in DRG 108. Therefore, we proposed to move annuloplasty from DRG 108 to DRGs 104 and 105. In order to more accurately reflect the cases assigned to DRGs 104 and 105, we proposed to retitle them as follows: DRG 104 Cardiac Valve and Other Major Cardiothoracic Procedures with Cardiac Catheterization DRG 105 Cardiac Valve and Other Major Cardiothoracic Procedures without Cardiac Catheterization. We received only supportive comments for our proposal to move annuloplasty to DRGs 104 and 105; therefore, that change is included in the final DRGs. Comment: Commenters generally appreciated any improvement in the payment for heart assist devices. However, some of them continue to urge HCFA to reclassify these cases to DRG 103 (Heart Transplant) or to their own DRG. Two commenters were unsure if we had proposed a classification change which was reflected in the proposed DRG weights or had merely requested comment on such a change. Another commenter was concerned that cases reassigned to DRG 105 (those in which there is no cardiac catheterization performed) would receive a lower payment than they currently do in DRG 108. Response: First, we note that the proposed DRG weights did include this change; that is, we moved over 2,000 heart assist implant cases from DRG 108 to DRGs 104 and 105 before recalibrating the proposed weights. In addition, although the final FY 1999 weight for DRG 105 is slightly lower than the weight for DRG 108 (5.7099 and 5.9764, respectively), the much higher DRG 104 weight (7.3690) results in an overall improvement in payment for these cases when reclassified. Using the FY 1997 MedPAR cases, we estimate that at least 40 percent of the heart assist implant cases will be assigned to DRG 104. Thus, as long as a hospital treats a mix of heart assist implant cases, with and without the cardiac catheterization procedure, its overall payment should be higher under the revised classification. We presume this will be the case for virtually all hospitals. With regard to the comments concerning reclassification of this procedure to DRG 103 or a new DRG, we refer the reader to our response to a similar comment in the August 29, 1997 final rule (62 FR 45967). 3. MDC 22 (Burns) Under the current DRG system, burn cases are assigned to one of six DRGs in MDC 22 (Burns), which have not been revised since 1986. In our FY 1998 hospital inpatient proposed rule (June 2, 1997; 62 FR 29912), in response to inquiries we had received, we indicated that we would conduct a comprehensive review of MDC 22 to determine whether changes in these DRGs could more appropriately capture the variation in resource use associated with different classes of burn patients. We solicited public comments on this issue, particularly asking for recommendations on ways to categorize related diagnosis and procedure codes to produce DRG groupings that would be more homogeneous in terms of resource use. In our May 8, 1998 proposed rule (63 FR 25579), we discussed in detail the results of our review of MDC 22. We received a proposal (endorsed by the American Burn Association (ABA)) for restructuring the DRGs based on several statistical and clinical criteria, including age, severity of the burn, and the presence of complications or comorbidities. Subsequently, we worked closely with representatives of the ABA and with the clinicians who developed the proposal in order to refine it for Medicare purposes. Based on this work, we proposed to replace the six existing DRGs in MDC 22 with eight new DRGs. For ease of reference and classification, the current DRGs in MDC 22, DRGs 456 through 460 and 472, would no longer be valid, and we would establish new DRGs 504 through 511 to contain all cases that currently group to MDC 22. (The complete titles of the new DRGs are set forth below.) In reviewing the Medicare burn cases, we found that the most important distinguishing characteristic in terms of resource use was the amount of body surface affected by the burn and how much of that burn was a 3rd degree burn. The second most important factor was whether or not the patient received a skin graft. Thus, a patient with burns covering at least 20 percent of body area, with at least 10 percent of that a 3rd degree burn, consumed the most resources. However, if a patient met these criteria and did not receive a skin graft, then the case was much less expensive and the average length of stay fell from over 30 days to 8 days. The first two proposed burn DRGs reflect these distinctions (DRGs 504 and 505). After classifying the most extensive burn cases, we found that the patients with 3rd degree burns that did not meet the criteria to be assigned to DRGs 504 and 505 were the most expensive of the remaining cases (that is, those patients whose burns did not meet the at least 20 percent body area or at least 10 percent 3rd degree criteria). These burns are referred to clinically as ``full-thickness burns.'' A subset of these full-thickness burn cases, those with skin graft or an inhalation injury, were much more expensive than the other cases. After dividing these patients into two groups, with or without skin graft or inhalation injury, we examined whether other factors had an influence on resource use. We found that patients who had a CC (complication or comorbidity) or a concomitant significant trauma consumed more resources whether or not they had a skin graft or inhalation injury. Thus, the next four proposed DRGs were defined as full-thickness burns with skin graft or inhalation injury with or without CC or significant trauma, or full-thickness burns without skin graft or inhalation injury with or without CC or significant trauma (DRGs 506 through 509). Finally, the last two proposed DRGs (510 and 511) were for cases with nonextensive burns. These cases are also split on the basis of CCs or concomitant significant trauma. Consistent with the recommendations of several commenters on last year's proposed rule, the new burn DRGs would no longer include a separate DRG for cases in which burn patients were transferred to another acute care facility. The specific diagnosis and procedure codes that were included in each of the eight proposed DRGs and their titles are as follows. DRGs 504 and 505--Extensive 3rd Degree Burns with and without Skin Graft. DRGs 504 and 505 include all cases with burns involving at least 20 percent of body surface area combined with a 3rd degree burn covering at least 10 percent of body surface area. Thus, these cases have diagnosis codes of 948.xx, with a fourth digit of 2 or higher (indicating that burn extends over 20 percent or more of body surface) and a fifth digit of 1 or higher (indicating a 3rd degree burn extending over 10 percent or more of body surface). Cases with the appropriate diagnosis codes are classified into DRG 504 if one of the following skin graft procedure codes is present: 85.82 Split-thickness graft to breast 85.83 Full-thickness graft to breast 85.84 Pedicle graft to breast 86.60 Free skin graft, NOS 86.61 Full-thickness skin graft to hand [[Page 40958]] 86.62 Other skin graft to hand 86.63 Full-thickness skin graft to other sites 86.65 Heterograft to skin 86.66 Homograft to skin 86.67 Dermal regenerative graft (new code in FY 1999--see Table 6A in section VI. of the Addendum) 86.69 Other skin graft to other sites 86.70 Pedicle of flap graft, NOS 86.71 Cutting and preparation of pedicle grafts or flaps 86.72 Advancement of pedicle graft 86.73 Attachment of pedicle or flap graft to hand 86.74 Attachment of pedicle or flap graft to other sites 86.75 Revision of pedicle or flap graft 86.93 Insertion of tissue expander DRGs 506 and 507--Full Thickness Burn with Skin Graft or Inhalation Injury with or without CC or Significant Trauma. These DRGs include all other cases of 3rd degree burns that also have either a skin graft or an inhalation injury. Thus, these cases have diagnosis codes of 941.xx through 946.xx, and 949.xx, with a fourth digit of 3 or higher, as well as cases with codes of 948.xx that did not group into DRGs 504 or 505 (that is, 948.00, 948.01, and 948.1x through 948.9x with a fifth digit of 0). In addition, cases classified into DRGs 506 and 507 must have either one of the skin graft procedure codes listed above or one of the following diagnosis codes for inhalation injuries: 518.5 Pulmonary insufficiency following trauma and surgery 518.81 Respiratory failure 518.84 Acute and chronic respiratory failure (new code in FY 1999-- see Table 6A in section VI. of the Addendum) 947.1 Burn of larynx, trachea, or lung 987.9 Toxic effect of gas, fume, or vapor, NOS Cases that meet both of these coding criteria are assigned to DRG 506 if there is a diagnosis code indicating either a CC (based on the standard DRG CC list) or concomitant significant trauma (based on the significant trauma diagnosis codes, listed by body site, used for classification in MDC 24). DRGs 508 and 509--Full Thickness Burn without Skin Graft or Inhalation Injury with or without CC or Significant Trauma. These DRGs include all other cases of 3rd degree burns. Thus, these DRGs include all cases without a skin graft or inhalation injury that have diagnosis codes of 941.xx through 946.xx, and 949.xx, with a fourth digit of 3 or higher, as well as cases with codes of 948.xx that did not group into DRGs 504 or 505. DRG 508 also requires a secondary diagnosis from the standard CC list or the trauma list based on the significant trauma diagnosis codes, listed by body site, used for classification in MDC 24. DRGs 510 and 511--Nonextensive Burns with and without CC or Significant Trauma. The remaining burn cases would be classified into one of these two proposed DRGs, depending on whether or not the claim included a diagnosis code reflecting the presence of a CC or a significant trauma, as explained above. Comment: We received five comments on this proposed change. In general, the commenters, including the ABA, strongly supported the proposed restructuring of MDC 22. The commenters agreed that the new burn DRGs should bring about meaningful improvements to the clinical coherency and payment equity for the cases assigned to the MDC 22 DRGs. One commenter noted that under the new DRGs, diagnosis codes in the 948.xx series (that is, the codes used to identify the extent of body surface involved in a burn and the percentage of the body surface with a 3rd degree burn) would take on added importance and emphasized the need for coder education in this area. Another commenter submitted several suggestions for additional procedure codes that should be added to the list of procedure codes that would result in assignment to DRG 504 and to DRGs 506 and 507. These codes include both additional codes that the commenter believes should be considered as skin grafts (such as procedure codes 08.61 through 08.69, reconstruction of eyelid with flaps or grafts) as well as codes for other procedures (for example, limb reattachments or eyeball enucleations) that, as the commenter pointed out, are now considered a related operating room procedure under existing DRG 472, Extensive Burns with Operating Room Procedure. This commenter also suggested that DRGs 506 and 507 be identified as surgical DRGs in Table 5 of the addendum to the final rule. Response: We appreciate the positive responses generated by this proposal. We agree that our proposed changes will place greater emphasis on the need for accurate use of the series 948.xx diagnosis codes. We note that this issue has been addressed in the American Hospital Association's quarterly publication, ``Coding Clinic for ICD- 9-CM.'' In the 1994, 4th quarter issue, Coding Clinic stated ``It is advisable to use category 948 as additional coding when needed to provide data for evaluating burn mortality, such as that needed by burn units. It is also advisable to use category 948 as an additional code for reporting purposes when there is mention of a third-degree burn involving 20 percent or more of the body surface.'' We believe the vast majority of burn cases already include the 948.xx coding if appropriate, especially those treated in burn centers. However, we will be pleased to work with other hospital groups that are interested in developing educational materials related to the accurate coding of burn cases. In developing the coding classifications used to assign cases under the burn DRGs, we worked closely with the ABA and its medical consultants to identify the most significant distinguishing characteristics in terms of resource use in burn cases. This process involved both grouping cases that were clinically similar as well as conducting a series of test runs to maximize the amount of variation in resource use that could be explained using varying groups of diagnosis and procedure codes. As stated in the May 8 proposed rule (63 FR 25579), we estimate that the proposed changes to the burn DRGs would increase by more than 25 percent the amount of variation in resource use explained by the DRGs in MDC 22, as well as improve the clinical coherence of the cases within each DRG. As recommended by the ABA, the procedure codes used to identify skin grafts coincide with the procedure codes now in use under existing DRG 458, Non-Extensive Burns with Skin Graft, and we believe that these codes represent the most resource-intensive skin grafts. Therefore, we are not adding the codes suggested by the commenter. We recognize that some procedures now listed under DRG 472 will no longer affect DRG assignment under the restructured burn DRGs. However, we believe that the substantially increased ability of the new DRGs to explain the variation in resource use among burn cases clearly indicates the appropriateness of narrowing the focus of the classification system to emphasize the extent and severity of the burn, in conjunction with skin grafts or inhalation injury. Our analysis indicated that the presence of skin grafts or inhalation injuries had a much more consistent effect on the consumption of hospital resources than the presence of one of the numerous operating room procedures now listed under DRG 472. We also note that, since the skin graft procedures now classified to DRG 504 were classified to former DRG 472, many DRG 472 cases will now be assigned to DRG 504, which has a higher weight than 472 did (14.1153 versus 10.2429). When the FY 1999 cases become available, we will review them to assess the revisions to MDC 22 and the possible need for the type of changes suggested by the commenter. [[Page 40959]] Finally, we note that we do not classify DRGs 506 and 507 as surgical DRGs because they include not only cases involving skin grafts, which are considered surgical procedures, but also cases involving inhalation injuries, which would not necessarily involve any surgical procedures. Thus, in this final rule, we are adopting the changes to the burn DRGs as proposed. 4. Legionnaires' Disease Effective with discharges occurring on or after October 1, 1997, a new diagnosis code was created for pneumonia due to Legionnaires' disease (code 482.84). In the August 29, 1997 final rule with comment period, we assigned this code to DRGs 79, 80, and 81 (Respiratory Infections and Inflammations) (62 FR 46090). However, we did not include this code as a human immunodeficiency virus (HIV) major related condition in MDC 25 (HIV Infections). Because pneumonia due to Legionnaires' disease is a serious respiratory condition that has a deleterious effect on patients with HIV, we proposed to assign diagnosis code 482.84 to DRG 489 (HIV with Major Related Condition) as a major related condition. In addition, we did not assign the code as a major problem in DRGs 387 (Prematurity with Major Problems) and 389 (Full Term Neonate with Major Problems). These DRGs are assigned to MDC 15 (Newborns and Other Neonates with Conditions Originating in the Perinatal Period). Again, as a part of the proposed rule, we assigned diagnosis code 482.84 as a major problem in DRGs 387 and 389 because of its effect on resource use in treating newborns. Commenters supported these proposed revisions, and we are incorporating them into the final DRGs. 5. Surgical Hierarchies Some inpatient stays entail multiple surgical procedures, each one of which, occurring by itself, could result in assignment of the case to a different DRG within the MDC to which the principal diagnosis is assigned. It is, therefore, necessary to have a decision rule by which these cases are assigned to a single DRG. The surgical hierarchy, an ordering of surgical classes from most to least resource intensive, performs that function. Its application ensures that cases involving multiple surgical procedures are assigned to the DRG associated with the most resource-intensive surgical class. Because the relative resource intensity of surgical classes can shift as a function of DRG reclassification and recalibration, we reviewed the surgical hierarchy of each MDC, as we have for previous reclassifications, to determine if the ordering of classes coincided with the intensity of resource utilization, as measured by the same billing data used to compute the DRG relative weights. A surgical class can be composed of one or more DRGs. For example, in MDC 5, the surgical class ``heart transplant'' consists of a single DRG (DRG 103) and the class ``major cardiovascular procedures'' consists of two DRGs (DRGs 110 and 111). Consequently, in many cases, the surgical hierarchy has an impact on more than one DRG. The methodology for determining the most resource-intensive surgical class involves weighing each DRG for frequency to determine the average resources for each surgical class. For example, assume surgical class A includes DRGs 1 and 2 and surgical class B includes DRGs 3, 4, and 5. Assume also that the average charge of DRG 1 is higher than that of DRG 3, but the average charges of DRGs 4 and 5 are higher than the average charge of DRG 2. To determine whether surgical class A should be higher or lower than surgical class B in the surgical hierarchy, we would weigh the average charge of each DRG by frequency (that is, by the number of cases in the DRG) to determine average resource consumption for the surgical class. The surgical classes would then be ordered from the class with the highest average resource utilization to that with the lowest, with the exception of ``other OR procedures'' as discussed below. This methodology may occasionally result in a case involving multiple procedures being assigned to the lower-weighted DRG (in the highest, most resource-intensive surgical class) of the available alternatives. However, given that the logic underlying the surgical hierarchy provides that the GROUPER searches for the procedure in the most resource-intensive surgical class this result is unavoidable. We note that, notwithstanding the foregoing discussion, there are a few instances when a surgical class with a lower average relative weight is ordered above a surgical class with a higher average relative weight. For example, the ``other OR procedures'' surgical class is uniformly ordered last in the surgical hierarchy of each MDC in which it occurs, regardless of the fact that the relative weight for the DRG or DRGs in that surgical class may be higher than that for other surgical classes in the MDC. The ``other OR procedures'' class is a group of procedures that are least likely to be related to the diagnoses in the MDC but are occasionally performed on patients with these diagnoses. Therefore, these procedures should only be considered if no other procedure more closely related to the diagnoses in the MDC has been performed. A second example occurs when the difference between the average weights for two surgical classes is very small. We have found that small differences generally do not warrant reordering of the hierarchy since, by virtue of the hierarchy change, the relative weights are likely to shift such that the higher-ordered surgical class has a lower average weight than the class ordered below it. Based on the preliminary recalibration of the DRGs, we proposed to modify the surgical hierarchy as set forth below. However, in developing the proposed rule, we were unable to test the effects of the proposed revisions to the surgical hierarchy and to reflect these changes in the proposed relative weights due to the unavailability of revised GROUPER software at the time the proposed rule was prepared. Rather, we simulated most major classification changes to approximate the placement of cases under the proposed reclassification and then determined the average charge for each DRG. These average charges then serve as our best estimate of relative resource use for each surgical class. We test the proposed surgical hierarchy changes after the revised GROUPER is received and reflect the final changes in the DRG relative weights in the final rule. We proposed to revise the surgical hierarchy for MDC 3 (Diseases and Disorders of the Ear, Nose, Mouth and Throat) as follows: We would reorder Sinus and Mastoid Procedures (DRGs 53-54) above Myringotomy with Tube Insertion (DRGs 61-62). We would reorder Mouth Procedures (DRGs 168-169) above Tonsil and Adenoid Procedure Except Tonsillectomy and/or Adenoidectomy Only (DRGs 57-58). We received two comments in support of our surgical hierarchy proposals. However, for this final rule, we tested the proposed changes using the most recent MedPAR file and the revised GROUPER software, and we found that the proposal to move Sinus and Mastoid Procedures (DRGs 53-54) above Myringotomy with Tube Insertion (DRGs 61-62) is not supported. Therefore, this change will not be made in this final rule. The proposed reordering of DRGs 53 and 54 above Cleft Lip and Palate Repair (DRG 52) (DRG 52 is currently ordered below DRGs 61 and 62 but above DRGs 53 and 54) is still supported and will be [[Page 40960]] incorporated in the final GROUPER, as will the proposed reordering of DRGs 168 and 169 above DRGs 57 and 58. 6. Refinement of Complications and Comorbidities List There is a standard list of diagnoses that are considered CCs. We developed this list using physician panels to include those diagnoses that, when present as a secondary condition, would be considered a substantial complication or comorbidity. In previous years, we have made changes to the standard list of CCs, either by adding new CCs or deleting CCs already on the list. We did not propose to delete any of the diagnosis codes on the CC list. In the September 1, 1987 final notice concerning changes to the DRG classification system (52 FR 33143), we modified the GROUPER logic so that certain diagnoses included on the standard list of CCs would not be considered a valid CC in combination with a particular principal diagnosis. Thus, we created the CC Exclusions List. We made these changes to preclude coding of CCs for closely related conditions, to preclude duplicative coding or inconsistent coding from being treated as CCs, and to ensure that cases are appropriately classified between the complicated and uncomplicated DRGs in a pair. In the May 19, 1987 proposed notice concerning changes to the DRG classification system (52 FR 18877), we explained that the excluded secondary diagnoses were established using the following five principles: Chronic and acute manifestations of the same condition should not be considered CCs for one another (as subsequently corrected in the September 1, 1987 final notice (52 FR 33154)). Specific and nonspecific (that is, not otherwise specified (NOS)) diagnosis codes for a condition should not be considered CCs for one another. Conditions that may not co-exist, such as partial/total, unilateral/bilateral, obstructed/unobstructed, and benign/malignant, should not be considered CCs for one another. The same condition in anatomically proximal sites should not be considered CCs for one another. Closely related conditions should not be considered CCs for one another. The creation of the CC Exclusions List was a major project involving hundreds of codes. The FY 1988 revisions were intended to be only a first step toward refinement of the CC list in that the criteria used for eliminating certain diagnoses from consideration as CCs were intended to identify only the most obvious diagnoses that should not be considered complications or comorbidities of another diagnosis. For that reason, and in light of comments and questions on the CC list, we have continued to review the remaining CCs to identify additional exclusions and to remove diagnoses from the master list that have been shown not to meet the definition of a CC. (See the September 30, 1988 final rule for the revision made for the discharges occurring in FY 1989 (53 FR 38485); the September 1, 1989 final rule for the FY 1990 revision (54 FR 36552); the September 4, 1990 final rule for the FY 1991 revision (55 FR 36126); the August 30, 1991 final rule for the FY 1992 revision (56 FR 43209); the September 1, 1992 final rule for the FY 1993 revision (57 FR 39753); the September 1, 1993 final rule for the FY 1994 revisions (58 FR 46278); the September 1, 1994 final rule for the FY 1995 revisions (59 FR 45334); the September 1, 1995 final rule for the FY 1996 revisions (60 FR 45782); the August 30, 1996 final rule for the FY 1997 revisions (61 FR 46171); and the August 29, 1997 final rule for the FY 1998 revisions (62 FR 45966)). We proposed a limited revision of the CC Exclusions List to take into account the changes that will be made in the ICD-9-CM diagnosis coding system effective October 1, 1998. (See section II.B.8, below, for a discussion of ICD-9-CM changes.) These proposed changes were made in accordance with the principles established when we created the CC Exclusions List in 1987. We received no comments on these proposed changes and we are incorporating them as final changes. Tables 6F and 6G in section VI of the Addendum to this final rule contain the revisions to the CC Exclusions List that would be effective for discharges occurring on or after October 1, 1998. Each table shows the principal diagnoses with changes to the excluded CCs. Each of these principal diagnoses is shown with an asterisk and the additions or deletions to the CC Exclusions List are provided in an indented column immediately following the affected principal diagnosis. CCs that are added to the list are in Table 6F--Additions to the CC Exclusions List. Beginning with discharges on or after October 1, 1998, the indented diagnoses will not be recognized by the GROUPER as valid CCs for the asterisked principal diagnosis. CCs that are deleted from the list are in Table 6G--Deletions from the CC Exclusions List. Beginning with discharges on or after October 1, 1998 the indented diagnoses will be recognized by the GROUPER as valid CCs for the asterisked principal diagnosis. Copies of the original CC Exclusions List applicable to FY 1988 can be obtained from the National Technical Information Service (NTIS) of the Department of Commerce. It is available in hard copy for $92.00 plus $6.00 shipping and handling and on microfiche for $20.50, plus $4.00 for shipping and handling. A request for the FY 1988 CC Exclusions List (which should include the identification accession number, (PB) 88-133970) should be made to the following address: National Technical Information Service; United States Department of Commerce; 5285 Port Royal Road; Springfield, Virginia 22161; or by calling (703) 487-4650. Users should be aware of the fact that all revisions to the CC Exclusions List (FYs 1989, 1990, 1991, 1992, 1993, 1994, 1995, 1996, 1997, and 1998) and those in Tables 6F and 6G of this document must be incorporated into the list purchased from NTIS in order to obtain the CC Exclusions List applicable for discharges occurring on or after October 1, 1998. Alternatively, the complete documentation of the GROUPER logic, including the current CC Exclusions List, is available from 3M/Health Information Systems (HIS), which, under contract with HCFA, is responsible for updating and maintaining the GROUPER program. Version 16.0 of this manual, which will include the final FY 1999 DRG changes, will be available in October 1998 for $225.00, which includes $15.00 for shipping and handling. This manual may be obtained by writing 3M/ HIS at the following address: 100 Barnes Road; Wallingford, Connecticut 06492; or by calling (203) 949-0303. 7. Review of Procedure Codes in DRGs 468, 476, and 477 Each year, we review cases assigned to DRG 468 (Extensive OR Procedure Unrelated to Principal Diagnosis), DRG 476 (Prostatic OR Procedure Unrelated to Principal Diagnosis), and DRG 477 (Nonextensive OR Procedure Unrelated to Principal Diagnosis) in order to determine whether it would be appropriate to change the procedures assigned among these DRGs. DRGs 468, 476, and 477 are reserved for those cases in which none of the OR procedures performed is related to the principal diagnosis. These DRGs are intended to capture atypical cases, that is, those cases not occurring with sufficient frequency to represent a [[Page 40961]] distinct, recognizable clinical group. DRG 476 is assigned to those discharges in which one or more of the following prostatic procedures are performed and are unrelated to the principal diagnosis. 60.0 Incision of prostate 60.12 Open biopsy of prostate 60.15 Biopsy of periprostatic tissue 60.18 Other diagnostic procedures on prostate and periprostatic tissue 60.21 Transurethral prostatectomy 60.29 Other transurethral prostatectomy 60.61 Local excision of lesion of prostate 60.69 Prostatectomy NEC 60.81 Incision of periprostatic tissue 60.82 Excision of periprostatic tissue 60.93 Repair of prostate 60.94 Control of (postoperative) hemorrhage of prostate 60.95 Transurethral balloon dilation of the prostatic urethra 60.99 Other operations on prostate All remaining OR procedures are assigned to DRGs 468 and 477, with DRG 477 assigned to those discharges in which the only procedures performed are nonextensive procedures that are unrelated to the principal diagnosis. The original list of the ICD-9-CM procedure codes for the procedures we consider nonextensive procedures, if performed with an unrelated principal diagnosis, was published in Table 6C in section IV. of the Addendum to the September 30, 1988 final rule (53 FR 38591). As part of the final rules published on September 4, 1990, August 30, 1991, September 1, 1992, September 1, 1993, September 1, 1994, September 1, 1995, August 30, 1996, and August 29, 1997, we moved several other procedures from DRG 468 to 477, as well as moving some procedures from DRG 477 to 468. (See 55 FR 36135, 56 FR 43212, 57 FR 23625, 58 FR 46279, 59 FR 45336, 60 FR 45783, 61 FR 46173, and 62 FR 45981, respectively.) a. Adding procedure codes to MDCs. We annually conduct a review of procedures producing DRG 468 or 477 assignments on the basis of volume of cases in these DRGs with each procedure. Our medical consultants then identify those procedures occurring in conjunction with certain principal diagnoses with sufficient frequency to justify adding them to one of the surgical DRGs for the MDC in which the diagnosis falls. Based on this year's review, we did not identify any necessary changes; therefore, we did not propose to move any procedures from DRGs 468 and 477 to one of the surgical DRGs. b. Reassignment of procedures among DRGs 468, 476, and 477. We also reviewed the list of procedures that produce assignments to DRGs 468, 476, and 477 to ascertain if any of those procedures should be moved from one of these DRGs to another based on average charges and length of stay. Generally, we move only those procedures for which we have an adequate number of discharges to analyze the data. Based on our review this year, we did not propose to move any procedures from DRG 468 to DRGs 476 or 477, from DRG 476 to DRGs 468 or 477, or from DRG 477 to DRGS 468 or 476. 8. Changes to the ICD-9-CM Coding System As discussed above in section II.B.1 of this preamble, the ICD-9-CM is a coding system that is used for the reporting of diagnoses and procedures performed on a patient. In September 1985, the ICD-9-CM Coordination and Maintenance Committee was formed. This is a Federal interdepartmental committee charged with the mission of maintaining and updating the ICD-9-CM. That mission includes 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 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 Committee is co-chaired by the National Center for Health Statistics (NCHS) and HCFA. The NCHS has lead responsibility for the ICD-9-CM diagnosis codes included in the Tabular List and Alphabetic Index for Diseases while HCFA has lead responsibility for the ICD-9-CM procedure codes included in the Tabular List and Alphabetic Index for Procedures. The Committee encourages participation in the above process by health-related organizations. In this regard, the Committee holds public meetings for discussion of educational issues and proposed coding changes. These meetings provide an opportunity for representatives of recognized organizations in the coding fields, such as the American Health Information Management Association (AHIMA) (formerly American Medical Record Association (AMRA)), the American Hospital Association (AHA), and various physician specialty groups as well as physicians, medical record administrators, health information management professionals, and other members of the public to contribute ideas on coding matters. After considering the opinions expressed at the public meetings and in writing, the Committee formulates recommendations, which then must be approved by the agencies. The Committee presented proposals for coding changes at public meetings held on June 5 and December 4 and 5, 1997, and finalized the coding changes after consideration of comments received at the meetings and in writing within 30 days following the December 1997 meeting. The initial meeting for consideration of coding issues for implementation in FY 2000 was held on June 4, 1998. Copies of the minutes of the 1997 meetings can be obtained from the HCFA Home Page @ http://www.hcfa.gov/ pubaffr.htm, under the ``What's New'' listing. Paper copies of these minutes are no longer available and the mailing list has been discontinued. We encourage commenters to address suggestions on coding issues involving diagnosis codes to: Donna Pickett, Co-Chairperson; ICD-9-CM Coordination and Maintenance Committee; NCHS; Room 1100; 6525 Belcrest Road; Hyattsville, Maryland 20782. Comments may be sent by E- mail to: dfp4@cdc.gov. Questions and comments concerning the procedure codes should be addressed to: Patricia E. Brooks, Co-Chairperson; ICD-9-CM Coordination and Maintenance Committee; HCFA, Center for Health Plans and Providers, Plan and Provider Purchasing Policy Group, Division of Acute Care; C4- 05-27; 7500 Security Boulevard; Baltimore, Maryland 21244-1850. Comments may be sent by E-mail to: pbrooks@hcfa.gov. The ICD-9-CM code changes that have been approved will become effective October 1, 1998. The new ICD-9-CM codes are listed, along with their proposed DRG classifications, in Tables 6A and 6B (New Diagnosis Codes and New Procedure Codes, respectively) in section VI. of the Addendum to this proposed rule. As we stated above, the code numbers and their titles were presented for public comment in the ICD- 9-CM Coordination and Maintenance Committee meetings. Both oral and written comments were considered before the codes were approved. Therefore, we solicited comments only on the proposed DRG classifications. Further, the Committee has approved the expansion of certain ICD-9- CM codes to require an additional digit for valid code assignment. Diagnosis codes that have been replaced by expanded codes, other codes, or have been deleted are in Table 6C (Invalid Diagnosis Codes). These invalid diagnosis codes will not be recognized by the GROUPER beginning with discharges occurring on or after October 1, 1998. The [[Page 40962]] corresponding new or expanded diagnosis codes are included in Table 6A. Procedure codes that have been replaced by expanded codes, other codes, or have been deleted are in Table 6D (Invalid Procedure Codes). Revisions to diagnosis code titles are in Table 6E (Revised Diagnosis Code Titles), which also include the proposed DRG assignments for these revised codes. For FY 1999, there are no revisions to procedure code titles. We received several comments about our proposed DRG assignments of new and revised codes. Comment: One commenter believes that revised diagnosis code 518.81 (acute respiratory failure) should be assigned as a ``major complication'' in DRG 121 since it was classified in this manner prior to the code revision. In addition, new diagnosis codes 518.83 (chronic respiratory failure) and 518.84 (acute and chronic respiratory failure) each should also be classified as a ``major complication'' in DRG 121. Several commenters stated that new procedure code 37.67 (implantation of cardiomyostimulation system) should not be classified to DRGs 442, 443, and 486 since the procedure is not performed for either injuries or trauma. Commenters also noted that the DRG assignments as set forth in Tables 6A through 6E in the May 8, 1998 proposed rule (63 FR 22576) were not always aligned properly with the appropriate MDC number. Response: We agree with the commenter that diagnosis codes 518.81, 518.83, and 518.84 should be included on the ``major complication'' list for DRG 121. As noted in the comment, code 518.81 is currently designated as a major complication and the assignment remains valid. In addition, the expanded codes 518.83 and 518.84 should be assigned to the major complication list because these conditions were formerly assigned to code 518.81. We also agree that procedure code 37.67 should not have been assigned to DRGs 442, 443, and 486 for the reasons cited by the commenter. We have revised Tables 6A, 6C, and 6E to reflect these changes. In addition, we have reformatted the tables to correct any alignment problems. Finally, we note that in Table 6B, the DRG assignment of procedure code 86.67 should list only DRGs 504, 506, and 507 under MDC 22. DRGs 458 and 472, which were listed in the proposed rule, have been deleted as a result of our restructuring of the burn DRGs (see section II.B.3 of this preamble). 9. Other Issues a. Palliative care. Effective October 1, 1996 (FY 1997), we introduced a diagnosis code to allow the identification of those cases in which palliative care was delivered to a hospital inpatient. This code, V66.7 (Encounter for palliative care), was unusual in that there had been no previous code assignment that included the concept of palliative care. Since this was a new concept, instructional materials were developed and distributed by the AHA as well as specialty groups on the use of this new code. With new codes, it sometimes takes several years for physician documentation to improve and for coders to become accustomed to looking for this type of information in order to assign a code. There is an inclusion note listed under V66.7 which indicates that this code should be used as a secondary diagnosis only; the patient's medical problem would always be listed first. Currently, use of diagnosis code V66.7 does not have an impact on DRG assignment. Consistent with prior practice, we have waited until the FY 1997 data became available for analysis before considering any possible modifications to the DRGs. As discussed in the proposed rule, in analyzing the FY 1997 bills received through September 1997, we found that 4,769 discharges included V66.7 as a secondary diagnosis. These cases were widely distributed throughout 199 DRGs. The vast majority of these DRGs included five or fewer discharges with use of palliative care. Only 12 DRGs included more than 100 cases. These were the following: ------------------------------------------------------------------------ Number of DRG Title cases ------------------------------------------------------------------------ 10............................... Nervous System Neoplasms 144 with CC. 14............................... Specific Cerebrovascular 272 Disorders Except TIA. 79............................... Respiratory Infections 139 and Inflammations Age >17 with CC. 82............................... Respiratory Neoplasms... 526 89............................... Simple Pneumonia and 200 Pleurisy Age >17 with CC. 127.............................. Heart Failure and Shock. 184 172.............................. Digestive Malignancy 226 with CC. 203.............................. Malignancy of 285 Hepatobiliary System or Pancreas. 239.............................. Pathological Fractures 218 and Musculosketal and Connective Tissue Malignancy. 296.............................. Nutritional and 173 Miscellaneous Metabolic Disorders Age >17 with CC. 403.............................. Lymphoma and Non-Acute 178 Leukemia with CC. 416.............................. Septicemia Age >17...... 147 ------------------------------------------------------------------------ Six of these DRGs are cancer-related; however, the other DRGs are quite diverse. Upon further analysis, we found that, for the most part, discharges with code V66.7 do not significantly differ in length of stay from the discharges in the same DRG without code V66.7. The length of stay for discharges with code V66.7 are sometimes longer and sometimes shorter and the comparative length of stay for a given DRG tends to vary by only one day. In general, the average charges for a palliative care case discharge with a secondary code of V66.7 were lower than the charges for other discharges within the DRG. However, these differences were relatively small and were well within the standard variation of charges for cases in the DRG. One approach we could take to revise the DRGs would be to divide those DRGs with a large number of cases coded with V66.7 into two different DRGs, with and without palliative care. However, the relatively small proportion of cases in each DRG argues against this approach; no DRG has more than 1 percent of its cases coded with palliative care and, in most cases, the percentage is well under 1 percent. An alternative approach would be to group all palliative care cases, regardless of the underlying disease or condition, into one new DRG. However, the charges of these cases are so varied that this is not a logical choice. In addition, there is a lack of clinical coherence in such an approach. The underlying diagnoses of these cases range from respiratory conditions to heart failure to septicemia. [[Page 40963]] Because there are so few cases in the FY 1997 data and they are so widely dispersed among different DRGs, we did not propose any DRG modification. We will make a more detailed analysis of these cases over the next year based on a more complete FY 1997 data file as well as review of the FY 1998 cases that will be available later this year. As time goes by, hospital coders and physicians should become more aware of this code and we hope that more complete data will assist our decision-making process. We received a few comments supporting our decision to make no DRG changes at this time for palliative care cases. One commenter agreed with our statement that it may take several years for use of this code to spread through the medical community. b. PTCA. Effective with discharges occurring on or after October 1, 1997, we reassigned cases of PTCA with coronary artery stent implant from DRG 112 (Percutaneous Cardiovascular Procedures) to DRG 116 (Other Permanent Cardiac Pacemaker Implant or PTCA with Coronary Artery Stent Implant). In the August 29, 1997 final rule with comment period, we responded to several commenters who contended that PTCA cases treated with platelet inhibitors were as resource intensive as the PTCA with stent implant cases and that these cases should also be moved to DRG 116. However, there is currently no code that describes the infusion of platelet inhibitors. Therefore, we were unable to make any changes in the DRGs for FY 1998. As set forth in Table 6B, New Procedure Codes in section VI. of the addendum to this final rule, a new procedure code for injection or infusion of platelet inhibitors (code 99.20) will be effective with discharges occurring on or after October 1, 1998. Our usual policy on new codes is to assign them to the same DRG or DRGs as their predecessor code. Because infusion of platelet inhibitors is currently assigned to a non-OR procedure code, we followed our usual practice and designated code 99.20 as a non-OR code that does not affect DRG assignment. We will not have any data on this new code until we receive bills for FY 1999. Thus, we would be unable to make any changes in DRG assignment until FY 2001. We note, however, that the Conference Report that accompanied the Balanced Budget Act of 1997 contained language stating that ``* * * in order to ensure that Medicare beneficiaries have access to innovative new drug therapies, the Conferees believe that HCFA should consider, to the extent feasible, reliable, validated data other than MedPAR data in annually recalibrating and reclassifying the DRGs.'' (H.R. Rep. No. 105-217 at 734 (1997)). At the time the proposed rule was published, we had received no data that would have allowed us to make an appropriate modification of DRG 112 for PTCA cases with platelet infusion therapy. In that rule, we stated that we would review and analyze any data we received during the comment period about the use of platelet inhibitors for Medicare beneficiaries. Since publication of the proposed rule, we received some data concerning the use of GPIIb/IIIa platelet inhibitor drug therapy as well as some comments on the issue. A discussion of the data and the comments and our responses are set forth below. Comment: The data we received were provided by the pharmaceutical company that manufactures a GPIIb/IIIa platelet inhibitor. In its comment accompanying the data, the company states its belief that the data conclusively demonstrate that procedure code 99.20 should be assigned to DRG 116 effective for discharges occurring on or after October 1, 1998. We received two other comments from hospitals supporting this reassignment in order to improve payment for a beneficial drug therapy. Another hospital urged HCFA not to make the reassignment because the commenter believes that there is no evidence that use of the drug decreases mortality or the risk of need for emergency coronary bypass in patients undergoing stent implantation. In addition, this commenter believes that the price charged for platelet inhibitor is exorbitant and that HCFA should not directly subsidize a pharmaceutical company through a DRG change. Finally, two commenters, a drug company and a pharmaceutical association, were encouraged by HCFA's willingness to consider data other than MedPAR data for analyzing possible DRG changes. The data we received comprise two different sets of Medicare beneficiaries who received PTCA, PTCA with implant of a coronary stent, PTCA with platelet inhibitor therapy, or PTCA with both implant of a stent and platelet inhibitor therapy. One set of data consists of just under 500 patients who received treatment in seven hospitals during a clinical trial conducted between January 1, 1996 and June 15, 1997. The other set consists of just over 6,200 patients treated in 83 hospitals between October 1, 1995 and December 31, 1996 (this is data from a health care information company that, among other products and services, performs clinical and financial analysis of data under contract with hospitals). For the first set of data, the hospitals are identified; however, for the second set of data, the hospital identifying information is confidential and was not released to HCFA. In order to provide HCFA with standardized charges, the information company obtained the HCFA provider-specific file and standardized the charges before providing them to HCFA. According to the commenter, based on the data provided the approximate average standardized charges for the different classes of patients are as follows: PTCA alone--$17,000. PTCA and stent--$22,000. PTCA and platelet inhibitor--$24,000. PTCA and both stent and platelet inhibitor--$29,000. Based on these data, the drug's manufacturer urges us to reassign procedure code 99.20 to DRG 116. The commenter also argues that failure to improve the payment for these cases may result in Medicare beneficiaries being denied equal access to potentially life-saving treatment. Response: We have reviewed the data submitted as well as considered the comments we have received. Based on the data provided, it appears that the cost of a PTCA case with platelet inhibitor drug therapy is at least as expensive as a PTCA case with stent implant. However, the vast majority of the cases (over 90 percent) cannot be linked to a hospital. In addition, although the large data set does constitute a sample of cases, as claimed by the commenter, it is not a random sample, but rather a sample of those hospitals that contract with the health information company. The pharmaceutical company states that the 83 hospitals are representative of all hospitals in the country, but we have no way to verify that claim. Because the data cannot be verified, and do not reflect a complete data set or a random sample, HCFA cannot use the data to make a change in the DRG assignment. The language that Congress included in the Conference Report that accompanied the Balanced Budget Act of 1997 stated that HCFA should ``* * * consider, to the extent feasible, reliable, validated data other than MedPAR data in annually recalibrating and reclassifying the DRGs.'' The data we have been given does not meet these requirements. We cannot validate whether the data are Medicare beneficiaries nor can we verify which hospitals provided the treatment or the amount of charges reported to Medicare. In addition, we do not believe that we [[Page 40964]] should base any DRG reclassification decisions that will increase payment for a set of cases on data that would not meet HCFA's strict requirements for making a DRG change that would lower the relative weight for a set of cases (see discussion below concerning radiosurgery procedures). As we have stated in several proposed and final rules (most recently in the August 30, 1996 final rule in a discussion of the coronary artery stent implant (61 FR 46170) and the August 29, 1997 final rule in response to a comment on the DRG assignment for new diagnosis code 686.01) (62 FR 45982), our longstanding practice is to assign a new code to the same DRG or DRGs as its predecessor code. Our compelling reason for this practice is our inability to move the cases associated with the new code to a new DRG assignment as part of the DRG reclassification and recalibration process. Consequently, our policy is to wait until we have a full year of Medicare data upon which to base an analysis of what the most appropriate DRG assignment would be. We can then move any cases that we would reassign so we can revise the DRG relative weights accordingly. If we were to assign procedure code 99.20 to DRG 116 at this time, we would be unable to move the cases associated with that code from DRG 112 into DRG 116 based on the data provided. Thus, the relative weight of DRG 112 would still reflect the cases with procedure code 99.20. Since these cases presumably have much higher charges than the other PTCA cases, the relative weight for DRG 112 would be overstated, which means the payments to those cases would be overstated. In addition, the charges for PTCA cases with platelet inhibitor drug therapy would not be reflected in the DRG 116 relative weight. Our practice of waiting until we have identifiable MedPAR data applies to all DRG changes, that is, both those changes that would enhance payment for a particular diagnosis or procedure, as well as, those that would decrease payment for a particular diagnosis or procedure. We note that, in FY 1996, when we created a new procedure code for stereotactic radiosurgery (92.3), we assigned the code to DRGs 1, 2, and 3, because that is where the predecessor procedure code was assigned. However, since code 92.3 is a nonoperating room procedure, we were relatively sure that the code would not remain assigned to DRG 1, 2, and 3 (which are the highest weighted surgical DRGs in MDC 1) once we had the actual charge data. As discussed in the August 29, 1997 final rule (62 FR 45971), procedure code 92.3 was reassigned to DRGs 7 and 8 once we had the FY 1996 data to analyze. Therefore, we ``overpaid'' those cases for 2 years; that is, their charges were much less than the average charges for DRGs 1, 2, and 3. We believe that any data we use to reclassify and recalibrate DRGs must be comprehensive and valid, as well as verifiable by HCFA. Concerning the commenter's argument that failure to change the DRG assignment for infusion of platelet inhibitor will compromise the availability of this treatment for Medicare beneficiaries, we note, as we have in several previous documents, that it is a violation of a hospital's Medicare provider agreement to place restrictions on the number of Medicare beneficiaries it accepts for treatment unless it places the same restrictions on all other patients. c. Implantation of Muscle Stimulator Comment: We received one comment arguing that the current DRG assignment for the implantation of a muscle stimulator and the associated tendon transfer for quadriplegics is inappropriate. The specific muscle stimulator device (an implanted neuroprosthesis that restores functional hand motion in people with quadriplegia who are 24 months post-injury) was approved by the Food and Drug Administration in August 1996. The device is designed to provide neuromuscular stimulation for certain patients with quadriplegia so that they can grasp with their hand and perform tasks such as holding eating utensils and pens and brushing their teeth. In many cases, the patient also undergoes a tendon transfer to the hand during the same admission or during a prior admission. The commenter notes that when the tendon transfer (procedure code 82.56 (other hand tendon transfer or transplantation)) and the insertion of the muscle stimulator (procedure code 83.92 (insertion or replacement of skeletal muscle stimulator)) are performed during the same admission, the case is assigned to DRG 7 or 8 (Peripheral and Cranial Nerve and Other Nerve System Procedures). However, when the procedures are performed during two separate admissions, the tendon transfer is assigned to DRGs 7 and 8 and the insertion of the muscle stimulator is assigned to DRG 468 (Extensive OR Procedure Unrelated to Principal Diagnosis). The commenter stated that although payment for DRGs 7, 8, and 468 are all significantly less than the cost of the hospital stay and the device, DRG 468 pays more and results in the hospital losing less money. The commenter noted that the device alone costs $24,500 and hospitals report losses of $11,000 to $26,000 when the device is inserted and a tendon transfer is performed during the same admission (resulting in assignment to DRGs 7 and 8). However, when the insertion of the device is performed in a separate admission, the cases are assigned to DRG 468 and hospitals' losses are limited to $4,000 to $18,000. The commenter believes that hospitals will refuse to perform this very useful surgery unless the DRG assignment is revised. If the insertion of the muscle stimulator were assigned to a surgical DRG in MDC 1 where the diagnosis codes for quadriplegia are assigned, the highest paying DRG assignment would be DRGs 1, 2, and 3 (Craniotomy). Besides being clinically inappropriate, the commenter believes the weights for these DRGs are too low to adequately pay for this procedure. The commenter recommended both a short and a long-term solution for this problem. For now, all cases with insertion of muscle stimulators performed in conjunction with tendon transfer should be assigned to DRG 468. In the long term, HCFA should establish a new DRG for the implantation of muscle stimulation devices and other stimulation devices as they become available. Response: In examining the latest FY 1997 MedPAR file (bills received through March 1998), we found only three cases for implantation of muscle stimulators for quadriplegics. One case was assigned to DRG 7 and the other two to DRG 8. The standardized charge and length of stay for each case is set forth below. ------------------------------------------------------------------------ Length of DRG Standardized stay charge (days) ------------------------------------------------------------------------ 7.............................................. $25,227 7 8.............................................. 8,849 2 8.............................................. 42,183 2 ------------------------------------------------------------------------ The average charge for all cases assigned to DRG 7 is approximately $21,000 and the average charge for DRG 8 cases is about $11,500. With so few cases, we would prefer to review the data in the FY 1998 MedPAR file before making any reclassification. Therefore, we will add these cases to our FY 2000 DRG reclassification analysis agenda. We note that the charges reported for two of the three cases are significantly less than the costs that the commenter believes would be incurred for this surgery (approximately $35,000). It would be inappropriate to assign the muscle stimulator insertions solely [[Page 40965]] to DRG 468. This DRG was created to capture a set of clinically unrelated cases where the only operating room procedures performed are unrelated to the patient's principal diagnosis. To permanently assign a procedure code only to DRG 468 would be contrary to the basic design and precepts of the DRG system. C. Recalibration of DRG Weights We proposed to use the same basic methodology for the FY 1999 recalibration as we did for FY 1998. (See the August 29, 1997 final rule with comment (62 FR 45982).) That is, we recalibrated the weights based on charge data for Medicare discharges. However, we used the most current charge information available, the FY 1997 MedPAR file, rather than the FY 1996 MedPAR file. The MedPAR file is based on fully-coded diagnostic and surgical procedure data for all Medicare inpatient hospital bills. The final recalibrated DRG relative weights are constructed from FY 1997 MedPAR data, based on bills received by HCFA through March 1998, from all hospitals subject to the prospective payment system and short- term acute care hospitals in waiver States. The FY 1997 MedPAR file includes data for approximately 11.3 million Medicare discharges. The methodology used to calculate the DRG relative weights from the FY 1997 MedPAR file is as follows: All the claims were regrouped using the DRG classification revisions discussed above in section II.B of this preamble. Charges were standardized to remove the effects of differences in area wage levels, indirect medical education costs, disproportionate share payments, and, for hospitals in Alaska and Hawaii, the applicable cost-of-living adjustment. The average standardized charge per DRG was calculated by summing the standardized charges for all cases in the DRG and dividing that amount by the number of cases classified in the DRG. We then eliminated statistical outliers, using the same criteria as was used in computing the current weights. That is, all cases that are outside of 3.0 standard deviations from the mean of the log distribution of both the charges per case and the charges per day for each DRG. The average charge for each DRG was then recomputed (excluding the statistical outliers) and divided by the national average standardized charge per case to determine the relative weight. A transfer case (including a postacute care transfer case as discussed in section IV.A of this preamble) is counted as a fraction of a case based on the ratio of its length of stay (plus one day to account for the double per diem payment for the first day) to the geometric mean length of stay of the cases assigned to the DRG. That is, a 5-day length of stay transfer case assigned to a DRG with a geometric mean length of stay of 10 days is counted as 0.6 of a total case. Transfers from DRGs 209, 210, or 211 to postacute care are counted as a fraction of a discharge based on the ratio determined by dividing the geometric mean length of stay for the DRG by the sum of half the geometric mean and half the length of stay for the case, plus one. We established the relative weight for heart and heart- lung, liver, and lung transplants (DRGs 103, 480, and 495) in a manner consistent with the methodology for all other DRGs except that the transplant cases that were used to establish the weights were limited to those Medicare-approved heart, heart-lung, liver, and lung transplant centers that have cases in the FY 1995 MedPAR file. (Medicare coverage for heart, heart-lung, liver, and lung transplants is limited to those facilities that have received approval from HCFA as transplant centers.) Acquisition costs for kidney, heart, heart-lung, liver, and lung transplants continue to be paid on a reasonable cost basis. Unlike other excluded costs, the acquisition costs are concentrated in specific DRGs (DRG 302 (Kidney Transplant); DRG 103 (Heart Transplant for heart and heart-lung transplants); DRG 480 (Liver Transplant); and DRG 495 (Lung Transplant)). Because these costs are paid separately from the prospective payment rate, it is necessary to make an adjustment to prevent the relative weights for these DRGs from including the effect of the acquisition costs. Therefore, we subtracted the acquisition charges from the total charges on each transplant bill that showed acquisition charges before computing the average charge for the DRG and before eliminating statistical outliers. When we recalibrated the DRG weights for previous years, we set a threshold of 10 cases as the minimum number of cases required to compute a reasonable weight. We proposed to use that same case threshold in recalibrating the DRG weights for FY 1999. Using the FY 1997 MedPAR data set, there are 37 DRGs that contain fewer than 10 cases. We computed the weights for the 37 low-volume DRGs by adjusting the FY 1998 weights of these DRGs by the percentage change in the average weight of the cases in the other DRGs. The weights developed according to the methodology described above, using the final DRG classification changes, result in an average case weight that is different from the average case weight before recalibration. Therefore, the new weights are normalized by an adjustment factor, so that the average case weight after recalibration is equal to the average case weight before recalibration. This adjustment is intended to ensure that recalibration by itself neither increases nor decreases total payments under the prospective payment system. Comment: One commenter was concerned about the general trend in the relative weights. This commenter calculated average relative weights for each MDC as well as the overall average DRG weight. Based upon this calculation, the commenter noted that the average weight for the pre- MDC DRGs and MDCs 8 (Diseases and Disorders of the Musculoskeletal system and Connective Tissue) and 24 (Multiple Significant Trauma) are decreasing. Concerning MDC 8, the commenter believes the average weight is decreasing because of the use of postacute care for these DRGs, noting that 4 of them are included in the list of 10 DRGs affected by the transfer to postacute care provision (see section IV.A of this preamble for a discussion of this provision). The commenter suggested that we leave the FY 1998 weights intact for MDC 8 until we can assess the effect of postacute care transfers on average standardized amounts. For the pre-MDCs and MDC 24, the commenter believes that the cases assigned to these categories are extremely resource-intensive and that the average weights should not be decreasing. Finally, the commenter noted that, although the total weight increased for MDC 22 (Burns), the average weight decreased. The commenter believes this is inconsistent with the statement in the proposed rule that the changes being made to MDC 22 would improve the explanation of variation in resource use in those DRGs (63 FR 25579). Response: We reviewed the table of average DRG weights presented in the comment, both overall and within MDCs, and we found that the commenter has mistakenly used a simple averaging methodology to determine the mean weight rather than a weighted averaging methodology, which is how the DRG relative weights are calculated. For example, suppose an MDC has three DRGs and there are 3 cases assigned to DRG 1, 6 cases assigned to DRG 2, and 7 cases assigned to DRG 3. The weights for the DRGs are [[Page 40966]] 1.000, 2.000, and 3.000, respectively. The simple average weight for the three DRGs would be calculated by adding the weights and dividing by the number of DRGs as follows: [GRAPHIC] [TIFF OMITTED] TR31JY98.051 However, the weighted average would be calculated by first multiplying the weights of each DRG by the number of cases in that DRG and dividing by the number of cases as follows: [GRAPHIC] [TIFF OMITTED] TR31JY98.052 Because of this mistake in average weight calculation, the commenter has made some incorrect conclusions. For example, the commenter states that the average DRG weight for FY 1998 is 1.3681 and the average of the proposed FY 1999 weights is 1.3895. In reality, the average FY 1998 weight is 1.4606 and the average of the proposed FY 1999 weights is 1.4673. (Note: These average weights are based on the MedPAR cases used to recalibrate the weights; that is, the FY 1998 weights are based on FY 1996 cases reclassified into the FY 1998 DRGs and the proposed FY 1999 weights are based on FY 1997 cases reclassified into the FY 1999 DRGS). The average weight of the final FY 1999 weights is 1.4679. Contrary to the commenter's assertion, the average weight of the proposed FY 1999 MDC 22 DRGs did not decrease compared to the FY 1998 MDC 22 weights (4.6663 and 4.5234, respectively). In addition, although all of the FY 1999 proposed pre-MDC DRG weights except DRG 483 decreased relative to FY 1998, the increase in DRG 483 was large enough (coupled with an increase in cases) to result in an overall higher average weight for the pre-MDC DRGs. We note that the weights for DRGs 481, 482, and 483 have increased between the proposed and final FY 1999 recalibrations. As we have noted in the past, the weights for the transplant DRGs (481, 482, and 495) have gradually decreased over the years. In addition, the transplant DRGs have a relatively small number of cases with a large range of reported charges. A few very low or high charge cases can make a relatively dramatic difference in the weights from year to year (August 29, 1997; 62 FR 45983). Finally, with regard to the commenter's request that we set the FY 1999 MDC 8 weights equal to the FY 1998 weights, we could refer the commenter to the discussion above concerning the steps we take in recalibrating the weights. Each year, when we recalibrate the DRG weights, we use charge data from the most recent Medicare cases available. That is, we use the charges reported by hospitals to establish the weights. In this way, we ensure that we are using the most recent hospital charging practices and patterns to set the new relative weights. Because each DRG weight is ``relative'' to all other DRG weights, we cannot arbitrarily freeze a set of those DRGs at the previous year's weights. In a relative system such as this, if some weights increase, others must decrease. Finally, as discussed above, when we recalibrate the weights, a transfer case is counted as a fraction of a case rather than a whole case. Section 1886(d)(4)(C)(iii) of the Act requires that beginning with FY 1991, reclassification and recalibration changes be made in a manner that assures that the aggregate payments are neither greater than nor less than the aggregate payments that would have been made without the changes. Although normalization is intended to achieve this effect, equating the average case weight after recalibration to the average case weight before recalibration does not necessarily achieve budget neutrality with respect to aggregate payments to hospitals because payment to hospitals is affected by factors other than average case weight. Therefore, as we have done in past years and as discussed in section II.A.4.b of the Addendum to this final rule, we make a budget neutrality adjustment to assure that the requirement of section 1886(d)(4)(C)(iii) of the Act is met. III. Changes to the Hospital Wage Index A. Background Section 1886(d)(3)(E) of the Act requires that, as part of the methodology for determining prospective payments to hospitals, the Secretary must adjust the standardized amounts ``for area differences in hospital wage levels by a factor (established by the Secretary) reflecting the relative hospital wage level in the geographic area of the hospital compared to the national average hospital wage level.'' In accordance with the broad discretion conferred under the Act, we currently define hospital labor market areas based on the definitions of Metropolitan Statistical Areas (MSAs), Primary MSAs (PMSAs), and New England County Metropolitan Areas (NECMAs) issued by the Office of Management and Budget (OMB). OMB also designates Consolidated MSAs (CMSAs). A CMSA is a metropolitan area with a population of one million or more, comprised of two or more PMSAs (identified by their separate economic and social character). For purposes of the hospital wage index, we use the PMSAs rather than CMSAs since they allow a more precise breakdown of labor costs. If a metropolitan area is not designated as part of a PMSA, we use the applicable MSA. Rural areas are areas outside a designated MSA, PMSA, or NECMA. Effective April 1, 1990, the term Metropolitan Area (MA) replaced the term Metropolitan Statistical Area (MSA) (which had been used since June 30, 1983) to describe the set of metropolitan areas comprised of MSAs, PMSAs, and CMSAs. The terminology was changed by OMB in the March 30, 1990 Federal Register to distinguish between the individual metropolitan areas known as MSAs and the set of all metropolitan areas (MSAs, PMSAs, and CMSAs) (55 FR 12154). For purposes of the prospective payment system, we will continue to refer to these areas as MSAs. Section 1886(d)(3)(E) of the Act also requires that the wage index be updated annually beginning October 1, 1993. Furthermore, this section provides that the Secretary base the update on a survey of wages and wage-related costs of short-term, acute care hospitals. The survey should measure, to the extent feasible, the earnings and paid hours of employment by occupational category, and must exclude the wages and wage-related costs incurred in furnishing skilled nursing services. We also adjust the wage index, as discussed below in section III.F, to take into account the geographic reclassification of hospitals in accordance with sections 1886(d)(8)(B) and 1886(d)(10) of the Act. B. FY 1999 Wage Index Update The final FY 1999 wage index (effective for hospital discharges occurring on or after October 1, 1998 and before October 1, 1999) is based on the data collected from the Medicare cost reports submitted by hospitals for cost reporting periods beginning in FY 1995 (the FY 1998 wage index was [[Page 40967]] based on FY 1994 wage data). The FY 1999 wage index includes the following categories of data, which were also included in the FY 1998 wage index: Total salaries and hours from short-term, acute care hospitals. Home office costs and hours. Direct patient care contract labor costs and hours. The wage index also continues to exclude the direct salaries and hours for nonhospital services such as skilled nursing facility services, home health services, or other subprovider components that are not subject to the prospective payment system. Finally, as discussed in detail in the August 29, 1997 final rule with comment period, we calculate a separate Puerto Rico-specific wage index and apply it to the Puerto Rico standardized amount. (See 62 FR 45984 and 46041) This wage index is based solely on Puerto Rico's data. For FY 1999 we proposed two changes to the categories of data included in the wage index: adding contract labor costs and hours for top management positions and replacing the fringe benefit category with the wage-related costs associated with hospital and home office salaries category. These two changes reflect changes to the Medicare cost report that were discussed in the September 1, 1994 final rule with comment period (59 FR 45355). The changes were made to the cost report for cost reporting periods beginning during FY 1995. Because we are using wage data from the FY 1995 cost report for the FY 1999 wage index, these two changes will be reflected in the wage index for the first time in FY 1999. As discussed in detail in the September 1, 1994 final rule with comment period (59 FR 45355), we expanded the definition of contract services reported on the Worksheet S-3 to include the labor-related costs associated with contract personnel in a hospital's top four management positions: Chief Executive Officer/Hospital Administrator, Chief Operating Officer, Chief Financial Officer, and Nursing Administrator. We also revised the cost report to reflect a change in terminology from ``fringe benefits'' to ``wage-related costs,'' to promote the consistent reporting of these costs. (See September 1, 1994 final rule with comment period (59 FR 45356-45359).) We made this change in terminology because we believed it would eliminate confusion regarding those wage-related costs that are incorporated in the wage index versus the broader definition of fringe benefits recognized under the Medicare cost reimbursement principles. Wage-related costs, which include core and other wage-related costs, are reported on the Form HCFA-339, the Provider Cost Report Reimbursement Questionnaire. Finally, we analyzed the wage data for the following costs, which were separately reported for the first time on the FY 1995 cost reports: Physician Part A costs. Resident and Certified Registered Nurse Anesthetist (CRNA) Part A costs. Overhead cost and hours by cost center. Our analyses and proposals concerning these data are set forth below in section III.C. Comment: MedPAC submitted a general comment on the wage index. First, the Commission stated that several of the issues raised in the proposed rule stem from the failure of the wage index to account for the mix of occupational categories employed by each hospital and that if the wage index reflected this mix it would be more accurate. In addition, MedPAC, noted that new measures are needed to implement each new prospective system as well as for Medicare+Choice plans and suggested that attention should be given to alternative strategies for obtaining labor prices that could eliminate the need to collect data separately for each type of provider affected. MedPac intends to examine this issue during the upcoming year. Response: We have addressed the issue of occupational mix in the past. In the May 27, 1994 Federal Register, we indicated we were not proposing to collect occupational mix data due to a lack of support from the hospital industry for an additional reporting burden with uncertain impact (59 FR 27724). However, certain segments of the industry continue to insist that an occupational mix would make the wage index fairer. We will continue to evaluate all the data and evidence that we receive on this issue. With respect to MedPAC's interest in examining alternative data collection strategies, we look forward to the results of its examination, and will provide whatever assistance we can. C. Issues Relating to the FY 1999 Wage Index 1. Physician Part A Costs Currently, if a hospital directly employs a physician, the Part A portion of the physician's salary and wage-related costs (that is, administrative and teaching services) is included in the calculation of the wage index. However, the costs for contract physician Part A services are not included. Our policy has been that, to be included in the wage index calculation, a contracted service must be direct patient care, or, beginning with the FY 1999 wage index, top level management (see discussion above). Because some States have laws that prohibit hospitals from directly hiring physicians, the hospitals in those States have claimed that they are disadvantaged by the wage index's exclusion of contract physician Part A costs. We began collecting separate wage data for both direct and contract physician Part A services on the FY 1995 cost report in order to analyze this issue. As we discussed in the September 1, 1994 final rule with comment period (59 FR 45354), our original purpose in collecting these data was to exclude all physician Part A costs from the wage index. When we made the change to the cost report, there were five States in which hospitals were prohibited from directly employing physicians. We understand that only two States currently maintain this prohibition: Texas and California. Thus, the number of hospitals affected by our current policy has decreased. Nevertheless, the fact that hospitals in these two States are still prohibited from directly employing physicians for Part A services and, therefore, must enter into contractual agreements with physicians for these services, perpetuates the perceived inequity. The main reason we planned to exclude all Part A physician costs rather than include the contract costs was our concern that it would be difficult to accurately attribute the Part A costs and hours of these contract physicians. In addition, we were concerned that including these costs could inappropriately inflate the hospitals' average hourly wages. That is, we anticipated that average costs for contract physicians would be significantly higher than the costs for those physicians directly employed by the hospital. However, our analysis of the data shows that the average hourly wages for contract physician Part A costs are very similar to, and, in fact slightly lower than, the costs for salaried physician Part A services. Based on this result, we believe that continuing to include the directly employed physician Part A costs and adding the costs for contract physicians is the better policy. Thus, we proposed to calculate the FY 1999 wage index including both direct and contract physician Part A costs. Of the 5,070 hospitals included in the FY 1995 wage data file, approximately 32 percent reported contract physician Part A costs. Including these costs would raise the wage index values for [[Page 40968]] 2 MSAs (4 hospitals) by more than 5 percent and 7 MSAs (43 hospitals) by between 2 and 5 percent. Two MSAs and one Statewide rural area (74 hospitals) would experience a decrease between 2 and 5 percent. The wage index values for the remaining 365 areas (4,949 hospitals) would be relatively unaffected, experiencing changes of between -2 and 2 percent. We received several comments regarding the inclusion of contract physician costs, and physician Part A costs generally. The specific comments and our responses are set forth below. Comment: A national hospital association noted its concern about the inclusion of teaching-related costs in the wage index because Medicare pays separately for the salaries of teaching physicians through direct graduate medical education (GME) payments. Nevertheless, the commenter supports the inclusion of contract physician costs in the FY 1999 wage index. The commenter indicated that it would work to develop a consensus among hospital and health system representatives on which physician salaries, if any, should be included in future wage indexes. Another commenter supported the inclusion of contract physician costs but recommended that HCFA take swift action to remove teaching physician costs ``to achieve geographic equity in payments.'' Several commenters believe that all physician Part A costs, including teaching physician costs, should be recognized in calculating the wage index. The commenters asserted that these are costs of doing business, and including them in the wage index appropriately measures the geographic variations in what hospitals pay for labor. However, numerous commenters argued that it is inappropriate to include teaching physician costs in the wage index because, in effect, it results in double payment to teaching hospitals for these costs. Recognizing that HCFA does not have the data available to separately identify the portion of physician costs attributable to teaching physicians, these commenters believe it would be preferable to remove all Part A physician costs from the wage index calculation. Response: As a conceptual matter, we believe that physician Part A costs other than teaching physician costs should be included in the wage index because these costs are paid under the prospective payment system. Further, in light of the data now available, we believe including contract physician Part A costs improves equity in the wage index by allowing hospitals that are prohibited by State law from directly employing physicians to include their costs of contracted physicians. With regard to teaching physician costs, the 1995 cost report does not separate teaching physician costs from other physician Part A costs. Consequently, we are unable to exclude teaching physician costs from the FY 1999 wage index. We believe the optimal approach is to consider this issue directly in developing the FY 2000 wage index. To facilitate evaluation of this issue, we will instruct the fiscal intermediaries to separate teaching physician costs from hospitals' FY 1996 wage data. We will carefully analyze those data, and any changes we propose to make based on that analysis will be included in the FY 2000 proposed rule. We do not agree with the commenters' suggestion that, in lieu of collecting data that would allow us to separately identify teaching physician costs, we should remove all physician salaries from the wage index. These physician Part A costs are incurred by the hospital for services related to such positions as medical director and clinical department heads. As such, they are legitimate labor costs included under the prospective payment system. Based on our analysis of the FY 1995 cost reports, we believe that the data reported for physician Part A costs are sufficiently reliable and complete that inclusion of physician Part A costs (direct as well as contract costs) for FY 1999 results in a wage index that better reflects relative hospital labor costs than a wage index that excludes all physician Part A costs. Moreover, as stated above, we believe the addition of contract physician Part A costs in the FY 1999 wage index improves the fairness and accuracy of the wage index relative to the FY 1998 wage index (which included direct physician Part A costs (salaries) but not contract physician Part A costs). Thus, rather than excluding all physician Part A costs, we believe the more responsible approach is to collect the necessary data as expeditiously as possible in order to analyze whether it is feasible to separate teaching physician costs from other physician Part A costs. Comment: Several commenters favored not only including physician salaries in the wage index, but also continuing to include teaching physician salaries. Commenters believe that if Congress had known about the payment redistributions that would result from eliminating teaching physician salaries from the wage index before it had enacted the reductions applicable to teaching hospitals in the Balanced Budget Act of 1997, it may not have enacted such deep cuts. One commenter also suggested that if we excluded physician salaries, we would need to restandardize the large urban standardized amount to reflect the new wage index. Another commenter stated that the costs of teaching physicians and residents should be included in the wage index because Medicare payments for GME are not sufficient to compensate hospitals for their GME costs. This commenter compared hospitals' direct GME costs on the Medicare cost report with the payments they receive and estimated a shortfall of $900 million. The commenter further noted that reductions in Medicare disproportionate share payments as a result of the Balanced Budget Act would have the effect of increasing the empirical estimate for the indirect graduate medical education adjustment, leading to a further shortfall in payments for GME. Response: We cannot know what Congress would or would not have done if it had known about the impacts of future changes to wage index policy. Rather, refinements to the wage data should be evaluated on their individual merits in terms of whether they contribute to or detract from the fairness and accuracy of the wage index. We disagree that changes to the wage index may require restandardization of the large urban standardized amount. The large urban standardized amount was not created by a separate standardization of the costs of hospitals in large urban areas, but by applying differential update factors established by Congress. We also disagree with the comment that the wage index should continue to include costs related to teaching physicians and residents because current and future GME payments are not fully compensating hospitals for their GME costs. The adequacy of direct GME payments is a separate issue by virtue of the fact that these costs are recognized separately and paid for through Medicare outside the prospective payment system. The amount Medicare pays for direct GME is based on policy considerations related to the nature of GME, and reflects Medicare's fair share of those costs. Similarly, indirect GME costs are distinct from hospitals' labor costs, and the level of IME payments is not relevant to the wage index. Comment: Many commenters referred to an analysis done by one commenter showing the projected payment impacts by State of our proposed policy of including physician (both direct and contract), resident, and CRNA costs in the wage index. These commenters [[Page 40969]] referred to the large losses that, according to this analysis, certain States will allegedly suffer because of this policy (California: $79 million; Florida: $36 million; Texas: $10 million). Corresponding gains were cited among northeast hospitals. The suggestion of these comments was that we should revise our proposed policy and exclude all of these costs to redistribute these losses and gains. Response: We disagree with the characterization of this analysis. With the exception of contract physician costs, all of these costs have been included in prior wage indexes. Therefore, the commenter's analysis does not reflect the impact of the proposed wage index relative to the current wage index. With respect to the losses in certain States cited by the commenter, our analysis indicates that, the projected payment impacts of including contract physician costs relative to a wage index without these costs are, respectively: a $13 million decrease, a $15 million decrease, and an $18 million increase. We note that these figures do not reflect the impact of changes to the wage indexes in these areas resulting from updating from the 1994 wage data to 1995 wage data, or the exclusion of allocated overhead. They do, however, present a clearer picture of the impacts in these States of including contract physician costs relative to current policy. Comment: One commenter vigorously opposes the inclusion of contract physician Part A costs, arguing we should instead exclude all physician Part A costs. The commenter, a national association of health systems, argued that this proposal contradicts the objectives we identified in the May 27, 1994 proposed rule (59 FR 27720) and the September 1, 1994 final rule (59 FR 45354), where we discussed the need to separately collect physician Part A costs. The commenter raises the following points and ultimately recommends excluding all physician Part A costs from the calculation of the wage index. First, the commenter contends that, by choosing to include physician Part A contract costs rather than exclude all physician Part A costs, we ``have expanded the unfair and unjustifiable policy tilt enjoyed by teaching hospitals.'' To emphasize this point, the commenter notes that over 70 percent of all contract physician costs stem from teaching hospitals (90 percent of salaried physician costs are also from teaching hospitals). Second, the commenter states that our rationale for proposing to include contract physician costs focused narrowly on whether these costs would inappropriately inflate the wage data. This narrow focus, according to the commenter, left out any explanation of why it is better to include contract physician costs rather than to exclude all Part A physician costs. Third, the commenter quotes liberally from our discussion in the proposed and final rules published in 1994, particularly our rationale for providing for separate reporting of physician Part A costs on the cost report. Referenced specifically are the three reasons why HCFA believed at that time that eliminating physician Part A costs would be preferable to including contracted physician costs. These reasons were: (1) Physician costs are not driven by normal labor market situations; (2) many hospitals indicated difficulties in accurately determining hours for these physicians' services; and (3) some hospitals have difficulty separating costs related to Part A from those related to Part B. The commenter specifically asks HCFA why it has changed its beliefs. Finally, the commenter surmises that one reason we proposed to include contract physician costs is that few areas would experience a significant change in their wage index values. To refute this, the commenter describes the results of analysis of the impacts of the proposed policy. The analysis found ``a dramatic and damaging impact on California, the largest state in the nation in terms of hospitals and number of Medicare discharges.'' The commenter believes that ``HCFA's wage index policy should be based not on whether the outcome will result in little change, but on whether it is the right policy in the first place.'' Response: We appreciate the considered arguments and detailed analysis presented by the commenter and understand the importance of this issue to the hospitals represented by the association. We agree with the commenter that the primary consideration in developing and refining the hospital wage index should be the ``right policy.'' In the context of the hospital wage index, we believe we should promote the fair and accurate measurement of relative hospital wage levels across geographic areas. At the same time, we believe it is appropriate to consider the potential impact of possible courses of action, though we agree with the commenter that the potential impact should not be the driving force in policy decisions. In the context of the hospital wage index, it is also critical to keep in mind that developing the ``right policy'' is a function not only of conceptual issues but also of data issues. If, for example, we believe as a conceptual matter that a certain type of cost should be included in the wage index, but the data on those costs are incomplete and unreliable, then including the costs in the wage index (which are conceptually right) could (because of the data problems) distort the measure of relative wage levels across geographic areas, and thus detract from the fairness and accuracy of the wage index; similarly, if we believe as a conceptual matter that a certain type of cost should be excluded from the wage index, but there is incomplete and unreliable data to separate those costs from other costs, then excluding the costs based on bad data could detract from the equity of the wage index. Thus, our ability to implement a ``conceptually right'' policy depends on the availability of reliable and complete data. As indicated above in the response to another comment, we believe there is good reason to include all physician Part A costs, rather than exclude all physician Part A costs as the commenter recommends. Among other things, with the exception of teaching physician costs, physician Part A costs are Part A costs that are paid under the prospective payment system. In addition, physician Part A costs represent above- average costs, although only a small percentage of the total for most hospitals; therefore, excluding all physician Part A costs might understate the relative wages of some hospitals. Based on our analysis of the FY 1995 cost reports, we believe that data reported for physician Part A costs are sufficiently reliable and complete that inclusion of the costs results in a wage index that is more fair and accurate, relative to a wage index which would exclude all physician Part A costs, even if the data are not perfect. As discussed above, although we have decided to adopt our proposal to include contract physician Part A costs in the wage index, we intend to direct the fiscal intermediaries to separately identify physician Part A costs (salaried and contracted) related to teaching for cost reports beginning during FY 1996. Although this information will not be reported separately on the Worksheet S, Part III until FY 1997 cost reports, we believe this issue merits undertaking a special auditing effort of the FY 1996 cost reports. With regard to the high proportion of physician costs attributable to teaching hospitals, although the distribution of costs seems disproportionate (and this is a large part of the reason we are expediting our efforts to separate teaching physician costs from other physician costs), our analysis of these [[Page 40970]] data indicates that, among hospitals reporting these costs, there is little difference between teaching and nonteaching hospitals in terms of the relative impact of these costs on hospitals' average hourly wages. That is, among both teaching and nonteaching hospitals reporting physician Part A costs, these costs make up between 3 and 4 percent of their total wage costs. Therefore, although more teaching hospitals report these costs than nonteaching hospitals (47 percent of teaching hospitals versus 30 percent of nonteaching hospitals), the average hourly wages of teaching hospitals are not more heavily weighted by these costs than they are for nonteaching hospitals. In fact, two of the MSAs that would be most negatively affected by excluding all physician costs from the wage data, Pittsburgh, PA and Rochester, NY, both have more nonteaching hospitals reporting physician costs than teaching hospitals. We believe the commenter's perception that we are tilting the wage index policy toward teaching hospitals is misguided and reflects an oversimplification of the issue. Based on our analysis of this issue, we are convinced the most prudent course is to focus on specifically developing data to further improve the fairness and accuracy of the wage index. In describing the perceived problems from our discussion of the physician cost data in the May 27, 1994 proposed rule, the commenter fails to acknowledge that the discussion was in relation to a proposed change. In fact, it was in response to public comments on this proposed change where we agreed to revise the cost report to collect data on contract physician costs. In addition, the September 1, 1994 final rule clearly stated that HCFA intended to evaluate the physician cost data prior to proposing any changes for the FY 1999 wage index. Regarding the problems associated with contract labor discussed in the FY 1995 proposed and final rules, we note that the separate physician cost data were not available at that time, and therefore the discussion was based on information provided from fiscal intermediaries and industry sources. Based on our analysis of the data available now, we believe that the problems are not as widespread as initially feared. Rather, these costs are similar to those reported for contracted medical providers that we do include, such as therapists and nursing staff. The commenter did not allege that there were widespread problems reporting these data. The commenter's characterization of the impact of this change on California's hospitals is inaccurate. No California MSA experiences a decrease in their wage index of more than 0.6 percent as a result of this change. The dramatic impacts referenced by the commenter occur only under the assumption that the comparative baseline excludes all physician Part A costs, the course recommended by the commenter. While excluding all physician Part A costs would result in a significant redistribution of payments to certain States such as California, other areas would experience dramatic payment decreases relative to last year. Comment: One commenter believes that, because the hospital wage index is used to adjust payments for various other types of providers, the wage data should be expanded to be as comprehensive as possible. Specifically, the commenter recommended that wage data related to excluded distinct part units, as well as all physician data, be included. Response: We have convened workgroups, both internally and externally, to focus on future wage index policies, and we anticipate that we will continue to focus on the appropriate scope of the wage data in those workgroups. In addition, any significant changes in the types of data to be included in the wage index will be implemented through the annual rulemaking process with opportunity for public comment, as has been our policy in the past. For the record, we believe that the hospital wage index should reflect, to the greatest degree possible, the wage costs associated with the prospective payment areas of the hospital. Comment: One commenter believes that there are ``evident problems with the quality and consistency of the physician contract labor data,'' which is evidenced by California's ranking as the 7th lowest State in terms of contract physician average hourly wage. This commenter also recommended that we begin a more rigorous audit mechanism of the wage data, stating that data reliability is still a problem. Response: We do not include hospitals' data (other than wage- related costs) if either the salaries or hours reported for contract labor are zero. Applying this edit to the wage data, California ranks as the 12th highest State in terms of contract physician average hourly wages. The analysis provided by the commenters did not include such an edit; therefore, their results are different. We disagree with the general point of this comment that there are quality problems with these data. These data have been subjected to the same review and edit process as are all wage data. We will continue to monitor the process for collecting wage data in the future, and make improvements as necessary. We also encourage hospitals and their associations to feel free to provide specific recommendations for potential improvements. Comment: One commenter noted that hospitals that acquire their physician Part A services through related organizations do not have an appropriate line on Worksheet S-3 to record these wage costs. Therefore, these hospitals are disadvantaged by the inclusion of costs only for directly employed and contract physician Part A services in the wage index calculation. The commenter recommended that we adjust the FY 1999 wage index to include related organization physician Part A costs for hospitals that were unable to include the costs on their Worksheet S-3s. Response: The commenter's statements about Worksheet S-3 are incorrect. The cost report instructions at section 2806.3 of the Provider Reimbursement Manual, Part II, allow hospitals to include the costs for physician Part A services from related organizations on line 33 of Worksheet S-3. These costs are also included on the trial balance, Worksheet A, in column 2 (with any adjustments in column 6). Regarding the commenter's recommendation, we cannot adjust the final FY 1999 wage index to include costs that hospitals did not properly report on their cost reports. 2. Resident and CRNA Part A Costs The wage index presently includes salaries and wage-related costs for residents in approved medical education programs and for CRNAs employed by hospitals under the rural pass-through provision. However, Medicare pays for these costs outside the prospective payment system. Removing these costs from the wage index calculation would be consistent with our general policy to exclude costs that are not paid through the prospective payment system, but, because they were not separately reported, we could not remove them. In the September 1, 1994 final rule with comment period (59 FR 45355), we stated that we would begin collecting the resident and CRNA wage data separately and would evaluate the data before proposing a change in computing the wage index. However, there were data reporting problems associated with these costs on the FY 1995 cost report. The original instructions for reporting [[Page 40971]] resident costs on Line 6 of Worksheet S-3, Part III, erroneously included teaching physician salaries and other teaching program costs from Worksheet A of the cost report. Although we issued revised instructions to correct this error, we understand these revisions may not have been uniformly instituted. Another issue relating to residents' salaries stems from apparent underreporting of these costs by hospitals and inconsistent treatment of the associated wage-related costs. In addition, the original Worksheet S-3 and reporting instructions did not provide for the separate reporting of CRNA wage- related costs. We believe that much of the CRNA Part A costs are reported under contract labor, rather than under salaried employee costs, due to the heavy use of contract labor by rural hospitals. We do not believe that it would be feasible at this time to try to remove these CRNA Part A costs from the contract labor costs in the FY 1995 cost report data. We improved the reporting instructions for CRNA costs on the FY 1996 cost report. Our analysis of the CRNA and resident wage data submitted on the FY 1995 cost report convinces us that these data are inaccurately and incompletely reported by hospitals. For example, although there are over 900 teaching hospitals receiving graduate medical education payments, only about 800 hospitals reported resident cost data. Because we do not want to make a relatively significant change in the wage index data calculation without complete and accurate data upon which to base our decision, we proposed to delay any decision regarding excluding resident and CRNA costs from the wage index until at least next year. In the May 8 proposed rule, we announced our intention to review the FY 1996 data when it becomes available later this year and present our analysis and any proposals in next year's proposed rule. Comment: Several commenters believe that HCFA should immediately exclude intern and resident and CRNA wage costs for the same reasons the commenters cited for excluding the teaching physician costs. One commenter objected to our statement that problems with the reporting of these data (stemming from inconsistent instructions) warranted a one- year delay. The commenter stated that ``it is better to exclude all clearly identified costs now rather than waiting some indeterminate time for all costs to be identified before excluding any of it.'' Analysis purporting to show a negative impact of $24 million on California due to including these data in the wage index was cited. Response: As we stated above, the instructions to the FY 1995 cost report Worksheet S-3 for reporting resident costs did not specifically separate teaching physician salaries and other GME program costs from residents' costs. This may have inappropriately inflated resident costs on Line 6 of Worksheet S-3. As a result, removing the costs reported on Line 6 from the FY 1999 wage index calculation would distort the wage index. Our reasoning with respect to retaining the CRNA costs is similar; that is, if Line 2 was removed, it would result in distortions since these costs were reported inconsistently. Therefore, because the data for these costs are not sufficiently reliable and complete, we maintain our position that the more responsible approach is to delay removing these costs until more accurate data are available for the FY 2000 wage index. With regard to the negative impact on California, any analysis based on this data will be skewed by the reporting flaws noted. The FY 1999 wage index calculation will continue to include intern and resident and CRNA wage costs. We also believe that several of the commenters are confused about the issue of CRNA costs. Currently, only the Part A portion of these costs are included in the wage index, and the only hospitals paid for these costs are small rural hospitals who employ the equivalent of no more than one full-time CRNA and are paid on the basis of reasonable costs. Therefore, they do not contribute to the concentration of physician costs in teaching hospitals. Comment: One commenter noted that the hourly wage rates for residents are lower than the overall average hourly wage of the hospitals that pay their salaries, and that the inclusion of residents' salaries and wage-related costs actually results in a decrease in teaching hospitals' average hourly wages rather than an increase, as suggested by most other commenters. The commenter suggested that removing residents from the data used to calculate the wage index would increase the wage index values in areas with a high concentration of teaching hospitals. Response: The FY 1995 data do not permit us to evaluate the accuracy of this comment because residents' salaries are commingled with teaching physicians' salaries for many hospitals. As with all changes to the wage data, the impacts cannot be evaluated properly until accurate data are available for all hospitals nationally. 3. Overhead Allocation In the proposed rule, we discussed in detail our proposal to remove from the calculation of the FY 1999 wage index the overhead costs associated with certain subprovider components that are excluded from the prospective payment system (63 FR 25586). Although the overall impact on hospitals of this change is relatively small, we believe it is an appropriate step toward improving the overall consistency of the wage index. In addition, we believe this change will significantly increase the accuracy of the wage data for individual hospitals, especially hospitals that have a relatively small portion of their facility devoted to acute inpatient care. We received several comments supporting this change, and none expressing opposition to it. One commenter referred to it as a step toward improving uniformity and overall consistency in the wage index process. We have, therefore, incorporated our proposal in the final wage index. D. Verification of Wage Data From the Medicare Cost Report The data for the FY 1999 wage index were obtained from Worksheet S- 3, Parts III and IV of the FY 1995 Medicare cost reports. The data file used to construct the final wage index includes FY 1995 data submitted to the Health Care Provider Cost Report Information System (HCRIS). As in past years, we performed an intensive review of the wage data, mostly through the use of edits designed to identify aberrant data. As a part of the August 29, 1997 final rule with comment period, we implemented a new timetable for requesting wage data corrections (62 FR 45990). We notified hospitals again of these changes through a February 1998 memorandum to the fiscal intermediaries and in the proposed rule. As noted in the proposed rule, beginning this year with the FY 1999 wage index, the wage index published in the final rule incorporates all corrections, including those to correct data entry or tabulation errors of the final wage data by the intermediary or HCFA. To allow hospitals an opportunity to evaluate the wage data to be used to construct the proposed and the final FY 1999 hospital wage index, we made available to the public data files containing the FY 1995 hospital wage data. In memoranda dated February 2 and April 21, 1998, we instructed all Medicare intermediaries to inform the prospective payment hospitals they serve of the availability of the wage data files and the process and timeframe for requesting revisions. The proposed and the final wage data files were made available February 6 and May 14, 1998, [[Page 40972]] through the Internet at HCFA's home page (http://www.hcfa.gov). The intermediaries were also instructed to advise hospitals of the alternative availability of these data through their representative hospital organizations or directly from HCFA. Table 3C in the Addendum to this final rule, as in the proposed rule, contains each hospital's adjusted average hourly wage used to construct the wage index values. A hospital can verify its adjusted average hourly wage, as calculated from Steps 4 and 5 of the computation of the wage index (see section III.E of this preamble) based on the wage data on the hospital's cost report (after taking into account any adjustments made by the intermediary), by dividing the adjusted average hourly wage in Table 3C by the applicable wage adjustment factors as set forth in Step 5 of the computation of the wage index. However, a hospital's average hourly wage using this calculation will vary from the average hourly wage shown on Line 32 of Worksheet S-3, Part III. (See Step 5 for a complete explanation.) We created the correction process, as detailed in the proposed rule, to resolve all substantive wage data correction disputes before finalizing the wage data for the FY 1999 payment rates. Hospitals had until June 5, 1998, to submit requests to correct errors in the final wage data (released May 14, 1998) due to data entry or tabulation errors by the intermediary or HCFA. The correction requests considered were limited to errors in the final wage data that the hospital could not have known about prior to the availability of the final wage data public use file. If hospitals availed themselves of these opportunities to timely identify and bring errors in their wage data to their intermediaries' attention, the wage index implemented on October 1 should be free of such errors. Nevertheless, in the unlikely event that errors should arise after that date, we retain the right to make midyear changes to the wage index under very limited circumstances. Specifically, in accordance with Sec. 412.63(w)(2), we may make midyear corrections to the wage index only in those limited circumstances where a hospital can show: (1) That the intermediary or HCFA made an error in tabulating its data; and (2) that the hospital could not have known about the error, or did not have an opportunity to correct the error, before the beginning of FY 1999 (that is, by the June 5, 1998 deadline). As indicated earlier, since a hospital will have had the opportunity to verify its data, and the intermediary will notify the hospital of any changes, we do not foresee any specific circumstances under which midyear corrections would be made. However, should a midyear correction be necessary, the wage index change for the affected area will be effective prospectively from the date the correction is made. E. Computation of the Wage Index The method used to compute the final wage index is as follows: Step 1--As noted above, we based the FY 1999 wage index on wage data reported on the FY 1995 Medicare cost reports. We gathered data from each of the non-Federal, short-term, acute care hospitals for which data were reported on the Worksheet S-3, Parts III and IV of the Medicare cost report for the hospital's cost reporting period beginning on or after October 1, 1994 and before October 1, 1995. In addition, we included data from a few hospitals that had cost reporting periods beginning in September 1994 and reported a cost reporting period exceeding 52 weeks. These data were included because no other data from these hospitals would be available for the cost reporting period described above, and particular labor market areas might be affected due to the omission of these hospitals. However, we generally describe these wage data as FY 1995 data. Step 2--For each hospital, we subtracted the excluded salaries (that is, direct salaries attributable to skilled nursing facility services, home health services, and other subprovider components not subject to the prospective payment system) from gross hospital salaries to determine net hospital salaries. To determine total salaries plus wage-related costs, we added the costs of contract labor for direct patient care, certain top management, and physician Part A services; hospital wage-related costs, and any home office salaries and wage- related costs reported by the hospital, to the net hospital salaries. The actual calculation is the sum of lines 2, 4, 6, 32, and 33 of Worksheet S-3, Part III. This calculation differs from the one computed on line 32 of Worksheet S-3, Part III. Therefore, a hospital's average hourly wage calculated under this step will be different from the average hourly wage shown on line 32, column 5. Step 3--For each hospital, we subtracted the reported excluded hours from the gross hospital hours to determine net hospital hours. To determine total hours, we increased the net hours by the addition of home office hours and hours for contract labor attributable to direct patient care, certain top management, and physician Part A salaries. Step 4--For each hospital reporting both total overhead salaries and total overhead hours greater than zero, we then allocated overhead costs. First, we determined the ratio of excluded area hours (Line 24 of Worksheet S-3, Part III) to revised total hours (Line 9 of Worksheet S-3, Part III, adding back CRNA Part A, physician Part A, and resident hours). Second, we computed the amounts of overhead salaries and hours to be allocated to excluded areas by multiplying the above ratio by the total overhead salaries and hours reported on Line 16 of Worksheet S-3, Part IV. Finally, we subtracted the computed overhead salaries and hours associated with excluded areas from the total salaries and hours derived in Steps 2 and 3. Step 5--For each hospital, we adjusted the total salaries plus wage-related costs to a common period to determine total adjusted salaries plus wage-related costs. To make the wage inflation adjustment, we estimated the percentage change in the employment cost index (ECI) for compensation for each 30-day increment from October 14, 1994 through April 15, 1996, for private industry hospital workers from the Bureau of Labor Statistics Compensation and Working Conditions. For previous wage indexes, we used the percentage change in average hourly earnings for hospital industry workers to make the wage inflation adjustment. For FY 1999 we used the ECI for compensation for private industry hospital workers because it reflects the price increase associated with total compensation (salaries plus fringes) rather than just the increase in salaries, which is what the average hourly earnings category reflected. In addition, the ECI includes managers as well as other hospital workers. We changed the methodology used to compute the monthly update factors. This new methodology uses actual quarterly ECI data to determine the monthly update factors. The methodology assures that the update factors match the actual quarterly and annual percent changes. The inflation factors used to inflate the hospital's data were based on the midpoint of the cost reporting period as indicated below. Midpoint of Cost Reporting Period ------------------------------------------------------------------------ Adjustment After Before factor ------------------------------------------------------------------------ 10/14/94...................................... 11/15/94 1.032882 11/14/94...................................... 12/15/94 1.030771 12/14/94...................................... 01/15/95 1.028721 [[Page 40973]] 01/14/95...................................... 02/15/95 1.026731 02/14/95...................................... 03/15/95 1.024776 03/14/95...................................... 04/15/95 1.022827 04/14/95...................................... 05/15/95 1.020886 05/14/95...................................... 06/15/95 1.018901 06/14/95...................................... 07/15/95 1.016822 07/14/95...................................... 08/15/95 1.014649 08/14/95...................................... 09/15/95 1.012446 09/14/95...................................... 10/15/95 1.010279 10/14/95...................................... 11/15/95 1.008146 11/14/95...................................... 12/15/95 1.006047 12/14/95...................................... 01/15/96 1.003981 01/14/96...................................... 02/15/96 1.001950 02/14/96...................................... 03/15/96 1.000000 03/14/96...................................... 04/15/96 0.998181 ------------------------------------------------------------------------ For example, the midpoint of a cost reporting period beginning January 1, 1995 and ending December 31, 1995 is June 30, 1995. An inflation adjustment factor of 1.016822 would be applied to the wages of a hospital with such a cost reporting period. In addition, for the data for any cost reporting period that began in FY 1995 and covers a period of less than 360 days or greater than 370 days, we annualized the data to reflect a 1-year cost report. Annualization is accomplished by dividing the data by the number of days in the cost report and then multiplying the results by 365. Step 6--Each hospital was assigned to its appropriate urban or rural labor market area prior to any reclassifications under sections 1886(d)(8)(B) or 1886(d)(10) of the Act. Within each urban or rural labor market area, we added the total adjusted salaries plus wage- related costs obtained in Step 5 for all hospitals in that area to determine the total adjusted salaries plus wage-related costs for the labor market area. Step 7--We divided the total adjusted salaries plus wage-related costs obtained in Step 6 by the sum of the total hours (from Step 4) for all hospitals in each labor market area to determine an average hourly wage for the area. Step 8--We added the total adjusted salaries plus wage-related costs obtained in Step 5 for all hospitals in the Nation and then divided the sum by the national sum of total hours from Step 4 to arrive at a national average hourly wage. Using the data as described above, the national average hourly wage is $20.7325. Step 9--For each urban or rural labor market area, we calculated the hospital wage index value by dividing the area average hourly wage obtained in Step 7 by the national average hourly wage computed in Step 8. We note that in June, 1998, OMB announced the designation of the Missoula, Montana MSA comprising Missoula, Montana. Step 10--Following the process set forth above, we developed a separate Puerto Rico-specific wage index for purposes of adjusting the Puerto Rico standardized amounts. We added the total adjusted salaries plus wage-related costs (as calculated in Step 5) for all hospitals in Puerto Rico and divided the sum by the total hours for Puerto Rico (as calculated in Step 4) to arrive at an overall average hourly wage of $9.5025 for Puerto Rico. For each labor market area in Puerto Rico, we calculated the hospital wage index value by dividing the area average hourly wage (as calculated in Step 7) by the overall Puerto Rico average hourly wage. Step 11--Section 4410 of Public Law 105-33 provides that, for discharges on or after October 1, 1997, the area wage index applicable to any hospital that is not located in a rural area may not be less than the area wage index applicable to hospitals located in rural areas in that State. Furthermore, this wage index floor is to be implemented in such a manner as to assure that aggregate prospective payments are not greater or less than those which would have been made in the year if this section did not apply. For FY 1999, this change affects 118 hospitals in 32 MSAs. The MSAs affected by this provision are identified in Table 4A by a footnote. F. Revisions to the Wage Index Based on Hospital Redesignation Under section 1886(d)(8)(B) of the Act, hospitals in certain rural counties adjacent to one or more MSAs are considered to be located in one of the adjacent MSAs if certain standards are met. Under section 1886(d)(10) of the Act, the Medicare Geographic Classification Review Board (MGCRB) considers applications by hospitals for geographic reclassification for purposes of payment under the prospective payment system. The methodology for determining the wage index values for redesignated hospitals is applied jointly to the hospitals located in those rural counties that were deemed urban under section 1886(d)(8)(B) of the Act and those hospitals that were reclassified as a result of the MGCRB decisions under section 1886(d)(10) of the Act. Section 1886(d)(8)(C) of the Act provides that the application of the wage index to redesignated hospitals is dependent on the hypothetical impact that the wage data from these hospitals would have on the wage index value for the area to which they have been redesignated. Therefore, as provided in section 1886(d)(8)(C) of the Act, the wage index values were determined by considering the following: If including the wage data for the redesignated hospitals would reduce the wage index value for the area to which the hospitals are redesignated by 1 percentage point or less, the area wage index value determined exclusive of the wage data for the redesignated hospitals applies to the redesignated hospitals. If including the wage data for the redesignated hospitals reduces the wage index value for the area to which the hospitals are redesignated by more than 1 percentage point, the hospitals that are redesignated are subject to that combined wage index value. If including the wage data for the redesignated hospitals increases the wage index value for the area to which the hospitals are redesignated, both the area and the redesignated hospitals receive the combined wage index value. The wage index value for a redesignated urban or rural hospital cannot be reduced below the wage index value for the rural areas of the State in which the hospital is located. Rural areas whose wage index values would be reduced by excluding the wage data for hospitals that have been redesignated to another area continue to have their wage index values calculated as if no redesignation had occurred. Rural areas whose wage index values increase as a result of excluding the wage data for the hospitals that have been redesignated to another area have their wage index values calculated exclusive of the wage data of the redesignated hospitals. The wage index value for an urban area is calculated exclusive of the wage data for hospitals that have been reclassified to another area. However, geographic reclassification may not reduce the wage index value for an urban area below the statewide rural wage index value. We note that, except for those rural areas where redesignation would reduce the rural wage index value, the wage index value for each area is computed exclusive of the wage data for hospitals that have been redesignated from the area for purposes of their wage index. As a result, several urban areas listed in Table 4a have no hospitals remaining in the area. This is because all the hospitals originally in these urban areas have been reclassified to another area by the MGCRB. These areas with no remaining hospitals receive the prereclassified wage index value. The prereclassified wage index value will apply as long as the area remains empty. The final wage index values for FY 1999 are shown in Tables 4A, 4B, 4C, [[Page 40974]] and 4F in the Addendum to this final rule. Hospitals that are redesignated should use the wage index values shown in Table 4C. Areas in Table 4C may have more than one wage index value because the wage index value for a redesignated urban or rural hospital cannot be reduced below the wage index value for the rural areas of the State in which the hospital is located. When the wage index value of the area to which a hospital is redesignated is lower than the wage index value for the rural areas of the State in which the hospital is located, the redesignated hospital receives the higher wage index value, that is, the wage index value for the rural areas of the State in which it is located, rather than the wage index value otherwise applicable to the redesignated hospitals. Tables 4D and 4E list the average hourly wage for each labor market area, prior to the redesignation of hospitals, based on the FY 1995 wage data. In addition, Table 3C in the Addendum to this final rule includes the adjusted average hourly wage for each hospital based on the FY 1995 data (as calculated from Steps 4 and 5, above). The MGCRB will use the average hourly wage published in the final rule to evaluate a hospital's application for reclassification for FY 2000, unless that average hourly wage is later revised in accordance with the wage data correction policy described in Sec. 412.63(w)(2). In such cases, the MGCRB will use the most recent revised data used for purposes of the hospital wage index. Although we did not propose any changes to the reclassification guidelines, we received two comments on that issue. Comment: One commenter was concerned that the number of hospitals participating in countywide reclassifications has declined over the years. The commenter believes that this is an indication that the criteria for hospitals in an urban county seeking reclassification to another urban county should be adjusted. Response: When we implemented the MGCRB process, we anticipated that, over the years, the number of hospitals that would continue to qualify for reclassification would decrease due to better data reporting and efforts by hospitals to constrain costs. The reclassification process is an annual process in which a hospital or group of hospitals must meet the defined criteria on an annual basis in order to remain reclassified to an alternative area for either the wage index, the standardized amount, or both. We note that hospitals that do not meet the countywide criteria under Sec. 412.234 may apply on an individual basis. Comment: One commenter supports the policy that allows rural hospitals to reclassify to another area for purposes of the disproportionate share adjustment even if the standardized amount is the same for both areas. However, this commenter is also concerned that separate criteria have not been developed for this type of reclassification and that we continue to rely on the criteria set forth in Sec. 412.230(d), which is the criteria for reclassification to another area for purposes of the standardized amount. Response: Section 4203(a) of the Balanced Budget Act of 1997 provided that, for a limited period of time, a rural hospital may apply for reclassification to another area for purposes of receiving disproportionate share payments whether or not the standardized amount is the same for both areas. Section 4203(b) provides that the MGCRB will apply the guidelines for reclassification for purposes of the standardized amount until the Secretary establishes other guidelines. We believe that the criteria in place for standardized amount reclassification are appropriate for determining whether hospitals should be reclassified for purposes of the disproportionate share payment. The criteria address the extent to which a hospital warrants reclassification by comparing the hospital's costs to its payments with and without reclassification. Nevertheless, we welcome specific suggestions for revising the DSH reclassification criteria. IV. Other Decisions and Changes to the Prospective Payment System for Inpatient Operating Costs A. Definition of Transfers (Sec. 412.4) Pursuant to section 1886(d)(5)(I) of the Act, the prospective payment system distinguishes between ``discharges,'' situations in which a patient leaves an acute care (prospective payment) hospital after receiving complete acute care treatment, and ``transfers,'' situations in which the patient is transferred to another acute care hospital for related care. If a full DRG payment were made to each hospital involved in a transfer situation, irrespective of the length of time the patient spent in the ``sending'' hospital prior to transfer, a strong incentive to increase transfers would be created, thereby unnecessarily endangering patients' health. Therefore, our policy, which is set forth in the regulations at Sec. 412.4, provides that, in a transfer situation, full payment is made to the final discharging hospital and each transferring hospital is paid a per diem rate for each day of the stay, not to exceed the full DRG payment that would have been made if the patient had been discharged without being transferred. Currently, the per diem rate paid to a transferring hospital is determined by dividing the full DRG payment that would have been paid in a nontransfer situation by the geometric mean length of stay for the DRG into which the case falls. Hospitals receive twice the per diem for the first day of the stay and the per diem for every following day up to the full DRG amount. Transferring hospitals are also eligible for outlier payments. Two exceptions to the current transfer payment policy are transfer cases classified into DRG 385 (Neonates, Died or Transferred to Another Acute Care Facility) and DRG 456 (Burns, Transferred to Another Acute Care Facility), which receive the full DRG payment instead of being paid on a per diem basis. Under section 1886(d)(5)(J) of the Act, which was added by section 4407 of the Balanced Budget Act of 1997, a ``qualified discharge'' from one of 10 DRGs selected by the Secretary to a postacute care provider will be treated as a transfer case beginning with discharges on or after October 1, 1998. Section 1886(d)(5)(J)(iii) confers broad authority on the Secretary to select 10 DRGs ``based upon a high volume of discharges classified within such group and a disproportionate use of'' certain postdischarge services. Section 1886(d)(5)(J)(ii) defines a ``qualified discharge'' as a discharge from a prospective payment hospital of an individual whose hospital stay is classified in one of the 10 selected DRGs if, upon such discharge, the individual-- Is admitted to a hospital or hospital unit that is not a prospective payment system hospital; Is admitted to a skilled nursing facility; or Is provided home health services by a home health agency if the services relate to the condition or diagnosis for which the individual received inpatient hospital services and if these services are provided within an appropriate period as determined by the Secretary. The Conference Agreement that accompanied the law noted that ``(t)he Conferees are concerned that Medicare may in some cases be overpaying hospitals for patients who are transferred to a post acute care setting after a very short acute care hospital stay. The Conferees believe that Medicare's payment system should [[Page 40975]] continue to provide hospitals with strong incentives to treat patients in the most effective and efficient manner, while at the same time, adjust PPS [prospective payment system] payments in a manner that accounts for reduced hospital lengths of stay because of a discharge to another setting.'' (H.R. Rep. No. 105-217, 740.) In its March 1, 1997 report, ProPAC expressed similar concerns: ``* * * length of stay declines have been greater in DRGs associated with substantial postacute care use, suggesting a shift in care from hospital inpatient to postacute settings' (pp. 21-22). In fact, based on the latest available data, overall Medicare hospital costs per case have decreased during FYs 1994 and 1995. This unprecedented real decline in costs per case has led to historically high Medicare operating margins (over 10 percent on average). Along with these declining lengths of stay and costs per case, there has been an increase in the utilization of postacute care. In 1990, the rate of skilled nursing facility services per 1,000 Medicare enrollees was 19. By 1995, it had grown to 33. Corresponding numbers for home health agency services are 58 per 1,000 Medicare enrollees during 1990 and 93 per 1,000 enrollees during 1995. Although home health services are not always directly related to a hospitalization episode, there does appear to be a trend toward increased use of home health for the provision of postacute care rehabilitation services. Previous analysis of the percentage of hospital discharges that receive postacute home health care showed a 10.3 percent increase in 1994 compared to 1992. In the May 8, 1998 proposed rule, we discussed our proposals to implement section 1886(d)(5)(J) of the Act. These proposals are set forth below. 1. Selection of 10 DRGs Section 1886(d)(5)(J)(iii)(I) of the Act provides that the Secretary select 10 DRGs based on a high volume of discharges to postacute care and a disproportionate use of postacute care services. Therefore, in order to select the DRGs to be paid as transfers, we first identified those DRGs with the highest percentage of postacute care. We used the FY 1996 MedPAR file because the complete FY 1997 MedPAR file was not available at the time we conducted our analysis. To identify postacute care utilization, we merged hospital inpatient bill files with postacute care bill files matching beneficiary identification numbers and discharge and admission dates. We created this file rather than depend on information concerning discharge destination on the inpatient bill because we have found that the discharge destination codes included on the hospital bills are often inaccurate in identifying discharges to a facility other than another prospective payment hospital. Section 1886(d)(5)(J)(ii)(III) of the Act requires the Secretary to choose an appropriate window of days in which the home health services start in order for the discharge to meet the definition of a transfer. In order to include postdischarge home health utilization in our analysis, we identified all hospital discharges for patients who received any home health care within 7 days after the date of discharge. (As described below in section IV.A.2., we ultimately decided to propose 3 days as the window for home health services.) Starting with the DRG with the highest percentage of postacute care discharges and continuing in descending order, we selected the first 20 DRG's that had a relatively large number of discharges to postacute care (our lower limit was 14,000 cases). In order to select 10 DRG's from the 20 DRG's on our list, for each of the DRG's we considered the volume and percentage of discharges to postacute care that occurred before the mean length of stay and whether the discharges occurring early in the stay were more likely to receive postacute care. The following table lists the 10 DRG's we proposed to include under our expanded transfer definition, their percentage of postacute utilization compared to total cases, and the total number of cases identified as going to postacute care. ------------------------------------------------------------------------ Percent of Number of DRG Title and type of DRG postacute postacute (surgical or medical) utilization cases ------------------------------------------------------------------------ 14................ Specific Cerebrovascular 49.5 186,845 Disorders Except Transient Ischemic Attack (Medical). 113............... Amputation for Circulatory 59.0 28,402 System Disorders Excluding Upper Limb and Toe (Surgical). 209............... Major Joint Limb 71.9 257,875 Reattachment Procedures of Lower Extremity (Surgical). 210............... Hip and Femur Procedures 77.8 111,799 Except Major Joint Age >17 With CC (Surgical). 211............... Hip and Femur Procedures 74.2 19,548 Except Major Joint Age >17 Without CC (Surgical). 236............... Fractures of Hip and 61.2 24,498 Pelvis (Medical). 263............... Skin Graft and/or 49.4 14,499 Debridement for Skin Ulcer or Cellulitis With CC (Surgical). 264............... Skin Graft and/or 39.3 1,328 Debridement for Skin Ulcer or Cellulitis W/O CC (Surgical). 429............... Organic Disturbances and 45.4 19,314 Mental Retardation (Medical). 483............... Tracheostomy Except for 45.3 18,254 Face, Mouth and Neck Diagnoses (Surgical). ------------------------------------------------------------------------ We included DRG 263 on the list because of its ranking in the top 20 DRG's in terms of postacute utilization and volume of discharges to postacute care. DRG's 263 and 264 are paired DRG's; that is, the only difference in the cases assigned to DRG 263 as opposed to DRG 264 is that the patient has a complicating or comorbid condition. If we included only DRG 263 in the list, it would be possible for a transfer case with a relatively short length of stay that should be assigned to DRG 263 and receive a relatively small transfer payment to be assigned instead to DRG 264, and receive the full DRG payment, simply by failing to include the CC diagnosis code on the bill. Therefore, our choice was to either delete DRG 263 from the list or add DRG 264. We decided to include DRG 264 in the proposed list because DRG 263 fully meets all the conditions for inclusion on the list of 10 DRG's. 2. Postacute Care Settings Section 1886(d)(5)(J)(ii) of the Act requires the Secretary to define and pay as transfers cases from one of 10 DRG's selected by the Secretary if the individual is discharged to one of the following settings: A hospital or hospital unit that is not a subsection [1886](d) hospital, that is, a hospital or unit excluded from the inpatient prospective payment system. A skilled nursing facility, that is, a facility that meets the definition of a skilled nursing facility set forth at section 1819 of the Act. Home health services provided by a home health agency, if the services are [[Page 40976]] related to the condition or diagnosis for which the individual received inpatient hospital services, and if the home health services are provided within an appropriate period (as determined by the Secretary). Section 1886(d)(1)(B) of the Act defines the hospitals and hospital units that are excluded from the prospective payment system as the following: psychiatric, rehabilitation, childrens', long-term care, and cancer hospitals and psychiatric and rehabilitation distinct part units of a hospital. Therefore, any discharge from a prospective payment hospital from one of the 10 proposed DRG's that is admitted to one of these types of facilities on the date of discharge from the acute hospital, on or after October 1, 1998, would be considered a transfer and paid accordingly under the prospective payment systems (operating and capital) for inpatient hospital services. We proposed that a discharge from a prospective payment hospital to a skilled nursing facility would include cases discharged from one of the 10 DRG's from an inpatient bed in the hospital to a bed in the same hospital that has been designated for the provision of skilled nursing care (a ``swing'' bed). The swing bed provision allows certain small rural hospitals to furnish services in inpatient beds which, if furnished by a skilled nursing facility, would constitute extended care services. In addition, any patient who receives swing-bed services is deemed to have received extended care services as if furnished by a skilled nursing facility. Thus, if swing beds were not included in the transfer policy, those hospitals with swing bed agreements could move patients assigned to one of the 10 selected DRG's from an inpatient bed to a swing bed and receive payment and receive the full DRG payment. In the proposed rule, we stated that we did not believe that this would be a fair policy in that it would create a payment advantage for swing bed hospitals. Therefore, we proposed that a discharge to a swing bed would be paid as a transfer when the patient is classified to one of the 10 selected DRG's. Section 1886(d)(5)(J)(ii)(III) of the Act states that the discharge of an individual who receives home health services upon discharge will be treated as a transfer if ``such services are provided within an appropriate period (as determined by the Secretary) * * *.'' As discussed above in section IV.A.1, we began our analysis using 7 days (one week) as the time period we would consider. However, after conducting further analysis, we proposed that 3 days after the date of discharge would be a more appropriate timeframe. Based on our analysis of the FY 1996 bills, approximately 90 percent of patients began receiving home health care within 3 days. With regard to an appropriate definition of ``home health services * * * relate[d] to the condition or diagnosis for which the individual received inpatient hospital services * * *'', we considered several possible approaches. Under one approach we could compare the principal diagnosis of the inpatient stay to the diagnosis code indicated on the home health bill, similar to our policy on the 3-day payment window for preadmission services. However, we believe that such a policy is far too restrictive in terms of qualifying discharges for transfer payment. In addition, a hospital would not know when it discharges a patient to home health what diagnosis code the home health agency will put on the bill. Therefore, the hospital would not be able to correctly code the inpatient bill as a transfer or discharge. We also considered proposing that any home health care that begins within the designated timeframe be included ``as related'' in our definition. However, this definition might be too broad and the hospital would not be able to predict which cases should be coded as transfers because the hospital often may not know about home health services that are provided upon discharge but were not ordered or planned for as part of the hospital discharge plan. We proposed that home health services would be considered related to the hospital discharge if the patient is discharged from the hospital with a written plan of care for the provision of home health care services from a home health agency. In this way, the hospital would be fully aware of the status of the patient when discharged and could be held responsible for correctly coding the discharge as a transfer on the inpatient bill. In general, this would mean that the home health service would qualify as a Part A home health benefit under section 1861(tt) of the Act as added by section 4611(b) of the BBA. In the proposed rule, we noted that we plan to compare inpatient bills with home health service bills for care provided within 3 days after discharge. If we find that home health services were provided within the postdischarge window, the hospital will be notified and the hospital payment adjusted unless the hospital can submit documentation verifying the discharge status of the patient. This will alert hospitals if there are problems with their discharge/transfer billing and allow them to adjust their discharge planning process and billing practices. If we find a continued pattern of a hospital billing for cases from the 10 DRG's as discharges and our records indicate that the patients are receiving postacute care services from an excluded hospital, a skilled nursing facility, or within the 3-day home health service window, the hospitals may be investigated for fraudulent or abusive billing practices. 3. Payment Methodology The statute does not dictate the payment methodology we must use for these transfer cases. However, section 1886(d)(5)(J)(i) of the Act provides that the payment amount for a case may not exceed the sum of half the full DRG payment amount and half of the payment amount under the current per diem payment methodology. Based on our analysis comparing the costs per case for the transfers in the 10 DRG's with payments under our current transfer payment methodology, we found that most of the 10 DRG's are appropriately paid using our current methodology (that is, twice the per diem for the first day and the per diem for each subsequent day). In fact, this payment would, on average, slightly exceed costs. However, this is not true of DRG's 209, 210, and 211. For those three DRG's, a disproportionate percentage (about 50 percent) of the costs of the case are incurred on the first day of the stay. Therefore, we stated in the proposed rule that we would pay DRG's 209, 210, and 211 based on 50 percent of the DRG payment for the first day of the stay and 50 percent of the per diem for the remaining days of the stay. The other seven DRG's would be paid under the current transfer payment methodology. We proposed to revise Sec. 412.4 to reflect these policies. In addition, we proposed to delete the reference in Sec. 412.4(d)(2) to DRG 456 (Burns, Transferred to Another Acute Care Facility) because we proposed to replace that DRG and there would no longer be any burn DRG with a transfer designation. As discussed in section II.B.3 of this preamble, we have adopted that DRG change effective for FY 1999. We received a large number of comments concerning this proposal. In general, commenters were opposed to the implementation of any postacute care transfer policy. Acknowledging that the policy is required by statute, most commenters also disagreed with the manner in which we proposed to implement the policy. However, one association representing postacute care providers was supportive of the proposed policy, in general, and our [[Page 40977]] various policy proposals. As discussed in the specific comments and responses that follow, we are implementing the discharge to postacute care provision as set forth in the proposed rule except that we are not including swing beds in the definition of a postacute care setting and we are clarifying the payment methodology for DRGs 209, 210, and 211. Comment: Commenters believed that the postacute care transfer provision penalizes hospitals for providing effective care and creates a perverse incentive for hospitals to keep patients longer. Some commenters suggested that this provision interferes with the practice of medicine by overriding the clinical decision-making process by physicians in determining the most appropriate level of care to provide to their patients. Many commenters stated that the postacute care transfer policy is contrary to the original intent of the prospective payment system. Several commenters urged us either to repeal the entire provision or to support efforts to have it repealed. Response: We disagree that this provision penalizes hospitals for effective care. As noted in the May 8 proposed rule, the Conference Agreement accompanying Public Law 105-33 states that ``Medicare's payment system should continue to provide hospitals with strong incentives to treat patients in the most effective and efficient manner, while at the same time, adjust PPS payments in a manner that accounts for reduced hospital lengths of stay because of a discharge to another setting.'' The transfer provision adjusts payments to hospitals to reflect the reduced lengths of stay arising from the shift of patient care from the acute care setting to the postacute care setting. In addition, because Medicare also often pays for the postacute care portion of beneficiaries' care, the transfer provision appropriately adjusts hospitals' payments to avoid duplicate payments for the care provided during a patient's episode of care. With respect to the payment incentives created by this provision, we would refer the reader to the tables set forth at Appendix D of this final rule. These tables graphically demonstrate payments compared to costs for transfer cases in each of the 10 selected DRGs. These tables show that, across virtually all lengths of stay for each of the DRGs, Medicare will pay in excess of costs even after the implementation of this provision. Thus, the argument that this provision creates perverse incentives and interferes with the appropriate practice of medicine is not persuasive. This policy does not require a change in physician clinical decision-making nor in the manner in which physicians and hospitals practice medicine; it simply addresses the appropriate level of payments once those decisions have been made. We believe a stronger argument can be made that the incentives of the current policy, where hospitals receive the full DRG payment for these DRGs regardless of how long patients remain in the acute care hospital prior to being transferred for postacute care, potentially have a greater impact on clinical decision-making. Simply put, as costs rise with each additional acute care day and payments are fixed, hospitals have a financial incentive to discharge patients as soon as possible. The incentive is less clear, and can be argued to be neutral, to the extent that the marginal payments for an additional acute inpatient care day increase in proportion to the marginal costs of that day. Thus, the postacute care transfer policy does not create perverse incentives for hospitals to keep patients longer; instead, it addresses current incentives to discharge patients as soon as possible. With respect to whether the provision is contrary to the original intent of the prospective payment system, we believe it is entirely consistent with the following statement made in the Federal Register during the first year of the prospective payment system in response to a comment concerning the hospital-to-hospital transfer policy: ``(t)he rationale for per diem payments as part of our transfer policy is that the transferring hospital generally provides only a limited amount of treatment. Therefore, payment of the full prospective payment rate would be unwarranted'' (49 FR 244). We also note that in its earliest update recommendations, the Prospective Payment Assessment Commission (MedPAC's predecessor organization) included what it called a site-of- service substitution adjustment to account for the shifting of portions of inpatient care to other settings. We believe this provision is an appropriate and consistent response to the changing treatment practice of the hospital industry. Comment: A commenter observed that our estimate of the impact of this transfer provision on hospitals' payments per case (a 0.6 percent decrease in payments) results in an overall payment reduction of $600 million for FY 1999. The commenter stated that the Congressional Budget Office (CBO) estimated the impact at $100 million for FY 1999. The commenter believed that this disparity in estimates substantiates claims that the new provision will have undesirable and unintended consequences. Response: We believe the commenter's estimate of the impacts of this provision are overstated. Based on the 0.6 percent decrease in payment per case estimated in our impact analysis, the projected impact of this transfer provision is approximately a $480 million decrease in overall payments. Although this savings estimate is higher than CBO's estimate, we would note that CBO assumed hospitals would change their behavior by keeping patients longer. As we describe in our impact analysis, we do not make any assumptions regarding changes in hospitals' behavior. We would also note that the precision with which one can estimate the savings associated with a provision such as this is highly dependent on the specifications of the provision and the data available to generate an estimate. Unlike the CBO estimate, our estimate reflects the 10 actual DRGs to be included and the latest discharge data to identify which cases would qualify as transfers. Comment: A large number of commenters objected to the inclusion of swing beds as a postacute care setting. Many of these commenters stated that they believed that Congress did not intend that discharges to swing beds be included in the postacute transfer provision. In addition, the commenters were concerned about the negative impact of this policy on rural hospitals and rural health care in general. Two commenters, including MedPAC, supported our proposal concerning swing bed discharges. Response: We proposed to include discharges to swing beds because the services provided in swing beds are exactly the same as the services provided in skilled nursing facilities. That is, a swing-bed hospital is equivalent to a skilled nursing facility when it provides a swing-bed service. Thus, the policy rationale for including discharges to skilled nursing facilities in the postacute care provision would apply equally to discharges to swing beds. Although we are not persuaded by the commenters that, from a payment policy perspective, our proposal to include swing beds in the transfer provision was inappropriate, we understand the commenter's concern that this policy could have an adverse impact on small rural hospitals. Although our analysis shows that the impact on these hospitals is negligible in the aggregate, the impact on individual hospitals may be more significant. We have decided not to include discharges to a swing bed in the expanded transfer definition at this [[Page 40978]] time. We will monitor these discharges closely and may reconsider this decision in the future. We note that section 1886(d)(5)(J)(iv) of the Act requires the Secretary to include a description of the effect of the postacute care transfer policy in the FY 2001 hospital inpatient prospective payment system proposed rule. Comment: Commenters requested clarification of our policy concerning transfers to skilled nursing facilities. First, the commenters questioned the Secretary's authority to include as transfers those discharges to nursing homes that are not certified by Medicare. In addition, the commenters believed that patients discharged to a Medicare-certified skilled nursing facility for custodial care should not be included. The commenters also urged us to limit application of the transfer policy to discharges to skilled nursing facilities in cases where the patient receives Medicare-covered postacute care. Response: Section 1886(d)(5)(J)(ii) of the Act defines a ``qualified discharge'' in part as a discharge of an individual from a prospective payment system hospital, if upon such discharge, the individual is ``* * * admitted to a skilled nursing facility. * * *'' There is no language in the statute that further defines skilled nursing facility. In the proposed rule, we stated that a discharge to a facility that meets the definition of a skilled nursing facility set forth at section 1819 of the Act would be considered a transfer. Discharges to nursing homes that are not certified by Medicare as skilled nursing facilities, or distinct parts of nursing homes that are not certified as skilled nursing facilities, would not be considered transfers. However, we do not believe it would be appropriate from either a legal or policy perspective to limit the transfer definition to situations where a patient is transferred to a skilled nursing facility for noncovered services. The statute does not limit application of the transfer definition to ``covered'' skilled nursing facility services. Moreover, there are several policy reasons why we would not adopt such a policy. First, it would place an added administrative burden upon the hospital to evaluate the patient's eligibility for covered skilled nursing services. Second, it would create incentives for providing noncovered postacute care that could potentially place beneficiaries at medical and financial risk. Third, it would be inconsistent with existing transfer policy (from one acute care hospital to another acute care hospital), which does not limit the definition of a transfer to those cases in which a patient receives Medicare-covered services at the receiving hospital. Finally, the basic rationale for the transfer policy (that is, adjusting hospital payments to reflect reduced hospital costs due to discharge to a postacute care facility) applies regardless of whether the postacute care is covered by Medicare. Therefore, our final regulations provide that all discharges from the 10 specified DRGs admitted to a skilled nursing facility will be defined as transfers, regardless of the coverage status of that admission. Comment: One commenter believes that patients who were admitted to a skilled nursing facility any time within 30 days after the date of discharge (the so-called 30-day skilled nursing facility eligibility window) and who received care related to the inpatient stay will be considered a transfer under this policy. The commenter is concerned that hospitals will be expected to track patients for this period of time and be held accountable for their actions in such situations. Response: In order to be considered a transfer, the patient must be admitted directly from the hospital to the skilled nursing facility. If the patient is not admitted directly to a skilled nursing facility, it would not constitute a transfer situation, even if care begins within the 30-day eligibility window and is related to the acute care hospital stay. Comment: One commenter suggested that the expanded transfer definition should apply only in cases where the patient is transferred within a hospital system, that is, the patient is discharged to an entity that is related to or owned by the hospital. A transfer to an independent postacute care entity would be defined as a discharge. Response: Section 1886(d)(5)(J)(ii) of the Act defines a qualified discharge from a prospective payment hospital as one in which the individual, upon discharge, ``* * * is admitted as an inpatient of a hospital or hospital unit excluded that is not a subsection (d) hospital * * * is admitted to a skilled nursing facility * * * is provided home health services from a home health agency. * * *'' The statute or the conference report does not limit the applicability of this provision to postacute care providers that are owned by or related to the discharging hospital. In addition, we do not believe that ownership of or affiliation with the postacute care providers is the overriding concern that led to the enactment of this provision. Although a hospital that owns or is related to the postacute care provider has an even greater financial incentive to transfer a patient early in the hospital stay to that facility, the current incentive to the hospital itself to discharge the patient as soon as possible is the same whether or not it owns or is related to the postacute care provider. Finally, if the transfer definition were based on a hospital's affiliation with the postacute provider, it would create a strong incentive to reconfigure the hospital's corporate structure to avoid being included under the provision. Comment: One commenter suggested that psychiatric hospitals and units be excluded from the provision because the postacute care services furnished by these facilities are unrelated to a medical hospitalization. Response: As a legal matter, section 1886(d)(5)(J)(ii)(I) of the Act includes all hospitals and hospital units excluded from the prospective payment system. This definition covers psychiatric hospitals and units. As a policy matter, we also strongly believe that transfers to psychiatric hospitals and units should be included under this provision. Inpatient care furnished by hospitals is not limited to diseases and disorders of the body, but is also furnished to patients with mental diseases and disorders as evidenced by the nine DRGs devoted solely to these conditions. Furthermore, exempting psychiatric hospitals and units from the provision could create an incentive to temporarily transfer patients who need postacute care to a psychiatric hospital or unit setting as a way of avoiding the transfer payment, thus delaying the appropriate medical care for the patient. Comment: Several commenters disagreed with our proposal to include as transfers all discharges from the 10 specified DRGs to home health care that begins within 3 days after the date of discharge. The commenters argued that postacute care that begins 3 days after discharge should not be considered a substitute for inpatient hospital care. Although MedPAC agreed with these commenters that home health services furnished after a delay of more than one day may not necessarily be regarded as substituting for inpatient acute care, they also noted that a 3-day window allows for the fact that most home health patients do not receive care every day as well as those occasions in which there may be a delay in arranging for the provision of planned care. The Commission also stated that a shorter period may create a stronger incentive to delay the provision of necessary treatment beyond the window so the hospital can receive the full DRG payment. Another commenter [[Page 40979]] supported 3 days as an appropriate period of time. Those commenters who recommended an alternative number of days for the home health window universally stated that a discharge to home health care should be considered a transfer only if the patients begin to receive home health care on the day of discharge. One commenter argued that a 3-day window would lead to either needlessly prolonged hospital stays or delayed home health care. Another commenter questioned why we would not want patients transferred to home health care as soon as possible. Response: The statute defines ``qualified discharge'' to include discharges where the individual is provided home health care ``within an appropriate period (as determined by the Secretary).'' We continue to believe a 3-day window for home health services is appropriate. Home health care is a less-structured and more flexible means of providing postacute care because it is provided not in an institutional setting but rather in the patient's home. We believe that a 3-day window provides flexibility in situations where home care may not be available or medically appropriate immediately upon discharge. It is also of sufficient length to discourage hospitals and physicians from delaying the initiation of necessary postacute care, while being short enough to avoid placing an undue burden upon hospitals that may want to delay submitting the inpatient hospital claim until they verify whether or not home health care has begun within the 3 days. We do not believe that it is appropriate to limit the transfer definition to situations where home care begins on the same day as the patient is discharged from the hospital. Our analysis indicates that currently less than 8 percent of discharged patients who receive home health services begin receiving those services on the date of discharge. It is unreasonable to expect that patients who are discharged in the late afternoon or evening would receive a home health visit that same day. Furthermore, we believe the financial incentive to delay needed home care for only a matter of hours would be overwhelming if we limited the definition to the same day. As we noted in the proposed rule, approximately 90 percent of patients who receive home health services after an inpatient hospital stay began their treatment within 3 days after the date of discharge. We believe 3 days accommodates current practices, while also being sufficiently narrow to allow hospitals to determine whether the care was actually delivered prior to submitting the bill. We intend to monitor this aspect of the policy through case review in order to track any changes in hospital practices that may indicate that we need to revise our window definition. Comment: One commenter argued that the best method to determine whether postacute home health services are related to the inpatient stay would be to match the principal diagnosis codes on the inpatient and home health bills. The commenter believed this would alleviate situations where the patient is discharged from the hospital with a written plan for the provision of home health services, but the services are related to a medical condition other than the condition responsible for the inpatient stay. In addition, the commenter noted that matching principal diagnosis codes would be consistent with current policy for the 3-day window for preadmission services. Response: We disagree that the determination of whether home health care is related to the acute hospitalization should be based on the presence of identical diagnosis codes on the inpatient and home health bills. This approach would rely on the coding practices of the providers involved. Providers, especially postacute care providers, frequently have the discretion to select from several possible diagnosis codes. A common practice of postacute care providers is to use the V57 diagnosis code category (care involving use of rehabilitation procedures) as principal because those codes best describe the reason for the postacute care. However, this code is seldom used by hospitals for acute care discharges because they are instructed by coding rules to code as principal the condition that required the hospital admission as determined at the time of discharge. In fact, if the hospitals coded discharges with the rehabilitation codes as principal, the discharges would never be included in the postacute care policy because those discharges would never be classified to one of the 10 selected DRGs. We believe our proposed policy on this issue is preferable. We note that hospitals that code a discharge to home health will be permitted to indicate through a condition code on the inpatient bill that the hospital's discharge plan does not call for home care related to the hospitalization, but that other nonrelated home care is appropriate. This way, the hospital will make a conscious selection that the home care the patient is to receive is not related to the hospitalization, and would be expected to have documentation in the patient's records to that effect. Comment: In the context of discussing the home health window, MedPAC questioned whether the same day requirement for admission to an excluded hospital or unit or a skilled nursing facility was too limited. The Commission suggested expanding the definition to account for a 24-hour period following discharge. Response: In describing which discharges to excluded hospitals and units or skilled nursing facilities should be treated as a transfer, the statute states that the patient is admitted to the facility upon discharge from the hospital. We believe that Congress intended that the policy apply to situations when the patient moves from the hospital directly to the excluded facility or the skilled nursing facility. Therefore, unless a patient is being transported from the hospital to the other facility in the middle of the night, the discharge and admission should occur on the same calendar day. We note that a direct transfer that spans midnight and results in a one-day difference in the discharge and admission dates will be considered a transfer for purposes of this policy. Comment: Many commenters indicated the discharge to postacute care provision will be an administrative burden for hospitals. Because Medicare beneficiaries are free to obtain services without a hospital referral, hospitals are concerned that they will be subject to allegations of fraud and abuse if they discharge a beneficiary to home with no plan of care for home health services and the beneficiary subsequently receives postacute care without the hospital's knowledge. These hospitals believe that they may be forced to hold bills for the 10 DRGS when they discharge a patient to self-care at home so they can follow-up and ensure that the patient did not receive postacute care. Another commenter is disturbed by our discussion in the proposed rule concerning future actions we may take if we find continued patterns of a hospital billing postacute transfer cases as discharges, including the possibility that hospitals may be investigated for fraudulent or abusive billing practices. The commenter believes that our language was too strong and that we are not allowing a period of transition in which hospitals may make honest billing errors as they adjust to this new policy. Finally, commenters suggested that we clarify when hospitals are responsible for knowing that a case is transferred for postacute care. [[Page 40980]] Response: We recognize there may occasionally be cases where a hospital believes it is discharging a patient to home or another setting not included in the postacute transfer definition, and a physician orders postacute care for the patient without notifying the hospital. Although these cases would be considered transfers under this provision, we do not believe that such instances, where they occur truly without knowledge of the hospital, constitute fraudulent actions. As we indicated in the proposed rule, we intend to monitor postacute care cases to evaluate whether such situations occur with unlikely frequency at specific hospitals and we will investigate the circumstances in those instances. Although we recognize honest mistakes will occur, we do not believe it is inappropriate to put hospitals on notice that we reserve the right to investigate those with aberrant patterns of inaccurate billing on these cases. While it is reasonable to assume there will be a learning curve in terms of hospitals' billing practices as these changes are implemented, we also take seriously our responsibility to protect the Medicare trust fund. Our intention in including a discussion of this issue in the proposed rule was an attempt to avoid any misunderstanding in terms of our commitment to ensure the correct implementation of this provision. In response to the request for clarification about the hospital's responsibility for knowing when a transfer occurs, the hospital is responsible for coding the bill based on its discharge plan for the patient, or if it finds out subsequently that postacute care occurred, it is responsible for either coding the original bill as a transfer or submitting an adjustment bill. We have consulted with the National Uniform Billing Committee (NUBC) to ensure that the appropriate changes are made on the claims form to enable hospitals to accurately code these cases and to submit corrections to them when additional information affecting the patient's discharge status code becomes available after the bill is submitted. Comment: One commenter recommended that we establish a hierarchical decision process for determining whether a discharge to home health services qualifies as a transfer. This commenter believed that the overriding consideration should be whether the services are related to the hospital stay. This commenter suggested that any home care ordered in the discharge plan should constitute related home health care, regardless of when it is provided. Response: Congress directed the Secretary to determine the appropriate time period within which the provision of home health services would trigger a transfer payment. Services provided outside that window, even if related to the hospital stay, would not result in the discharge being considered a transfer. In addition, we believe that a time limit is consistent with the concern that these transfer cases are predominantly situations where care is being shifted from the acute setting to a postacute care setting. If a patient is discharged to home and does not need home health care for several days, there may be little, if any, shift of acute care services to the postacute care setting. Comment: One commenter stated that we should specify that the written plan of care for home health services should be defined clearly as ``a specific order by the patient's physician in the hospital medical record that directs the hospital to arrange for home health services upon discharge.'' Response: We do not believe that it is necessary to specify the precise definition of what a written plan of care for home health services must entail. We note that we would deem a case to be a transfer if care related to the discharge was provided within 3 days after the date of discharge even if the hospital had no written plan of care. Comment: A representative of physical therapists expressed concern that the 3-day window for home health services may influence hospitals to wait until after the 3 days to initiate home health services. This commenter is also concerned that our proposal to identify related home health services based on the written plan of care by the hospital at the time of discharge may discourage hospitals from planning for home health, resulting in uncoordinated and delayed postacute care following discharge. Response: We believe there are sufficient protections against hospitals inappropriately delaying home health care. First, the provision of home health care is ordered by the patient's physician orders. We believe physicians will be reluctant to compromise their patients' treatment by inappropriately delaying home health care. In addition, we will monitor hospitals' discharge patterns to home health for evidence that care is being routinely delayed until the fourth day after discharge and intend to aggressively pursue situations where abuse is evident. If evidence of a pattern of abuse is found, we will address it through appropriate policy changes in the FY 2001 proposed rule. With respect to the commenter's concern that identifying related home health services based on the hospital's written plan of care may create a disincentive to plan home care, we will also be able to identify those cases where home health services were received within 3 days of discharge and the hospital indicated that the patient was discharged home with no plan for home health services. As we noted above, we recognize there will be a certain percentage of cases where home care is arranged after release from the hospital; however, we would expect such situations to be relatively rare. Comment: One commenter, representing medical rehabilitation providers, expressed concern that this provision may change hospitals' referral patterns, delaying the initiation of rehabilitation services. The commenter suggested that we collect the following information from prospective payment hospitals to monitor their referral patterns: Site of referral for cases in the 10 DRGs, including discharge to home without postacute care. Date from onset and length of stay prior to referral, by DRG. General medical condition and functional status of the patient if the hospital normally collects functional information. In addition, HCFA should collect the following information from postacute care providers: The DRG assigned to the acute care hospitalization. The date from onset and date of referral to the postacute care provider. For patients referred for rehabilitation services to a rehabilitation hospital or unit, the functional status of the patient on admission to and at discharge from the rehabilitation provider. The commenter noted that over 90 percent of rehabilitation providers already use functional assessment tools, therefore, this data collection would not be excessively burdensome. Response: We appreciate this commenter's concerns regarding any potentially adverse effects of this provision with respect to beneficiaries' health. We already collect most of the hospital data suggested by the commenter (with the exception of patients' functional status and medical condition, though even this could be accessed on a limited basis). Similarly, for postacute providers, the first two items of data are already readily available in our system. As we have described above, we intend to use these data to monitor providers' behavior after implementation of this policy. Comment: Commenters requested that we require the fiscal intermediaries to [[Page 40981]] automatically adjust the payments received by the hospital when the hospital codes a case as a discharge and no bill is ever received for postacute care services. In making this request, the commenters referred to the process we described in the proposed rule in which we would compare the discharge status coded on the hospital bills with postacute care bills received to determine whether qualifying postacute care was provided when the hospital billed the case as a discharge. Response: As noted above, hospitals will be able to submit corrections to their discharge status codes when they determine that previously submitted bills are incorrect. It would be impractical to require the fiscal intermediaries to adjust payments for cases coded as transfers when no matching postacute care bill is identified. Such a requirement raises a potential scenario where a case may be inappropriately adjusted upward because the matching postacute bill has not entered the claims system at the time the bill comparison is made. The prescribed period of time within which a provider may submit a bill for Medicare payment is relatively long and we believe it would be impractical for each intermediary to reprocess already paid bills based solely on the absence of a matching postacute care bill. In addition, we note that there may be occasions when no postacute care bill is submitted even though the patient was discharged to that care. For example, as we discussed above, if a patient is transferred to a skilled nursing facility and receives noncovered care, there will be no bill in the Medicare claims files. We believe it is preferable to require hospitals to submit bill adjustments. Comment: One commenter was unclear about how postacute care transfers will be identified in the billing process. Specifically, the commenter questioned whether the hospital will indicate a transfer by the discharge status code or whether the identification will occur by matching the acute hospitalization bill against a postacute bill at the fiscal intermediary. Response: Transfer cases will be identified based on the discharge status code listed on the hospital claim form (the HCFA-1490, also known as the UB-92). As noted above, we have consulted with the NUBC to ensure that the appropriate changes are made on the claims form to enable hospitals to accurately code these cases. The language in the proposed rule concerning a process of matching the date of discharge from the acute hospital stay with the date that postacute care services begin was a description of the process that HCFA will use as a check to verify the accuracy of the discharge codes. Comment: One commenter asked whether the discharge destination code ``08,'' which is described as ``Discharged/transferred to home under care of a Home IV (intravenous) provider,'' would be used to identify a transfer. This commenter was also concerned about whether code ``05,'' which is described as ``Discharged/transferred to another type of institution for inpatient care or referred for outpatient services to another institution'' would be sufficient to identify transfers to excluded hospitals or units. Response: Discharge code ``08'' will not trigger a transfer payment because it should not be used in situations where a patient is receiving IV services under the Medicare home health benefit. Rather, code ``06'' should be used to signify all care provided by a home health agency under the Medicare home health benefit. With respect to discharge code ``05,'' the NUBC is discussing what additional codes need to be added or what current codes may be revised to allow for more specific coding to distinguish transfer situations from nontransfers. Instructions on the discharge codes will be provided to the fiscal intermediaries and, thereafter, to the hospitals before the effective date of the postacute transfer provision (that is, October 1, 1998). Comment: Several commenters suggested that DRG 483 should not be included as one of the 10 DRGs under this provision. The commenters believed that this DRG is not clinically homogeneous and includes many different conditions with different expected lengths of stay. They also stated that our analysis showed that transfers from this DRG would be paid below costs for almost every day below the mean length of stay. One commenter indicated it appeared this DRG was singled out for specific treatment. MedPAC commented that the criteria we used to select the 10 DRGs was reasonable, although it indicated that the list is fairly narrow in the types of conditions or procedures represented. Therefore, when we consider an expansion of this list, MedPAC recommended that we include coronary surgery DRGs, such as the coronary bypass DRGs (106, 107, and 109), and the pneumonia DRGs (89, 90, or 91). Response: As described in the proposed rule and above in this section of the preamble, the 10 DRGs were selected based on the criteria specified in the statute, that is, the DRGs exhibit a high volume and disproportionate percentage of postacute cases. None of the 10 DRGs were predetermined. With respect to DRG 483, a significant percentage of discharges (over 45 percent are transferred to postacute care. This places it in the top 25 DRGs in terms of postacute care utilization. Of those 25 DRGs, it is ranked 9th in terms of the volume of cases receiving postacute care. We believe these factors justify its inclusion. In addition, contrary to the commenter's statement, our analysis of payments and costs for transfers in this DRG shows that average payments exceed average costs for all but those cases transferred very early in the stay (before the 6th day in a DRG with an average length of stay of 34 days). (See the table for DRG 483 in Appendix D of this final rule.) The marginal costs per day for this DRG are consistent with and are accommodated almost perfectly by the transfer per diem payment methodology. We appreciate MedPAC's support regarding our selection criteria and will take its recommendations regarding additional DRGs into consideration in our future analysis. Comment: Some commenters believe that the process we used to select the 10 DRGs did not reflect the intent of Congress. They suggested that, in selecting the 10 DRGs, we should include an evaluation of whether a DRG was prone to inappropriate use of postacute care. Response: Section 1886(d)(5)(J)(iii)(I) of the Act provides that the affected DRGs are ``* * *10 diagnosis-related groups selected by the Secretary based on a high-volume of discharges classified within such groups and a disproportionate use of post discharge [sic] services * * *.'' This language does not direct the Secretary to select the 10 DRGs based upon their vulnerability to inappropriate use of postacute care. As stated earlier, the postacure care transfer provision adjusts hospital payments to reflect the reduced lengths of stay arising from the shift of care to the postacute care setting. Comment: One commenter was offended by the rationale stated in the proposed rule for including DRG 264 (Skin Graft and/or Debridement for Skin Ulcer or Cellulitis without complication or comorbidity (CC)) in the list of 10 DRGs. The commenter argued that no medical record coder would intentionally fail to list a CC in order to avoid the transfer payment for a case that groups to DRG 263 (Skin Graft and/or Debridement for Skin Ulcer or Cellulitis With CC). The commenter noted that this would be an illegal, [[Page 40982]] fraudulent act on the part of the coder and should not be used as a deciding factor in the methodology for selecting the 10 DRGs. Response: In making our selection of the 10 DRGs, we decided to include paired DRGs if one of them met our criteria. While we do not believe that medical record coders will exclude a CC code in their list of diagnosis codes, the hospital claim is not generally submitted to HCFA by the coder, but rather by a billing office where information included on the claim is frequently subject to additional review. By including DRG 264, we hope to avoid any questions or issues concerning the accurate coding of a particular case involving skin graft and debridement. Comment: Several commenters stated that the alternative payment methodology for DRGs 209, 210, and 211 described in the proposed rule would not pay the full DRG amount until one day after the geometric mean length of stay for the DRG. This result is contrary to the usual per diem payment methodology where the full DRG payment is received one day before the geometric mean length of stay. Response: The alternative payment methodology in the proposed rule was described as ``50 percent of the DRG payment for the first day of the stay, and 50 percent of the per diem for the remaining days of the stay.'' This wording imprecisely described our proposed policy. The alternative payment methodology proposed for DRGs 209, 210, and 211 is equal to 50 percent of the DRG payment plus 50 percent of the amount which would be paid under our per diem methodology. Under this formula, on day one of a postacute care transfer, hospitals would receive one- half the DRG payment amount plus the per diem payment for the DRG (one- half the usual transfer payment of double the per diem for day one). For each subsequent day prior to transfer, hospitals receive one-half the per diem up to the full DRG payment, which is reached one day prior to the geometric mean length of stay for the DRG. We note that, although we inaccurately described the methodology, we used the correct formula in calculating the budget neutrality factors and outlier thresholds in the proposed rule. Comment: One commenter believed that the alternative payment methodology used for DRGs 209, 210, and 211 should be used for all 10 of the DRGs selected under the postacute care transfer provision. The commenter argued that for postacute care transfers, unlike transfers under our current transfer policy, the hospital provides all necessary acute care services to the patient, regardless of length of stay, before transferring the patient to postacute care. Response: As noted above, we believe care previously provided in the acute care setting increasingly has been shifted to the postacute setting. Therefore, we do not agree with the commenter's belief that these cases are significantly different from those considered transfers under our current definition of transfers; in both situations, the length of stay is reduced and presumably a hospital furnished fewer services and incurs lower costs relative to a typical discharge. Furthermore, as demonstrated in the tables comparing average payments and costs for these DRGs in Appendix D, the seven DRGs that will be paid under the current per diem methodology have a gradual increase in costs as length of stay rises, consistent with the gradual increase in payments under our current per diem methodology. Therefore, we are not expanding the application of the alternative payment methodology beyond the three DRGs identified in the proposed rule. Comment: MedPAC suggested we may wish to evaluate whether the alternative payment methodology for postacute transfers in DRGs 209, 210, and 211 should be expanded to our policy for transfers between two acute care hospitals. Response: We have evaluated our transfer payment formula for our current transfer policy in the past and revised it to pay double the per diem amount for the first day of a transfer stay. Because the majority of cases that are transferred from one acute care hospital to another result in the case being assigned to a medical DRG, our analysis indicated that the current per diem payment (with a double payment on the first day) accurately reflects the costs of these cases, as it does for the seven DRGs paid under the per diem methodology under the postacute transfer provision. Although we do not plan further changes in the payment methodology for transfers to another acute care hospital, we will continue to evaluate the potential for further refinements in this policy, particularly in light of the changes introduced in this final rule. Comment: One commenter requested clarification of how the indirect medical education (IME) and disproportionate share hospital (DSH) adjustments are treated under the transfer payment methodology. This commenter also requested clarification regarding the outlier payment calculation for transfer cases and recommended that the transfer payment rather than the DRG payment serve as the comparative basis for determining whether a transfer case qualifies as an outlier. Commenters also indicated some confusion as to when full payment would be made under the transfer methodology in situations where the geometric mean length of stay for a DRG is not a whole number, for example, 9.8 days. Response: The IME and DSH payments are determined in accordance with Secs. 412.105(e) and 412.106(a)(2), respectively, by applying the IME and DSH adjustment factors calculated under those sections to the DRG revenue. In the case of a transfer occurring before the average length of stay, the applicable IME or DSH factor would be applied to the DRG revenue determined under the applicable transfer payment methodology. With respect to outliers for transfer cases, the methodology suggested by the commenter is actually the methodology we use to determine outliers for these cases. In the September 1, 1995 Federal Register, we described how the cost outlier threshold is calculated for transfers (60 FR 45804). The outlier threshold for transfer cases reflects the fact that transfer cases receive a reduced payment amount. Specifically, the threshold for transfers paid under the current per diem methodology is equal to the fixed loss outlier threshold for all cases ($11,100 for FY 1999) divided by the geometric mean length of stay for the DRG, multiplied by the length of stay prior to transfer, plus one day. For postacute transfers in DRGs 209, 210, and 211, the outlier threshold is determined by dividing the fixed loss outlier threshold for all cases by the geometric mean length of stay for the DRG, multiplied by the sum of half the geometric mean and half the length of stay for the case, plus one. We note that we are making a conforming change in Sec. 412.80(b), which describes outlier payments for transfers, to incorporate the revisions we are making in the transfer policy. Finally, in the case of a DRG with a geometric mean length of stay of 9.8 days, full payment would be received on day 9. The following table illustrates this point, using DRGs 209 and 236 with geometric mean lengths of stay of 4.9 and 4.1 days, respectively. [[Page 40983]] ------------------------------------------------------------------------ DRG 209 236 ------------------------------------------------------------------------ Full DRG Payment Amount \1\............. $8,400.32 $2,790.60 Per Diem Amount......................... 2,048.86 680.63 Payment for Transfer on Day 1 \2\....... 6,249.02 1,361.26 Payment for Transfer on Day 2........... 7,273.45 2,041.89 Payment for Transfer on Day 3........... 8,297.88 2,722.52 Payment for Transfer on Day 4 \3\....... 8,400.32 2,790.60 ------------------------------------------------------------------------ \1\ This amount is determined using the other areas national standardized amount from Table 1A in Section VI of the addendum to this final rule. The respective relative DRG weights are taken from Table 5. For DRG 209, the relative weight is 2.1803, and for DRG 236 it is 0.7243. It assumes a wage index of 1.0, and no IME or DSH payments. Any IME or DSH payments would be factored into the transfer amount as described above. \2\ For DRG 209, the payment amount is equal to one-half of the full DRG payment amount ($4,200.16) plus the per diem amount ($2,048.86). For DRG 236, the payment amount is equal to double the per diem amount. \3\ Total payment is limited to the full DRG amount (with the exception of outlier cases), rather the result of an additional per diem amount (or half the per diem). Comment: A few commenters stated that because average lengths of stay vary by geographic region, the transfer policy punishes those regions with average lengths of stay less than the mean. They recommended that an adjustment factor be developed to recognize this disparity or that regional averages should be used to compute the per diem amount. Response: We recognize that lengths of stay vary by region and are generally lower in the west, particularly compared to the northeast. In addition, regions with shorter lengths of stay tend to also have lower average costs due to the fewer number of days that patients spend in the hospital. One of the reasons for this variation is the greater reliance on postacute care earlier in the stay in those areas with lower average lengths of stay. We do not believe it would be appropriate to base the transfer payment methodology on regional average lengths of stay. The national standardized amounts, which apply across all regions, reflect costs and lengths of stay across all regions. If a hospital in one region has a case with certain patient characteristics and a hospital in another region has a case with identical patient characteristics (including the same length of stay), we see no reason to have a rule under which one hospital would receive the full DRG payment but the other hospital would receive a transfer payment. Comment: One commenter believed that, in lieu of expanding the transfer definition, it would make more sense to recalibrate the 10 DRGs to better reflect the recent reductions in lengths of stay and costs for these categories. Response: All of the DRGs are recalibrated annually, using the latest available charge data for Medicare beneficiaries. Because of the recalibration process, a reduction in the relative weights of certain DRGs results in an increase in the weights of other DRGs. Therefore, there are no overall reductions in Medicare payments to hospitals. That is, although the hospital will receive a reduced payment through lower weights for the DRGs affected by the shift toward greater utilization of postacute care early in a stay, it will receive greater payment for the DRGs in which the weight is increased because there is no reduction in overall costs. In addition, any reduction in payment for the selected DRGs is shared by all hospitals including those that have not reduced their average length of stay and costs through the increased use of postacute care. We believe that any change in Medicare payment because of the early transfer of acute care patients to postacute care should be targeted at those hospitals that have actually incorporated this practice into their patient care. Comment: Another commenter noted that, if these cases are to be treated as transfers for payment, they should be treated that way for recalibration as well. Response: We agree. In the proposed rule, we did not revise the discussion of the recalibration process to specifically mention the postacute transfers, but we did treat these cases as transfers during the recalibration process that resulted in the DRG weights set forth in that rule. For purposes of the DRG recalibration, transfer cases, including the postacute transfer cases, are counted as a fraction of a discharge based on the length of stay, thereby reducing proportionately the contribution of the charges for the case toward the average charges for the DRG. This process effectively inflates the charges of a transfer case to what they would have been had the patient's length of stay been equal to the geometric mean length of stay. If we do not perform this calculation, these cases would receive reduced payment because they are transfers, but be treated as discharges in recalibration, lowering the relative weights for affected DRGs. Comment: One commenter questioned whether the postacute care transfer provision will have any effect on the payments made by Medicare to the postacute providers. Response: The only payment implication of this provision is to affect the prospective payment for the acute inpatient hospitalization. Medicare payment to any postacute providers involved in the stay are not affected by this policy. B. Rural Referral Centers (Sec. 412.96) Under the authority of section 1886(d)(5)(C)(I) of the Act, Sec. 412.96 sets forth the criteria a hospital must meet in order to receive special treatment under the prospective payment system as a rural referral center. For discharges occurring before October 1, 1994, rural referral centers received the benefit of payment based on the other urban rather than the rural standardized amount. As of that date, the other urban and rural standardized amounts were the same. However, rural referral centers continue to receive special treatment under both the disproportionate share hospital payment adjustment and the criteria for geographic reclassification. One of the criteria under which a rural hospital may qualify as a rural referral center is to have 275 or more beds available for use. A rural hospital that does not meet the bed size criterion can qualify as a rural referral center if the hospital meets two mandatory criteria (specifying a minimum case-mix index and a minimum number of discharges) and at least one of the three optional criteria (relating to specialty composition of medical staff, source of inpatients, or volume of referrals). With respect to the two mandatory criteria, a hospital may be classified as a rural referral center if its-- Case-mix index is at least equal to the lower of the median case-mix index for urban hospitals in its census region, excluding hospitals with approved teaching programs, or the median case-mix index for all urban hospitals nationally; and [[Page 40984]] Number of discharges is at least 5,000 discharges per year or, if fewer, the median number of discharges for urban hospitals in the census region in which the hospital is located. (The number of discharges criterion for an osteopathic hospital is at least 3,000 discharges per year.) 1. Case-Mix Index Section 412.96(c)(1) provides that HCFA will establish updated national and regional case-mix index values in each year's annual notice of prospective payment rates for purposes of determining rural referral center status. The methodology we use to determine the proposed national and regional case-mix index values, is set forth in regulations at Sec. 412.96(c)(1)(ii). The proposed national case-mix index value included all urban hospitals nationwide, and the proposed regional values were the median values of urban hospitals within each census region, excluding those with approved teaching programs (that is, those hospitals receiving indirect medical education payments as provided in Sec. 412.105). These values were based on discharges occurring during FY 1997 (October 1, 1996 through September 30, 1997) and include bills posted to HCFA's records through December 1997. Therefore, in addition to meeting other criteria, for hospitals with fewer than 275 beds, we proposed that to qualify for initial rural referral center status for cost reporting periods beginning on or after October 1, 1998, a hospital's case-mix index value for FY 1997 would have to be at least-- 1.3578; or Equal to the median case-mix index value for urban hospitals (excluding hospitals with approved teaching programs as identified in Sec. 412.105) calculated by HCFA for the census region in which the hospital is located. (See the table set forth in the May 8, 1998 proposed rule at 63 FR 25593.) Based on the latest data available (FY 1997 bills received through March 31, 1998), the final national case-mix value is 1.3590 and the median case-mix values by region are set forth in the table below: ------------------------------------------------------------------------ Case-mix Region index value ------------------------------------------------------------------------ 1. New England (CT, ME, MA, NH, RI, VT)...................... 1.2490 2. Middle Atlantic (PA, NJ, NY).............................. 1.2519 3. South Atlantic (DE, DC, FL, GA, MD, NC, SC, VA, WV)....... 1.3474 4. East North Central (IL, IN, MI, OH, WI)................... 1.2711 5. East South Central (AL, KY, MS, TN)....................... 1.3042 6. West North Central (IA, KS, MN, MO, NE, ND, SD)........... 1.2325 7. West South Central (AR, LA, OK, TX)....................... 1.3326 8. Mountain (AZ, CO, ID, MT, NV, NM, UT, WY)................. 1.3726 9. Pacific (AK, CA, HI, OR, WA).............................. 1.3427 ------------------------------------------------------------------------ For the benefit of hospitals seeking to qualify as referral centers or those wishing to know how their case-mix index value compares to the criteria, we are publishing each hospital's FY 1997 case-mix index value in Table 3C in section VI. of the Addendum to this final rule. In keeping with our policy on discharges, these case-mix index values are computed based on all Medicare patient discharges subject to DRG-based payment. 2. Discharges Section 412.96(c)(2)(I) provides that HCFA will set forth the national and regional numbers of discharges in each year's annual notice of prospective payment rates for purposes of determining referral center status. As specified in section 1886(d)(5)(C)(ii) of the Act, the national standard is set at 5,000 discharges. However, we proposed to update the regional standards. The proposed regional standards were based on discharges for urban hospitals' cost reporting periods that began during FY 1996 (that is, October 1, 1995 through September 30, 1996). That is the latest year for which we have complete discharge data available. Therefore, in addition to meeting other criteria, we proposed that to qualify for initial rural referral center status for cost reporting periods beginning on or after October 1, 1998, the number of discharges a hospital must have for its cost reporting period that began during FY 1997 would have to be at least-- 5,000; or Equal to the median number of discharges for urban hospitals in the census region in which the hospital is located. (See the table set forth in the May 8, 1998 proposed rule at 63 FR 65594.) Based on the latest discharge data available for FY 1996, the final median numbers of discharges for urban hospitals by census region areas are as follows: ------------------------------------------------------------------------ Number of Region discharges ------------------------------------------------------------------------ 1. New England (CT, ME, MA, NH, RI, VT)..................... 6,672 2. Middle Atlantic (PA, NJ, NY)............................. 8,676 3. South Atlantic (DE, DC, FL, GA, MD, NC, SC, VA, WV)...... 7,753 4. East North Central (IL, IN, MI, OH, WI).................. 7,346 5. East South Central (AL, KY, MS, TN)...................... 6,741 6. West North Central (IA, KS, MN, MO, NE, ND, SD).......... 5,346 7. West South Central (AR, LA, OK, TX)...................... 5,251 8. Mountain (AZ, CO, ID, MT, NV, NM, UT, WY)................ 7,992 9. Pacific (AK, CA, HI, OR, WA)............................. 5,993 ------------------------------------------------------------------------ We note that the number of discharges for hospitals in each census region is greater than the national standard of 5,000 discharges. Therefore, 5,000 discharges is the minimum criterion for all hospitals. We reiterate that, to qualify for rural referral center status for cost reporting periods beginning on or after October 1, 1998, an osteopathic hospital's number of discharges for its cost reporting period that began during FY 1996 would have to be at least 3,000. We received no comments on the rural referral center criteria. C. Payments to Disproportionate Share Hospitals: Conforming Change Regarding Interpretation of Medicaid Patient Days Included in Disproportionate Patient Percentage (Sec. 412.106) Effective for discharges beginning on or after May 1, 1986, hospitals that treat a disproportionately large number of low-income patients receive additional payments through the disproportionate share (DSH) adjustment. One means of determining a hospital's DSH payment adjustment for a cost reporting period requires calculation of its disproportionate patient percentage for the period. The disproportionate patient percentage is the sum of a prescribed Medicare fraction and a Medicaid fraction for the hospital's fiscal period. Under clause (I) of section 1886(d)(5)(F)(vi) of the Act and Sec. 412.106(b)(2), the Medicare fraction is determined by dividing the number of the hospital's patient days for patients who were entitled (for such days) to benefits under both Medicare Part A and Supplemental Security Income (SSI) under Title XVI of the Act, by the total number of the hospital's patient days for the patients who were entitled to Medicare Part A. The Medicaid fraction is determined, in accordance with clause (II) of section 1886(d)(5)(F)(vi) of [[Page 40985]] the Act and Sec. 412.106(b)(4), by dividing the number of the hospital's patient days for patients who (for such days) were eligible for medical assistance under a State Medicaid plan approved under Title XIX of the Act but who were not entitled to Medicare Part A, by the total number of the hospital's patient days for that period. Initially, HCFA calculated the Medicaid fraction by interpreting section 1886(d)(5)(F)(vi)(II) of the Act to recognize as Medicaid patient days only those days for which the hospital received Medicaid payment for inpatient hospital services. See 51 Fed. Reg. 31454, 31460 (1986). The agency's interpretation was declared invalid by four Federal circuit courts of appeals. See Cabell Huntington Hosp., Inc. v. Shalala, 101 F.3d 984, 990-91 (4th Cir. 1996) (following three other circuits). These courts held that the statute requires, for purposes of calculating the Medicaid fraction, inclusion of each patient day of service for which a patient was eligible on that day for medical assistance under an approved State Medicaid plan. Specifically, the statute requires inclusion of each hospital patient day for a patient eligible for Medicaid on such day, regardless of whether particular items or services were covered or paid under the State Medicaid plan. On February 27, 1997, the HCFA Administrator issued HCFA Ruling 97- 2, which acquiesced in the four adverse appellate court decisions. The Ruling changed the agency's statutory construction to comport with those decisions, in order to facilitate nationwide uniformity in the calculation of the Medicaid fraction. Like the court decisions, the Ruling provides that a hospital's Medicaid patient days include each patient day of service for which a patient was eligible on such day for medical assistance under an approved State Medicaid plan, regardless of whether particular items or services were covered or paid under the State plan. The Ruling also reflects the hospital's burden of furnishing data adequate to prove each claimed Medicaid patient day, and of verifying with the State that a patient was eligible for Medicaid during each day of the inpatient hospital stay. The Ruling further provides that the agency's new interpretation is effective February 27, 1997 for each cost reporting period that: (1) Begins on or after that effective date; (2) was not settled, as of that date, on the Medicaid patient days issue, by means of an applicable notice of program reimbursement (NPR) (see Sec. 405.1803); or (3) was settled through such an NPR as of the Ruling's effective date and is the subject of a pending administrative appeal or civil action that satisfies all applicable jurisdictional requirements of the Medicare statute and regulations. The Ruling also provides, however, that the change in statutory interpretation effected by the Ruling is not a basis for reopening a hospital cost reporting period (see Secs. 405.1885-405.1889) that was finalized previously on the same matter at issue. We proposed to revise Sec. 412.106(b)(4) in order to conform the Medicare regulations to the new statutory construction issued in HCFA Ruling 97-2. These revisions are necessary to ensure that the regulations comport with the four appellate court decisions that declared invalid the agency's prior interpretation and led to the issuance of the HCFA Ruling. The proposed revisions would further facilitate nationwide uniformity in the calculation of the Medicaid fraction. Since the proposed revisions were intended simply to conform the regulations to HCFA Ruling 97-2 (and hence to the four adverse court decisions), revised Sec. 412.106(b)(4) would reiterate the Ruling's change of interpretation that the Medicaid fraction under section 1886(d)(5)(F)(vi)(II) of the Act includes each hospital patient day for a patient eligible for Medicaid on such day, regardless of whether particular items or services were covered or paid under the State Medicaid Plan. Our proposed revisions to Sec. 412.106(b)(4), like the Ruling, would continue to place on the hospital the burdens of production, proof, and verification as to each claimed Medicaid patient day. Under our proposal, revised Sec. 412.106(b)(4) would apply to cost reporting periods beginning on or after October 1, 1998. HCFA Ruling 97-2, which includes the same provisions as proposed Sec. 412.106(b)(4), would continue to apply to any cost reporting period beginning before October 1, 1998 provided that, as of February 27, 1997, there is for such period: no submitted cost report; no cost report settled on the Medicaid patient days issue through an applicable NPR; or a cost report settled on that issue, which is also the subject of a jurisdictionally proper administrative appeal or civil action on the issue. We received no comments in response to this proposal. Therefore, we are incorporating the proposed conforming change in this final rule. D. Payment for Bad Debts (Sec. 413.80) Section 4451 of the Balanced Budget Act of 1997 reduces the payment for enrollee bad debt for hospitals. Specifically, this provision reduces the amount of bad debts otherwise treated as allowable costs, attributable to the deductibles and coinsurance amounts under this title, by 25 percent for cost reporting periods beginning during fiscal year 1998, by 40 percent for cost reporting periods beginning during fiscal year 1999, and by 45 percent for cost reporting periods beginning during a subsequent fiscal year. We proposed to conform the regulations to the statute. Section 4451 of the Balanced Budget Act of 1997 also provides that in determining such reasonable costs for hospitals, any copayments reduced under the election available for hospital outpatient services under section 1833(t)(5)(B) of the Act will not be treated as a bad debt. This provision will be implemented in the outpatient prospective payment system regulation that implements sections 4521, 4522, and 4523 of the Balanced Budget Act of 1997, to be published later this year. We received one comment regarding the reduction in Medicare bad debt reimbursement which is discussed below. Comment: One commenter requested that the regulations and/or cost report forms (HCFA 2552-96) be modified to clarify that hospital-based skilled nursing facility bad debts will continue to be 100 percent reimbursable since freestanding skilled nursing facilities are not subject to the reduction in reimbursement and skilled nursing facilities are not mentioned in the law at section 1861(v)(1)(T). The commenter believed that in the BBA committee reports describing changes in reimbursement for Medicare bad debts, it seemed clear the changes were to apply to all providers, yet the law clearly stated that hospitals are the sole provider type subject to reductions in reimbursement. The commenter also noted that in reviewing the new hospital cost report forms, HCFA 2552-96, the commenter believed that the forms would apply the reduction in reimbursement to hospital-based skilled nursing facilities. Response: The HCFA 2552-96 hospital cost report forms do not apply the reduction in bad debt reimbursement to hospital-based skilled nursing facilities. Page 36-159, Line 26 and Page 36-164, Line 40 require entering the reduction for ``hospitals only''. Section 4451 of the BBA, and these implementing regulations, apply only to hospitals and any subprovider units settled through the hospital cost report, whether or not they have a separate provider number. Included in this are rehabilitation units, psychiatric [[Page 40986]] units, and childrens' hospitals, which are considered hospital providers. Cost reports for skilled nursing facilities, home health agencies, outpatient therapy, comprehensive outpatient rehabilitation facilities, community mental health centers, federally qualified health centers, and rural health clinics (after January 1, 1998) are separately settled and bad debts for these providers are not reduced. The bad debt reduction does not apply to ambulatory surgical centers because they are paid on another basis (fee schedule). End stage renal disease bad debts are computed separately and are not reduced. E. Payment for Direct Costs of Graduate Medical Education to Hospitals and Qualified Nonhospital Providers (Secs. 405.2468, 413.85, and 413.86) 1. Statutory Background Since its inception in 1965, Medicare has provided payment only to hospitals for the costs of graduate medical education (GME) training. The BBA allows for direct GME payment to qualified nonhospital providers to encourage training of future physicians in nonhospital settings. Under section 1886(k) of the Act, as added by section 4625 of the BBA, the Secretary is now authorized, but not required, to pay qualified nonhospital providers for the direct costs of GME training. The Conference Report also notes that the Conferees believe this authority may help alleviate physician shortages in underserved rural areas. We believe that providing Medicare payment directly to qualified nonhospital providers may facilitate more training and better quality training in nonhospital sites. Section 1886(k) of the Act states: ``For cost reporting periods beginning on or after October 1, 1997, the Secretary may establish rules for payment to qualified nonhospital providers for their direct costs of medical education, if those costs are incurred in the operation of an approved medical residency training programs described in subsection (h).'' The statute further provides that, to the extent the Secretary exercises this broad discretionary authority, the rules ``shall specify the amounts, form, and manner in which such payments will be made and the portion of such payments that will be made from each of the trust funds under this title.'' a. Payments only to ``qualified nonhospital providers''. The statute confers broad discretion on the Secretary regarding whether and how to pay qualified nonhospital providers for direct GME costs. However, the statute does specify the entities whom the Secretary can pay--``qualified nonhospital providers.'' Section 1886(k)(2) of the Act defines ``qualified nonhospital providers'' to include: Federally Qualified Health Centers (FQHCs), as defined in section 1861(aa)(4); Rural Health Centers (RHCs), as defined in section 1861(aa)(2); Medicare+Choice organizations; and such other providers (other than hospitals) as the Secretary determines to be appropriate. b. Payments only for the ``direct costs'' of training. The statute also specifies the costs the Secretary can pay for under section 1886(k) of the Act. Medicare pays hospitals for both the direct and indirect costs of medical education under sections 1886(h) and 1886(d)(5)(B) of the Act respectively, but section 1886(k) of the Act provides for payment to qualified nonhospital providers only for the direct costs of medical education. In addition, section 1886(k) of the Act provides for payment for the direct costs of training medical residents only if those costs are incurred in the operation of an ``approved medical residency training program.'' Accordingly, the statute authorizes Medicare payments to qualified nonhospital providers only for the costs of training medical residents, not for the costs of training other health professionals. In addition to adding section 1886(k) of the Act, section 4625 of the BBA amends section 1886(h)(3)(B) of the Act to prohibit double payments for direct GME to a hospital and a qualified nonhospital provider. This prohibition on double payments requires that the Secretary reduce a hospital's GME payments (the ``aggregate approved amount'' as defined in section 1886(h)(3)(b) of the Act) to the extent we pay a qualified nonhospital provider for GME under section 1886(k) of the Act. 2. Payment to Hospitals for GME Under sections 1886(d)(5)(B)(iv) and 1886(h)(4)(E) of the Act, a hospital may include the time a resident spends in nonprovider settings in its indirect medical education (IME) and direct GME full-time equivalent count if it incurs ``all or substantially all'' of the costs of training residents in the nonhospital site. Under Secs. 412.105(f) and 413.86(f)(1)(iii), a hospital may count resident training time in nonhospital sites for indirect and direct GME respectively if the resident is involved in patient care and there is a written agreement between the hospital and the nonhospital site that states that the resident's compensation for training time spent outside the hospital setting is to be paid by the hospital. 3. Proposed Policies Pursuant to section 4625 of the BBA, we proposed to provide Medicare payment to qualified nonhospital providers for the direct costs of GME training, effective for portions of cost reporting periods occurring on or after January 1, 1999. We proposed Medicare would make GME payments to the following ``qualified nonhospital providers''-- FQHCs, RHCs, and Medicare+Choice organizations. Under the authority of section 1886(k)(2)(D) of the Act, the Secretary may expand the definition of a ``qualified nonhospital provider'' to include such other providers (other than hospitals) as the Secretary determines to be appropriate. Once we have gained experience providing direct GME payments to FQHCs, RHCs, and Medicare+Choice organizations, we may consider including other types of nonhospital providers in the definition of a ``qualified nonhospital provider.'' Additionally, we proposed that, under certain circumstances, a hospital may continue to receive GME payments for residents who train in the nonhospital setting. In those instances where a hospital is eligible to continue receiving GME payments for residents who train in the nonhospital setting, the nonhospital site could receive payment from the hospital for costs they incur in training medical residents. Thus, our proposed policy would promote the intent of section 4625 of the BBA to provide financial support, either directly from Medicare or through the hospital, to qualified nonhospital providers for the direct costs of training residents in the nonhospital site. a. ``All or substantially all'' of the costs of training. Similar to our current policy of paying hospitals for training in nonhospital sites, we proposed that a qualified nonhospital provider may receive payment for the direct costs of GME if it incurs ``all or substantially all'' of the training costs. Although we proposed to pay the qualified nonhospital provider only when it incurred ``all or substantially all'' of the costs of training, we solicited comment on possible methods for allocating the GME payments for training in the nonhospital site where neither the hospital nor the qualified nonhospital provider is incurring ``all or substantially all'' of the costs of the training program. Under the proposed system, we would pay either the hospital or the qualified nonhospital provider for the cost of training in the nonhospital site, depending on which [[Page 40987]] entity incurs ``all or substantially all'' of the costs of training in the nonhospital site. We proposed to revise the definition of ``all or substantially all'' of the costs, which currently applies only to hospitals. Under the proposed redefinition, a hospital or qualified nonhospital provider would incur ``all or substantially all'' of the costs for the training program in the nonhospital setting if it pays for, at a minimum: that portion of the costs of the teaching physicians' salaries and fringe benefits that are related to the time spent in teaching and supervision of residents; and residents' salaries and fringe benefits (including travel and lodging expenses where applicable). b. Definition of ``direct costs'' of medical education for qualified nonhospital providers. Section 4625 of the BBA provides for payment to qualified nonhospital providers only for the direct costs of training residents. Our proposed definition of ``direct costs'' for qualified nonhospital providers is comparable to the direct costs for hospitals under section 1886(h) of the Act. Under our proposed policy, direct GME costs include costs incurred by the nonhospital site for the education and training of medical residents in approved programs. We proposed to include the following costs in the definition of direct costs: residents' salaries and fringe benefits (including related travel and lodging expenses where applicable); that portion of costs of the teaching physicians' salaries and fringe benefits that are related to the time spent in teaching and supervision of residents; and other related GME overhead costs. Consistent with our policies on direct GME costs for hospitals, we proposed direct GME costs for qualified nonhospital providers will not include normal operating costs or the marginal increase in costs that the nonhospital site experiences as a result of having an approved medical residency training program. For example, a decrease in productivity and increased intensity in treatment patterns as the result of a training program do not constitute ``direct costs'' of training residents in the nonhospital setting; rather, these are the ``indirect costs'' of such training. Also consistent with our policies for direct GME payments to hospitals, we proposed to pay qualified nonhospital providers only for training that is related to the delivery of patient care services. We also proposed that direct GME costs for qualified nonhospital providers, like direct GME costs for hospitals, would include only that portion of costs of the teaching physicians' salaries and fringe benefits associated with time spent in teaching and supervising residents. Specifically, a physician's time spent on teaching of a general nature would constitute a direct GME cost while activities spent in direct patient care which involve residents do not constitute direct costs. In addition, we proposed that direct costs in the qualified nonhospital provider would include that portion of teaching physicians' salaries and fringe benefits associated with time spent developing resident schedules and evaluating or rating the residents. Direct costs may also include the portion of a teaching physician's office costs allocated to GME. We stated that direct GME costs for qualified nonhospital providers would not include the following: a teaching physician's time spent in the care of individual patients which results in billable services; teaching physicians' activities that are related to the education of other health professionals (i.e., classroom instruction in connection with approved activities other than GME such as provider-operated nursing programs); teaching physicians' time spent on administrative and supervisory services to the qualified nonhospital provider that are unrelated to approved educational activities (i.e. operating costs); and teaching physician activities that involve nonallowable costs such as research and medical school activities that are not related to patient care in the nonhospital setting. Costs associated with the providing teaching services to undergraduate medical students are also not include in direct graduate medical education costs. GME overhead costs include only those costs that are allocable to direct GME and that are not used in patient care. For example, a portion of administrative and general costs could be appropriately allocated to an RHC's or FQHC's GME cost center. Similarly, a conference room that is dedicated specifically for the training of residents could be appropriately allocated to an RHC or FQHC's GME cost center. By contrast, patient care rooms added to an RHC or an FQHC cannot be appropriately allocated to an RHC's or FQHC's GME cost center. One of the advantages of the proposed definition of ``direct costs'' is that it is administratively feasible. Our definition of ``direct costs'' for qualified nonhospital providers is comparable to the direct costs that are included in the per resident amount paid to hospitals under section 1886(h) of the Act. At present, there is limited information regarding the actual costs of training residents in nonhospital sites. After we gain experience providing direct GME payments to qualified nonhospital providers and have reviewed the GME costs separately reported by these qualified nonhospital providers, we may revise the definition of ``direct costs.'' We solicited comments on other elements that may constitute direct costs of GME in the qualified nonhospital provider that can be identified, reported, and verified as directly attributable to GME activities through the cost reporting process. We were interested in comments on whether we should include other costs in the definition of ``direct costs'' for qualified nonhospital providers and on the administrative feasibility of identifying the GME portion of those costs. c. Determining direct costs. One of our major concerns in developing policies for paying qualified nonhospital providers for the direct costs of GME is the administrative feasibility of determining the amount of direct costs incurred by the qualified nonhospital provider. It is our understanding that, currently, hospitals and nonhospital sites often share, to varying degrees, the costs of training residents in the nonhospital site. Because of the difficulty in apportioning costs between the hospital and the nonhospital for the training in the nonhospital site, we believe that it is not administratively feasible to pay both the hospital and the nonhospital site for the cost of training in the nonhospital site. We have been unable to devise a method for accurately apportioning costs between the two entities. Furthermore, the potential for both the hospital and the qualified nonhospital provider to be paid for the same direct GME expenses poses a significant problem for complying with section 1886(h)(3)(B) of the Act, as amended by the BBA, which specifically prohibits double payments. Under this provision, the Secretary shall reduce the hospital's GME payment (the ``aggregate approved amount'') to the extent we pay the qualified nonhospital provider for GME costs under section 1886(k) of the Act. Consequently, our policy must ensure that Medicare does not pay two entities for the same training time in the nonhospital site. Given that the hospital's per resident amount can include, but is not necessarily based on the costs of training in the nonhospital site, we were not able to devise an equitable way of reducing the hospital's per resident payment to reflect payments made [[Page 40988]] under section 1886(k) of the Act. It may not be equitable to subtract the exact amount of payment made to the qualified nonhospital provider from the hospital's per resident payment because the payment made to the nonhospital site may be unrelated to the hospital's per resident amount. We believe that the residents' salaries, teaching physicians' salaries, and overhead costs for the nonhospital setting will constitute a different proportion of the total GME costs in the nonhospital setting as compared with the hospital setting. Rather, it may be more equitable to determine the proportion of costs incurred by each entity and reduce the hospital's per resident payment by the proportion of GME costs incurred by the nonhospital site; however, since specific components of the per resident amount were not identified in the hospital's GME base year (1984), we cannot accurately determine the appropriate amount to reduce the current year hospital per resident payment amount. Moreover, to reduce the hospital's GME payments based solely on the amount paid to the qualified nonhospital provider could result in inequitable payments to the hospital, which has ongoing costs even when the resident is training in the nonhospital site. In fact, it could leave the hospital at risk of receiving no payment for the GME costs it has incurred. In order to encourage training in nonhospital sites, it is important to develop a policy that, while providing payment to qualified nonhospital providers, would also be equitable to hospitals. We believe that paying only the qualified nonhospital provider for the training costs could result in hospitals choosing not to rotate their residents to the nonhospital site. We have been unable to devise an equitable and accurate method for dividing the GME payment for training in the nonhospital site if neither the hospital, nor the nonhospital site incurs ``all or substantially all'' of the costs. As such, we solicited comment on possible methods for allocating the GME payments for training in the nonhospital site where neither the hospital nor the qualified nonhospital provider agrees who is incurring ``all or substantially all'' of the costs for the training program. We believe that the policies discussed below are equitable to both hospital and qualified nonhospital providers and will achieve Congress' objective of encouraging and supporting training in the nonhospital setting. Given our concerns about administrative feasibility, the statutory prohibition on double payments, and developing policies that are equitable to hospitals as well as qualified nonhospital providers, we believe the only feasible way to pay for training in nonhospital settings is to pay either the hospital or the nonhospital provider. Currently, hospitals may receive payment for the time residents spend in the nonhospital setting if the hospital incurs ``all or substantially all'' of the training costs. We proposed to adopt a similar policy for qualified nonhospital providers; that is, a qualified nonhospital provider may receive payment for the direct costs of GME if it incurs ``all or substantially all'' of the training costs. d. Payment to FQHC's and RHC's. We proposed to pay FQHC's or RHC's for direct GME costs based on reasonable costs if the FQHC or RHC incurs ``all or substantially all'' of the costs of training the resident in the nonhospital setting. The FQHC or RHC would have to report direct GME costs in a reimbursable cost center on its cost report under the proposal. Conversely, where an FQHC or RHC did not incur ``all or substantially all'' of the costs of training residents in the nonhospital site, the FQHC or RHC would report direct GME costs in a nonreimbursable cost center on the cost report. We proposed that the FQHC's and RHC's allowable direct GME costs be subject to reasonable cost principles in 42 CFR part 413 and other relevant provisions referenced in part 413. In addition, the FQHC's and RHC's direct GME costs would be subject to the Reasonable Compensation Equivalency limits under Secs. 415.60 and 415.70. Also, Medicare would pay only for its share of the direct costs of training in the qualified nonhospital provider. We proposed that the FQHC's and RHC's Medicare share equal the qualified nonhospital provider's ratio of Medicare visits to total visits. Thus, the amount of Medicare payment would equal the product of the clinic's Medicare allowed reasonable direct GME costs and the clinic's ratio of Medicare visits to total visits. For FQHC's and RHC's that incur ``all or substantially all'' of the costs for the training program in the nonhospital setting, we proposed that the direct GME costs would not be subject to the existing per visit payment caps for reimbursement under sections 505.1 and 505.2 of the Medicare Rural Health Clinic and Federally Qualified Health Centers Manual. We also proposed that, where payment is available under section 1886(k) of the Act for residents working in either an FQHC or an RHC, the FQHC's and RHC's do not need to include residents as health care staff in the calculation of productivity standards under section 503 of the Manual. e. Payment to Medicare+Choice organizations. We proposed making direct GME payment to Medicare+Choice organizations which incur ``all or substantially all'' of the costs of training in a nonhospital site. The Medicare+Choice organization would be eligible to receive payment on a reasonable costs basis for residents' salaries and fringe benefits only for the time that the resident spends in the nonhospital setting. In addition, we proposed that the Medicare+Choice organization's allowed costs include only that portion of the teaching physician salaries and fringe benefits that is related to training in the nonhospital setting. We proposed limiting payment to Medicare+Choice organizations to residents' salaries and fringe benefits and supervisory teaching physician compensation which can be allocated to direct GME. We did not propose to pay Medicare+Choice organizations for the costs of overhead that may be associated with a GME program. We solicited suggestions for creating a methodology for allocating and reporting overhead costs for Medicare+Choice organizations and suggestions for mechanisms for the audit and review of the costs for Medicare+Choice organizations. Similar to our proposed policy for paying FQHCs and RHCs for direct costs of GME, we proposed that the Medicare+Choice organization's reimbursement for residents' salaries and fringe benefits (including related travel and lodging expenses where applicable) would be subject to the reasonable cost principles in 42 CFR part 413 and any other relevant provisions referenced in part 413. In addition, we proposed the Medicare+Choice organization's GME reimbursement would also be subject to the Reasonable Compensation Equivalency limits under Secs. 415.60 and 415.70. We proposed to allow the Medicare+Choice organization to receive direct GME payment only for the direct costs of training in the nonhospital site that are associated with the delivery of patient care services. In determining the amount of direct GME payments to Medicare+Choice organizations, we proposed adjusting for Medicare's share of those education costs. Medicare's share would equal the ratio of the total number of Medicare enrollees in the Medicare+Choice organization to total enrollees in the Medicare+Choice organization. We proposed that, in order to receive the direct GME payment, the [[Page 40989]] Medicare+Choice organization must produce a contractual agreement between itself and the nonhospital sites. Medicare+Choice organizations may contract with any nonhospital patient care site, including freestanding clinics, nursing homes, and physicians' offices in connection with approved programs. The contract between the Medicare+Choice organization and the nonhospital site must indicate that, for the time that residents spend in the nonhospital site, the Medicare+Choice organization agrees to pay for the cost of residents' salaries and fringe benefits. In addition, the contract must indicate that the Medicare+Choice organization agrees to pay the portion of the costs of teaching physicians' salaries and fringe benefits that is related to the time spent in teaching and supervision of residents and is unrelated to the volume of services provided by the physician. The contract must stipulate the portion of each teaching physician's time that will be spent training residents in the nonhospital setting. Moreover, the contract must indicate that the Medicare+Choice organization agrees to identify an amount for the cost of the teaching physician's salary based on the time that the resident spends in the nonhospital setting, not based upon a capitated rate for the delivery of physician services. f. Payment to hospitals. A hospital may include a resident's training time in a nonhospital setting in its FTE counts for direct GME and for IME if the hospital incurs ``all or substantially all'' of the costs for training in the nonhospital setting. We proposed that, in order for a hospital to include residents' training time in a nonhospital setting, the hospital and the nonhospital site must have a written contract which indicates the hospital is assuming financial responsibility for, at a minimum, the cost of residents' salaries and fringe benefits (including travel and lodging expenses where applicable) and the costs for that portion of teaching physicians' salaries and fringe benefits related to the time spent in teaching and supervision of residents. The contract must indicate that the hospital is assuming financial responsibility for these costs directly or that the hospital agrees to reimburse the nonhospital site for such costs. The contract must also contain an acknowledgment on the part of the qualified nonhospital provider if the nonhospital site is an FQHC or RHC that, since the residents' time is being counted by the hospital, the nonhospital site must report GME costs on the Medicare cost report in a nonreimbursable GME costs center. In addition, in order to determine teaching physician compensation that may be allocated to direct GME, the FQHC and RHC will have to specify the portion of the teaching physicians' time that will be spent training residents in the nonhospital setting. Under Sec. 413.86(f)(1)(iii), hospitals may contract with any nonhospital patient care site such as freestanding clinics, nursing homes, and physicians' offices in connection with approved programs. Payment to the hospital for the direct costs of GME training in the nonhospital setting will continue to reflect Medicare's share, which equals the hospital's ratio of Medicare inpatient days to total inpatient days. 5. Trust Funds Under section 1886(k)(1) of the Act, the rules established by the Secretary for paying qualified nonhospital providers for GME must specify the portion of Medicare payments that will be made from each of the Medicare trust funds. We proposed that GME payments made directly to an FQHC, RHC, or Medicare+Choice organization would be made from the Federal Supplementary Medical Insurance Trust Fund. 6. Proposed Effective Dates We proposed that the effective date of these provisions for FQHCs, RHCs, Medicare+Choice organizations, and hospitals would be January 1, 1999. Of the provisions affecting hospitals, the policies for IME payments would apply to discharges occurring on or after January 1, 1999. The policies concerning medical education payments to FQHCs, RHCs, and hospitals would apply to portions of cost reporting periods occurring on or after January 1, 1999. We proposed that Medicare+Choice organizations could begin receiving payments for direct GME costs incurred on or after January 1, 1999. 7. Responses to Comments Received on Proposed Policies and Final Rule Provisions Below we are summarize the comments we received on the proposed policies and provide our responses to those comments. a. Definition of qualified nonhospital provider. Comment: One commenter stated that HCFA should expand the definition of a qualified nonhospital provider to include preventive medicine residencies. This commenter quoted the Conference Report statement: The Conferees also note that preventive medicine residency training occurs most often in nonhospital settings and the Conferees encourage the Secretary to examine carefully the opportunities to provide support to such training programs. The commenter further noted that a small number of residency programs would benefit if we adopted the suggestion. Response: Consistent with the direction of the Conference Report, we have examined how to encourage preventive medicine training through the Medicare program. We understand that preventive medicine training consists of one year of clinical training, one year of academic study, and a practicum year. To the extent that the one year of clinical training is provided in patient care sites that qualify to receive medical education payments, Medicare provides payment for training much in the same way we provide payment for all other specialty programs. A hospital can count a preventive medicine resident who receives training in all areas of the hospital complex. The hospital may also count a preventive medicine resident who receives training in a nonhospital site if the resident is involved in direct patient care and there is a written agreement between the hospital and the nonhospital site that the hospital is incurring ``all or substantially all'' of the costs of training the resident in the nonhospital site. FQHCs, RHCs, and Medicare+Choice organizations can receive payment on a reasonable cost basis for costs associated with training preventive medicine residents if the entity incurs ``all or substantially all'' of the costs. Since the year of academic study does not involve direct patient care, a hospital or qualified nonhospital provider cannot receive Medicare payment for that year of preventive medicine training. A fundamental principle of Medicare payment for education is that the residents must participate in patient care services to patients at the health care site. Although we believe that preventive medicine residents are engaging in activities that will benefit all patients, not just Medicare patients in general, the year of academic study does not constitute patient care services which would qualify for Medicare payment for GME. We understand the clinical training that preventive medicine residents receive may also occur in patient care sites that do not receive payments from Medicare, such as public health clinics. Even if the clinics were included under the definition of qualified nonhospital provider, Medicare payment to clinics for GME would likely still be very low because it would reflect the share of services provided by the clinic to [[Page 40990]] Medicare beneficiaries as compared to all services it provides. We do not believe that Medicare beneficiaries make significant use of public health clinics for Medicare covered services since these services are also available through their regular doctor. If we were to provide payments to public health clinics associated with the training of preventive medicine residents, we would also have to resolve technical problems related to providing payments to entities that have never had a relationship with Medicare. As we stated above, where a hospital or qualified nonhospital provider incurs ``all or substantially all'' of the costs of the clinical training in that nonhospital site, Medicare will make payments for GME costs associated with training preventive medicine residents. Comment: One commenter urged HCFA to consider including nonhospital dental clinics in the definition of qualified nonhospital providers. One commenter urged us to expand the definition of a qualified nonhospital provider to make payment of both direct and indirect GME directly to nursing homes and hospices. One commenter requested clarification as to whether our definition of a qualified nonhospital provider includes community mental health centers. If not, the commenter requested that we consider including community mental health centers in the definition of qualified nonhospital provider. Response: As we stated in the proposed rule, we believe that it is appropriate to have more experience with providing payments to the qualified nonhospital providers listed in the statute before we expand the definition to include other sites such as those stated by these commenters. We note that even if nonhospital dental clinics were included in the definition of a qualified nonhospital provider, a dental clinic's low Medicare share means the benefit of the provision would be small. Dental clinics are likely to have a low Medicare share because Medicare covers few dental services. Currently, our definition of qualified nonhospital provider does not include community mental health centers per se, but it may be possible for a community mental health center to meet the criteria for being designated as a rural health clinic under section 1861(aa)(2) of the Act and section 405.2402. We would note that a hospital or Medicare+Choice organization may receive payment associated with resident rotations through the nonhospital sites suggested by these commenters if the hospital or Medicare+Choice organization incurs ``all or substantially all'' of the costs at the clinic. In this way the clinic will be paid by the hospital for GME costs. Comment: One commenter argued that Congress specified that a qualified nonhospital provider includes FQHC's, RHC's, and managed care plans to ensure that these organizations were included but that Congress did not intend to limit qualified nonhospital providers to these organizations. The commenter believed that excluding other nonhospital sites from the definition of a qualified nonhospital provider is contrary to Congress' intent. Response: As we have stated, we will consider other nonhospital sites in the definition of qualified nonhospital providers once we have experience with these policies. We disagree that the proposal to limit the definition of a qualified nonhospital provider at this time to the entities listed in the statute is inconsistent with Congressional intent. The statute defines qualified nonhospital provider to include ``such other providers (other than hospitals) as the Secretary determines to be appropriate.'' Thus, the statute authorizes but does not require the inclusion of other entities. Comment: One commenter stated that educational consortia are becoming important models for community-based graduate medical and nursing training and suggested that we expand the definition of qualified nonhospital provider to include consortia. Response: We are interested in learning more about the development of GME programs through educational consortia. Section 4628 of the BBA requires the Secretary to establish a demonstration project under which GME payments will be made to consortia. We will consider changes to our GME payment policies based on our evaluation of any future demonstration projects. Comment: One commenter urged us to expand the definition of a qualified nonhospital provider to include Osteopathic Postdoctoral Training Institutions (OPTIs), community based health care consortia consisting of one or more colleges accredited by the American Osteopathic Association (AOA), one or more AOA accredited hospitals, and other health care facilities such as nursing homes, ambulatory clinics, community health centers, and managed care organizations. The commenter suggested that payments be made directly to the OPTI based on the number of residents participating in OPTI hospitals or a national average payment. The commenter stated that the OPTI would distribute the payments among the consortia members. Response: An OPTI includes hospital and nonhospital sites as well as educational institutions and we believe an OPTI is a consortium. As we stated above, we will be studying GME payments to consortia in a demonstration project required by section 4628 of BBA. b. Definition of direct costs. Comment: One commenter suggested that direct costs of training in nonhospital sites should include mileage associated with travel between multiple clinic sites. The commenter also stated that direct costs should include the costs of telemedicine, including telephone, fax, videoconference, and the internet because these electronic communication mechanisms enable primary care residents in nonhospital sites to be trained for practice outside of the resource-rich, multispecialty hospital setting. Response: We agree that travel costs may be an element of direct costs when residents work in multiple nonhospital sites or when residents travel from a hospital training site to remote clinics. We disagree that the cost associated with telecommunication services should be allowable as training costs. Although telecommunication services may be integral to providing services to patients while residents are training in nonhospital sites, these services are not principally designed to be used as GME training tools. Rather, the telecommunication services to which the commenter is referring, like the use of a stethoscope or an examining room, are compensated as operating costs through Medicare's payments for patient care services. Comment: Several commenters stated that the effect of training on indirect costs is similar in nonhospital clinics and hospitals. One commenter suggested that indirect costs are easily identifiable and should be separately reimbursable in nonhospital settings. Response: The statute states that the ``Secretary may establish rules for payment to qualified nonhospital providers for the direct costs of medical education if those costs are incurred in operation of an approved medical residency training program described in subsection (h).'' The statute clearly limits payment to qualified nonhospital providers under section 1886(k) of the Act for the direct costs of GME. Comment: One commenter stated that the proposed regulations fail to reflect that FQHCs are eligible for Part B payments for allowable teaching costs even without the new methodology [[Page 40991]] established pursuant to the new BBA provision. Because FQHCs are governed by cost reimbursement principles that include teaching costs, FQHCs are already allowed to claim all training-related costs, including direct faculty and resident costs. This commenter suggested that FQHCs that participate in teaching programs should be able to recapture higher operating costs caused by lower productivity and increased overhead. According to this commenter, we should consider including the following in direct costs: --Slowdown in productivity; --Facilities and space for training; --Transportation and living costs for residents; --Availability of lab and radiology equipment and services; --Administrative overhead; --Increased intensity in treatment patterns used in training; --Equipment costs; --Library (either onsite or electronic access); --Capital costs for startup of residency program; --Increased complexity at teaching FQHCs; and --Increased social complexity of patient case mix. Response: The costs of resident salaries and fringe benefits and supervising physicians may be allowable costs under Sec. 405.2470. If the RHC or FQHC were to have a written agreement with a hospital where the hospital provides compensation for these costs to the clinic, these costs would become nonreimbursable costs. However, FQHCs and RHCs that have an all-inclusive rate that exceeds the cap under sections 505.1 and 505.2 of the Medicare Rural Health Clinic and Federally Qualified Health Centers Manual would still benefit from the proposed policy in that costs above the cap that would otherwise be nonreimbursable by Medicare can now be compensated as direct GME costs through the agreement with the hospital. That is, if the FQHC or RHC incurs ``all or substantially all'' of the costs and receives payment directly from Medicare, these costs are GME costs that are treated separately in applying the caps on the all-inclusive rate under sections 505.1 and 505.2 of the Medicare Rural Health Clinic and Federally Qualified Health Centers Manual. An additional benefit in the situation where we pay the FQHC or RHC directly for GME is that residents do not need to be included as health care staff in the calculation of productivity standards under section 503 of the Manual. We further believe that residents should be excluded from productivity standards in situations where the hospital is being paid for training time and GME costs are not reimbursable costs for the FQHC or RHC. We are adopting this policy in this final rule and will modify section 503 of the Manual accordingly. Among the items listed in this comment, we believe that costs which are directly related to the operation of a medical residency training program (facilities and space exclusively dedicated to training, resident travel costs between remote clinic sites) in addition to facility overhead which can be allocated to a medical education cost center constitute allowable direct GME costs for which the FQHC or RHC can receive payment directly from Medicare. We believe the remaining items listed are either indirect costs of training or allowable cost for patient care services under Sec. 405.2468(a) through (e) which can only be reimbursed as non-GME operating costs. Comment: One commenter was opposed to the application of reasonable compensation equivalents to physicians in FQHCs and RHCs. The commenter stated that the BBA required HCFA to subject RHCs to productivity standards and the per-visit cost limit. According to the commenter, if Congress had intended for the RCE limits to be imposed on RHCs, the BBA would have required such a policy. The commenter stated that, by definition, RHCs and FQHCs are located in areas where it is difficult to attract physicians and that the providers must pay compensation that exceeds the RCE limits to attract qualified physicians. The commenter requested that the limits not be imposed on FQHC and RHC services to individual patients. Response: For purposes of making indirect GME payments to FQHCs and RHCs, the RCE limits will only apply to the portion of a teaching physician's compensation that is attributable to direct GME. We are not applying the RCE limit to physician compensation that is related to providing services to individual patients. Because we intend to pay for these GME costs on a reasonable cost basis, it is necessary to apply the RCE limits to assure that GME costs will be reasonable. Comment: One commenter stated that if HCFA intends to compute the fixed cost for nonhospital training of all health professionals from the cost reimbursement data received over the next few years from qualified nonhospital providers, costs associated with training of nonphysician health practitioners should also be reported. This commenter stated that it will be difficult to collect these data at a later date. Response: FQHCs and RHCs seeking payment from Medicare for direct GME must appropriately classify those costs to a GME cost center on the cost report. These payments are limited to the direct costs the FQHC or RHC incurs for an approved medical residency training program as described under section 1886(h) of the Act. Training of non-physician health professionals are not included in these programs. Therefore, in submitting costs reports, FQHCs and RHCs must clearly distinguish the costs of training residents from the cost of training other health professionals in nonhospital sites. Although FQHCs and RHCs will need to document costs of approved medical residency programs to be allocated to the GME cost center, we do not believe the information benefit associated with obtaining data on training of other health practitioners would justify imposing an additional administrative burden on FQHCs and RHCs to report costs for which they will receive no payment. c. Revised definition of ``all or substantially all'' of the costs. Comment: A number of commenters felt the proposed redefinition of ``all or substantially all'' of the costs will be counterproductive and result in less training in nonhospital settings. One commenter stated that the current standard of ``or substantially all'' has helped to facilitate resident training in nonhospital sites. This commenter stated that there is strong anecdotal evidence that resident training in ambulatory sites has been increasing and recommended that any changes to existing policies be tested for the likelihood that they promote expanded ambulatory GME. Response: We disagree with the commenters who suggested that the proposed redefinition of ``all or substantially all'' of the costs of training residents in the nonhospital sites will result in less training in nonhospital settings. First, we do not believe that hospitals themselves will be discouraged from continuing to rotate residents to nonhospital sites. Hospitals must consider accreditation and other program requirements in addition to purely financial considerations. We have reviewed the program requirements for residency education in family practice and internal medicine in the 1997-1998 GME Directory. The Directory specifies that family practice residents must spend specified amounts of time and see a minimum number of patients in the family practice center in each residency program year. Similarly, the Directory specifies that at least 25 [[Page 40992]] percent of the 3 year residency program for internal medicine must be in an ambulatory care setting. Given these requirements for primary care training programs, we do not believe that hospitals will respond to the revised definition of ``all or substantially all'' of the costs by rotating fewer residents to nonhospital sites. Moreover, a hospital that meets the ``all or substantially all'' criterion may count the resident's training time in the nonhospital site for direct GME as well as IME. Second, we believe that our proposal will encourage more ambulatory sites to participate in training. To the extent our policies would allow qualified nonhospital providers to receive payments directly from Medicare, more qualified nonhospital providers may be willing to become training sites. In addition, the hospital may incur supervisory teaching physician costs that previously might have been borne by the nonhospital site. Therefore, the nonhospital site either will receive revenues for costs that the site itself incurs or will no longer incur those costs. Comment: Several commenters agreed that it is appropriate to provide GME payment to the entity that incurs ``all or substantially all'' of the costs whether it be the hospital or the qualified nonhospital provider. Many of these commenters, however, believe that ``all or substantially all'' of the costs should be limited to resident salaries and fringe benefits. Response: We disagree. Section 1886(h)(4)(E) of the Act states that hospitals may include residents in their FTE counts for direct GME if the hospital incurs ``all or substantially all of the costs of the training program in that setting.'' Section 1886(d)(5)(B)(iv) of the Act allows hospitals to count residents for IME effective October 1, 1997 if the hospital ``incurs all or substantially all of the costs for the training program in that setting.'' As we stated previously and in the preamble to the proposed rule (63 FR 25597), we reviewed data on resident costs from recent Medicare hospital cost reports and found that, on average, resident salaries and fringe benefits account for less than half of total direct GME costs. We believe that the revised policy, which requires hospitals to incur a higher percentage of total training costs in the nonhospital setting than are accounted for by resident compensation reflect a better measure of ``all or substantially all'' of the costs than current policy. Comment: One commenter argued that the rationale for the proposal is insufficient to merit a change in current policy. This commenter stated that our proposal focused only cost data from hospitals and not nonhospital sites. This commenter believed that, because our proposal addressed training in nonhospital sites, it would be more appropriate to analyze resident salaries and fringe benefits as a share of overall training costs at nonhospital sites. The commenter acknowledged that these data are not available at the present time, but believed that resident compensation is likely to be a substantial component of overall training costs in nonhospital sites. The commenter noted that the preamble to the proposed rule indicates that residents' salaries and supervisory costs would likely ``constitute a different proportion of the total GME costs in the nonhospital setting as compared with the hospital setting.'' (63 FR 25597). The commenter added that direct GME payments to hospitals are based on 1984 hospital costs that may not accurately reflect current costs. Response: Our analysis is based on recent cost report data submitted to us by hospitals. That data shows that resident salaries and fringe benefits are less than half of total resident costs for hospitals. At this time, based on available data as well as a desire to treat hospitals and nonhospital sites equitably, we believe the hospital cost report data is a useful proxy for purposes of applying a standard of ``all or substantially all'' to nonhospital sites. We agree that it would be appropriate to analyze data on the cost of training from nonhospital sites and we will consider revisions to our policies as we obtain cost data from nonhospital sites. We note that, if resident compensation is, in fact, a larger percentage of total costs in the nonhospital site relative to the hospital, as suggested by this commenter, this would mean that costs other than resident compensation are a smaller proportion of total costs. The hospital would have to assume relatively modest additional costs through arrangements with nonhospital sites to continue counting the residents for indirect and direct GME. We also note that preliminary data by researchers studying costs incurred by a nonhospital site to train residents has shown that resident salary and fringe benefits are a smaller ratio of total costs at the nonhospital site relative to the hospital. If this conclusion is accurate, it would provide additional evidence that our revised definition is a better measure of ``all or substantially all'' of the costs. Comment: One commenter acknowledged that we revised the definition of ``all or substantially all'' to address a concern that nonhospital sites do not have sufficient resources to support their medical education activities, but argued that the proposed change in policy will not improve the ability of nonhospital sites to support training and may compromise existing and developing relationships between hospital and nonhospital GME sites. This commenter stated that the relationship between the hospital and nonhospital site should be voluntary and that it is up to the parties to define the appropriate parameters of their relationships, including how costs beyond the resident stipend and benefits should be accommodated. Response: As we stated earlier, we do not believe that this revised policy will compromise existing training relationships between hospitals and nonhospital sites. We agree with the commenter that arrangements between hospitals and nonhospital sites for training should be voluntary and the entities should be responsible for negotiating the parameters of their relationship. If a hospital and nonhospital site cannot agree on an arrangement regarding costs, the hospital may pursue an agreement with another nonhospital site for training. Similarly, if a nonhospital site cannot reach agreement with a hospital, it does not have to allow its facility to be used as a training site and can pursue a training arrangement with another hospital. Comment: One commenter asked why a nonhospital site would claim costs, and report an offset to those costs, if the hospital incurs the GME costs for training in the nonhospital site. Response: In response to this comment, in this final rule we are modifying the requirements for both hospitals and qualified nonhospital providers. As stated previously, hospitals are required to furnish a written agreement between the hospital and the nonhospital site that indicates that the hospital is incurring the cost of the resident's compensation in the nonhospital site and that the hospital is providing reasonable compensation for teaching activities to the nonhospital site. The agreement must also indicate the amounts being furnished to the nonhospital site for teaching activities. If the resident is working at an FQHC or RHC and there is a written agreement that allows the hospital to count the resident for indirect and direct GME, the FQHC or RHC must report its direct GME costs in a nonreimbursable cost center. The FQHC or RHC is not required to offset from those GME costs revenues received from the hospital. [[Page 40993]] We are requiring the FQHC or RHC to report its direct GME costs in a nonreimbursable cost center because these costs will no longer be allowable costs under Sec. 405.2468(a) through (e). As stated earlier, direct GME costs will not be subject to the cap on the all-inclusive rate under section 503 of the RHC and FQHC Manual. The reporting of direct GME costs in a separate cost center on the FQHC and RHC cost report will also allow us to receive data on the costs of training in nonhospital sites. Comment: Some commenters argued that our proposal would impose undue administrative burden on hospitals and nonhospital sites by requiring them to report all of the GME costs they incur. One commenter stated that HCFA should retain the current definition of ``all or substantially all'' of the costs because it is logical, straightforward, and appropriate. This commenter asserted that it is difficult to isolate and quantify costs other than resident salaries and fringe benefits are incurred in nonhospital sites. According to this commenter, resident salaries and fringe benefits are easy to identify and their administration and recordkeeping can be monitored uniformly across the GME community. The commenter suggested that in assuming responsibility for resident compensation, the teaching hospital assumes responsibility for assuring that all residents are provided appropriate educational environments, supervision, and support for their training. Another commenter argued that the proposed redefinition of ``all or substantially all'' of the costs does not reflect certain services or costs (e.g. house staff credentialing and related functions) just as the per resident amounts do not reflect services or costs that are included in the proposal (e.g. resident travel and lodging). These commenters suggested that resident salaries and fringe benefits should suffice as a proxy that appropriate educational services at an appropriate cost are being delivered by the hospital for the nonhospital training. Another commenter stated that it is a managed care organization that pays the resident salaries and fringe benefits and that this should be sufficient for receiving GME payment in the nonhospital site. According to these commenters, the entity that incurs the costs of the resident compensation should be considered to be incurring ``all or substantially all'' of the costs and be eligible to count the resident for direct and indirect GME. Response: We do not believe that we are establishing a burdensome regulatory structure with tremendous documentation requirements. For hospitals seeking to count the time of residents training in the nonhospital site, we are requiring a written agreement between the hospital and the nonhospital site stating that the hospital will incur ``all or substantially all'' of the costs. The written agreement must indicate that the hospital is incurring the cost of the resident salaries and providing compensation for supervisory teaching physician costs. The agreement must also specify the amounts paid to the nonhospital site. These agreements and amounts paid by the hospital to the nonhospital site may be the product of negotiation between the hospital and nonhospital site. The hospital does not have to report the nonhospital site's GME costs. We anticipate that in the course of any negotiation between the hospital and nonhospital site, the nonhospital site may need to identify its training costs. However, this is a matter between the hospital and nonhospital. If a hospital seeks to count the time of residents training in FQHC's and RHC's, the FQHC or RHC must identify its training costs in a nonreimbursable GME cost center. FQHC's and RHC's must separately report GME costs in order to distinguish these costs from other patient care costs that are paid for by Medicare on the basis of reasonable costs through the all inclusive rate. Under this final rule, we are not requiring FQHC's and RHC's to report the offset to those costs for payments received from the hospital. Requiring FQHC's and RHC's to report costs without offsetting revenues received from the hospital will allow us to obtain gross cost data on the costs of training in nonhospital sites. RHC's and FQHC's must identify teaching physician costs and allocate overhead to the direct GME cost center, in addition to the current cost reporting requirements for these entities. These entities are currently paid on the basis of costs, and we do not believe the additional cost reporting requirements will be substantial. We disagree with the comment that resident compensation should suffice as a proxy that appropriate educational services, at an appropriate cost, are being delivered and should be the sole criterion for determining which entity receives payment. Our concern in developing this policy is not whether we are paying for appropriate educational services but whether the entities that incur training costs are appropriately paid. Regardless of which entity incurs the cost of the resident's compensation, Medicare should only pay for appropriate educational services. Other regulations independent of the ``all or substantially all'' criterion ensure that Medicare pays for accredited educational programs. Comment: One commenter stated that teaching physicians in nonhospital sites may be remunerated through a variety of different arrangements, including ``in kind'' compensation for continuing education or through voluntary contributions. According to this commenter, the proposed policies would require hospitals and nonhospital sites to identify financial transactions which may not exist. The commenter further stated that there is no established methodology for defining or quantifying supervisory costs. The commenter noted that even if the costs could be identified, the costs would vary depending upon specialty and the year of residency training, which would require a sophisticated accounting infrastructure. The commenter also asserted that community-based physicians would be discouraged from training residents because of the administrative burden of documenting the precise number of hours they spend teaching or supervising residents. Response: We recognize that there could be a variety of financial arrangements between hospitals and nonhospital sites with regard to training. The hospital and the nonhospital site can take into account those types of arrangements in negotiating an agreement. Although there will be some additional cost reporting requirements imposed on FQHC's and RHC's that receive payment for direct GME through the hospital or directly from Medicare, there are established cost reporting principles for identifying these costs in providers. Medicare+Choice organizations, in addition to producing a written agreement with nonhospital sites, will have to report GME costs when they incur ``all or substantially all'' of the costs. We are developing a modest one page cost statement that will allow the Medicare+Choice organizations to claim direct GME costs that are eligible for payment. If an FQHC or RHC incurs ``all or substantially all'' of the costs of the program, and is therefore eligible to be paid directly for GME, we do not believe the burden of documenting supervisory physician time spent in GME activities will be substantial. Our expectation is that physicians will need to estimate the number of hours they will spend in GME and non-GME activities during the course of the year and verify the estimates with a limited time study. This is similar to the documentation that was required of hospitals to allocate teaching physician costs between Part A [[Page 40994]] and Part B and between operating costs and direct medical education. Comment: Several commenters suggested that we initiate demonstration projects addressing payment for GME in nonhospital sites. One commenter suggested that we analyze our proposed revision to ``all or substantially all'' of the costs through a demonstration project before implementing the changes on a nationwide basis. Such a demonstration project would indicate whether the proposed change would encourage or discourage training in nonhospital sites. Another commenter suggested that our proposed policy may adversely affect many GME programs and should be tested prior to being implemented on a national basis. Response: Congress established a provision in the BBA authorizing the Secretary to provide payment to nonhospital sites and we do not believe a demonstration project is necessary. Furthermore, since this policy is more stringent than existing regulations, we are doubtful that hospitals would participate voluntarily in a demonstration project. Comment: One commenter objected to the revision of the ``all or substantially all'' criteria and stated that the proposed policy would constrain the ability of teaching hospitals and Medicare+Choice organizations to develop reasonable rotations in hospitals and managed care plans. The commenter suggested an alternative under which a Medicare+Choice organization could submit a short application that would contain agreements between hospitals and Medicare+Choice organization addressing, among other things, the amount of time residents would spend at each site. Under this approach, we would pay the qualified nonhospital provider based on the product of a per resident amount, the number of FTE residents, and the Medicare share. Each resident would be counted as a partial FTE based for the hospital and for the qualified nonhospital provider based on the percentage of time worked at each site. A Medicare+Choice organization would be paid its FTE percentage times a portion of the hospital per resident payment amount or a national average per resident amount. This commenter argued that this approach would meet the Congressional objective of allowing residents to receive training in hospitals and Medicare+Choice organizations while prohibiting double payment without establishing a cumbersome new set of cost reporting requirements. Response: We considered the approach suggested by this commenter but we believe it would not facilitate training in qualified nonhospital providers. FQHC's, RHC's, and Medicare+Choice organizations generally provide a low percentage of total services to Medicare beneficiaries. The commenter's approach would to some extent substitute the Medicare share of the qualified nonhospital provider for the Medicare share of the hospital, and we believe this would result in lower Medicare payments overall for training in nonhospital sites. Also, we believe this approach would be inequitable to hospitals in that they would lose both the direct and indirect medical education payments for the proportion of time residents spend in the qualified nonhospital provider even though they have ongoing training costs while the residents train in the nonhospital site. We believe that it is reasonable to pay the hospital or qualified nonhospital provider which incurs ``all or substantially all'' of the costs. Furthermore, the revised definition reflects a better measure of ``all or substantially all'' of the costs and will result in appropriate payment to hospitals for training in qualified nonhospital providers and other nonhospital sites. As we stated in the May 8 proposed rule (63 FR 25597), we also have concerns that it would not be equitable to eliminate the hospital's payment entirely for the time resident's spend in nonhospital sites because the hospital may continue to incur some of the costs associated with training residents in nonhospital sites. We believe that the policies we are adopting are equitable to both hospital and nonhospital sites and will achieve Congress' objective of encouraging training in nonhospital sites. Comment: One commenter stated that there might be important differences in the accounting and administrative systems of various categories of qualified nonhospital providers that might present some difficulties in identifying the cost data necessary to accurately complete cost reporting forms. Other commenters stated that hospitals will have difficulty obtaining the necessary data from the nonhospital sites to complete the agreements or that the revised definition of ``all or substantially all'' would impose undue administrative burden. Another commenter stated that the revised definition of ``all or substantially all'' creates a major problem in identifying the portion of time office physicians spend in teaching and supervising residents and is another administrative burden placed on physicians. Response: As stated before, we do not believe we are imposing undue administrative burden. Direct GME costs for FQHC's and RHC's will have to be separately identified and reported. Although this will require the development of a mechanism for FQHC's and RHC's to allocate overhead and supervisory physician costs to the GME costs center, we do not believe that our policy will create significant administrative difficulties for FQHC's and RHC's, which already prepare cost reports for Medicare. As stated previously, we do not believe this process will generate a substantial burden on supervising physicians in FQHC's and RHC's beyond a written agreement between the clinic and the physician regarding the amount of time the physician expects to spend in GME activities and a time study verifying the allocation. The submission of a cost statement for GME will be a new responsibility for Medicare+Choice organizations which do not have experience with reporting costs. However, as stated above, we are developing a one page cost statement of GME expenses to limit the administrative burden on Medicare+Choice organizations. With regard to the concern expressed about creating a burdensome set of new cost reporting requirements, we reiterate that a condition of payment to the hospital for training in the nonhospital site is the production of the written agreement between the hospital and the nonhospital site. We are not requiring hospitals to submit cost data to Medicare as a precondition to counting the resident for indirect and direct GME. Comment: One commenter noted that some arrangements between hospitals and nonhospital settings for the training of residents predate the GME base year. This commenter stated that hospitals did not compensate nonhospital sites for supervisory teaching physician costs and it would not be fair to shift these costs to teaching hospitals. The commenter also stated that teaching hospitals have already entered into written agreements with nonhospital sites under the existing rules. According to the commenter, the proposed rule would necessitate renegotiation of thousands of agreements, imposing tremendous transaction costs upon the academic medical community. The commenter noted that if the agreements are not renegotiated prior to the effective date, the hospital will be unable to count the residents for direct and indirect GME, and this will have a lasting effect because of the 3 year averaging rules. Another commenter stated that there are many complex [[Page 40995]] contractual arrangements between hospital based programs and nonhospital sites regarding the placement, training and patient service utilization of residents, and any change in Medicare GME payment policy could have significant and unknown impacts on these current training structures. Response: The GME provisions of this final rule will be effective January 1, 1999. All other provisions of this final rule are effective October 1, 1998. By making a later effective date for the GME provisions, hospitals and nonhospital sites will have 5 months following publication of this final rule to negotiate agreements that will allow hospitals to continue counting residents training in nonhospital sites for indirect and direct GME. These agreements are related solely to financial arrangements for training in nonhospital sites. We do not believe that the agreements regarding these financial transactions will necessitate changes in the placement and training of residents. In response to the comment that it is unfair to shift costs to the hospital, we believe it is appropriate to include supervisory costs in the nonhospital site as part of ``all or substantially all'' of the costs that hospitals must incur to count the resident. Currently, the hospital is able to count the resident even though its costs for that resident may be lower during the time the resident trains outside the hospital. At the same time, the nonhospital site may have incurred costs for which it received no compensation. We believe that requiring the hospital to incur the costs associated with training in the nonhospital site is equitable to both the hospital and nonhospital site and is consistent with the statutory requirement that the hospital must incur ``all or substantially all'' of the costs. Comment: One commenter argued that we should not use reasonable costs as the basis for making payment to qualified nonhospital providers. This commenter stated that Medicare+Choice organizations do not submit cost reports and it would be extraordinarily expensive and cumbersome to report accounting costs. Several commenters also objected to our proposal to the extent we would allow overhead costs for FQHCs, RHCs, and hospitals but not Medicare+Choice organizations. These commenters believed that the policy cannot be justified on the basis that Medicare+Choice organizations do not submit cost reports. One commenter suggested that HCFA use predetermined payment amounts that do not require the subsequent submission of cost reports. The commenter noted that the proposed rule itself notes that direct GME payments are based on average per resident costs from 1984 that might bear little or no relation to accounting costs in 1998. Another commenter suggested that Medicare+Choice organizations should be paid an overhead factor for direct GME costs based on square footage of the clinic and a number of other factors. Alternatively, this commenter suggested use of an average overhead factor based on the number of residents trained until actual overhead expenses for Medicare+Choice organizations can be identified. Response: Medicare+Choice organizations will typically contract with clinics for the provision of services to beneficiaries. In these situations, we can make payment directly to the Medicare+Choice organization if the plan produces a written agreement with the clinics where training occurs that the plan will incur ``all or substantially all'' of the costs associated with training in the nonhospital site. We are requiring a written agreement between the Medicare+Choice organization and the nonhospital sites. We believe that the primary components of GME costs are resident compensation and supervisory teaching physician costs and that facility overhead costs which can be allocated to direct GME are a smaller component of direct GME costs. Nevertheless, we agree that we should not limit allowable direct GME costs for Medicare+Choice organizations to resident compensation and supervisory physician costs. If the Medicare+Choice organization can document other direct GME costs that directly relate to a training program, we will allow these costs. We note that, at this time, it is not feasible to develop an average overhead factor which can be paid to Medicare+Choice organizations that incur ``all or substantially all'' of the costs of a training program in a nonhospital site. This is because our data systems on hospital GME costs do not distinguish between supervisory teaching physician costs and overhead costs attributable to direct GME. In response to the comment that we use square footage or other mechanisms as a basis for allocating overhead to GME costs for Medicare+Choice organizations, we are concerned about developing a sophisticated cost allocation process for determining Medicare+Choice allowable direct GME costs since Medicare+Choice organizations do not submit cost reports. However, we are revising our proposal to require the written agreement to state that the Medicare+Choice organization will incur the costs of residents' salaries and fringe benefits and provide reasonable compensation for the remaining costs of the training program in the nonhospital site. Based on the statement of costs, the Medicare+Choice organization will report its costs to HCFA and we will provide payment based on the lower of the Medicare+Choice organization's cost per resident or a national average of the hospital per resident amounts. Comment: Several commenters were concerned that if neither the hospital or nonhospital site incurs ``all or substantially all'' of the costs, neither setting would receive payment even though each entity incurs a portion of the training costs. One commenter suggested that there will be difficulty allocating costs under our proposed definition of ``incurring costs'' and stated that we should encourage affiliations and provide simpler and clearer guidance for institutions. Response: Under this final rule, an entity must incur ``all or substantially all'' of the costs to receive payments for the time the resident spends in the nonhospital site. Since we do not conduct cost- finding to determine who bears ``all or substantially all'' of the graduate medical education costs, we are generally dependent on hospital and non-hospital provider agreements to determine who bears them. As stated earlier in this final rule as well as in the proposed rule, we do not believe it would be administratively feasible to apportion payments appropriate to the hospital and nonhospital site in situations where neither the hospital or nonhospital site agree on who incurs ``all or substantially all'' of the costs. We must also consider the statutory prohibition on double payments in these situations. Furthermore, although it may be appropriate to provide payment for GME costs where the nonhospital site incurs only a portion of the training costs, we do not believe it would be equitable to allow a nonhospital site to be paid where it was incurring only a portion of the costs but only allow payment to a hospital when it incurs ``all or substantially all'' of the costs. In response to the commenter who suggested that we should encourage ``affiliations,'' we believe the revised definition of ``all or substantially all'' of the costs provides incentives for hospitals and nonhospital sites to reach agreement with regard to financial arrangements for training in nonhospital sites to avoid the situation where neither entity receives payment for GME. Comment: One commenter asked whether hospitals would be eligible to receive payments in situations where the teaching faculty volunteers their [[Page 40996]] services and neither the hospital or nonhospital entity incurs costs for supervisory teaching physicians, but the hospital incurs the costs of resident salaries and fringe benefits (including travel and lodging expenses where applicable). The commenter asked whether the contract should state that there are no teaching physician costs incurred and the remainder of the costs represent ``all or substantially all'' of the costs. Another commenter stated that the ``all or substantially all'' definition creates special problems where community physicians voluntarily serve in a teaching capacity without compensation. The commenter stated that the implication of the proposed policy is that some portion of the community physician's earnings must be included in the calculation and asked that we either delete the proposed change or specify that voluntary supervision of training residents does not need to be included in the definition of ``all or substantially all'' of the costs. Response: We have received anecdotal information that some supervisory teaching physicians participate in teaching activities without compensation in nonhospital clinics. Although there may be situations where a supervising physician is participating in teaching, we do not believe that lack of explicit compensation for teaching activities means that physicians are necessarily volunteering their time. Rather, we believe that the physician's compensation in the clinic encompasses both teaching and nonteaching activities. Nevertheless, for purposes of satisfying the requirement of a written agreement, the written agreement between a hospital and a nonhospital site may specify that there is no payment to the clinic for supervisory activities because the clinic does not have these costs. Comment: One commenter stated that a hospital was permitted to include, within in its GME base period costs, teaching physician costs related to the hospital by common ownership or control under Sec. 413.17. Citing the GME consistency principle at Sec. 412.113(b)(3), this commenter requested that we clarify that the same policy applies in the context of GME payment to nonhospital sites. That is, the regulation should include specific language which states that costs incurred by an organization related to the hospital under Sec. 413.17 will be recognized as if incurred by the hospital in applying the expanded definition of ``all or substantially all'' of the costs. Response: The consistency principle under Sec. 412.113(b)(3) required consistent treatment of medical education costs during the transition to the inpatient hospital PPS during the 1980s. This rule was intended to prevent medical education costs from being included in hospital payments for operating costs and also being paid on a reasonable costs basis to hospitals as GME during the early years of the PPS. We do not see a relationship between the consistency rule and our proposed policies with regard to payment for GME training in nonhospital sites. With regard to the costs of related parties under Sec. 413.17, our policy was not to include costs associated with training in nonhospital clinics in the per resident amount even though certain direct GME costs of related parties could have been allowable. We also do not believe that Sec. 413.17 has applicability to our proposed policy. We are requiring a written agreement between hospitals and nonhospital sites for purposes of this final rule, even where the hospital and nonhospital site are related organizations under Sec. 413.17. In practice, since we are requiring an agreement between hospitals and nonhospital sites that are under common ownership or control, the agreements should be a formality. Comment: One commenter stated that the necessary statutory and regulatory incentives do not exist for teaching hospitals to provide compensation to nonhospital sites for their GME costs. Response: We disagree. The proposed rule requires a written agreement between the hospital and nonhospital site that the hospital will provide compensation to the nonhospital site for certain types of GME costs. Without this agreement, the hospital will be unable to count the resident for indirect and direct GME. As stated earlier, the agreements must also indicate the amounts the hospital will actually pay to the nonhospital site for GME training. Comment: One commenter stated the definition of ``all or substantially all'' of the costs should not include residents' travel and lodging costs. This commenter stated that there is no rationale for this change and that the criteria imposes significant reporting burdens with no offsetting benefits. The commenter also stated that the phrase ``where applicable'' is vague and requires additional definition language (related to distance, means of travel) if entities are to understand their reporting obligations. Response: Our intent in adding the phrase ``including residents travel and lodging costs, where applicable'' was to provide for the inclusion of direct GME costs that may be more prevalent in a nonhospital setting than in the hospital setting. The phrase ``where applicable'' means that depending on the specific arrangement in some cases, residents will be responsible for paying their own travel and lodging costs while serving at the nonhospital site. In other cases, it is possible that the site will pay for the residents to travel to the site and for lodging while at the site. This is basically a fringe benefit paid by the site for the resident. Therefore, in situations where travel and lodging is an expense of the nonhospital site while the resident is training there, the written agreement must indicate that the hospital will incur these costs. In determining whether the hospital has incurred ``all or substantially all'' of the costs of the program, the hospital must include this ``unique'' fringe benefit if it was paid for by the nonhospital site. Comment: One commenter stated that the proposed regulations effectively deny payments to FQHC's unless they incur ``all or substantially all'' of the costs of the program. The commenter stated that since the FQHC does not typically pay the residents' salaries, the proposed rule does not significantly increase the ability of the FQHC to recover GME costs. This commenter stated that it is eminently possible to devise a method under which hospitals that utilize qualified nonhospital providers would report costs showing allowable FQHC costs. In these situations, costs would be apportioned to the proper cost center. Response: We disagree. The FQHC can recover its GME costs either directly from Medicare if it incurs ``all or substantially all'' of the costs, or from the hospital through the written agreement. Without a written agreement that specifies the amounts the hospital will pay the nonhospital site for training in the nonhospital site, the hospital will be unable to count the resident for indirect and direct GME. d. Medicare share. Comment: One commenter stated that the limitation of direct GME payments to FQHC's based on Medicare's share at the FQHC will seriously constrain participation because only 8 percent of FQHC patients are Medicare patients. The commenter quoted the Conference Report which states that ``the Conferees believe this authority may help alleviate physician shortages in rural areas.'' According to this commenter, the combination of requiring the FQHC to incur ``all or substantially all'' of the costs in order to receive payment and the limitation to Medicare share does little to provide sufficient resources to allow FQHC's to train physicians in underserved rural areas. The commenter believed the limitation of payments based on Medicare's share is not [[Page 40997]] required by the BBA provision authorizing GME to qualified nonhospital providers and is contrary to the intent of the law. Response: It is a fundamental and longstanding principle that, to the extent Medicare pays for certain types of costs, the Medicare program should pay only its fair share. This principle applies not only in the context of Medicare payment for medical education, but also to Medicare payment in general. Comment: One commenter stated that Medicare enrollees use 3.5 times the number of outpatient services as non-enrollees. The commenter suggested that it would be more equitable to base Medicare's share for Medicare+Choice organizations on the ratio of outpatient expenses for Medicare enrollees to total enrollees. As an alternative, this commenter suggested using Medicare visits to total visits to calculate Medicare share, consistent with the calculation in the inpatient setting of Medicare inpatient days to total inpatient days. Response: We believe that either of the proposals suggested by this commenter would impose significant additional reporting responsibilities on Medicare+Choice organizations which receive payment from Medicare for direct GME. Basing the Medicare share calculation on the ratio of outpatient expenses attributable to Medicare beneficiaries to total expenses would require Medicare+Choice organizations to provide a sophisticated report of expenses not unlike the Medicare cost report. In situations where the Medicare+Choice organization is contracting for services provided in a clinic, this would require the Medicare+Choice organization to document costs which are not even its own. We considered using the ratio of Medicare enrollee to total enrollee visits in the Medicare share calculation, but have concerns that this approach would also be burdensome in that it would require Medicare+Choice organizations to furnish utilization data for clinics or physician offices that they do not own or control. e. National average per resident amounts. Comment: One commenter argued that national average per resident amounts are not appropriate for the nonhospital setting. According to the commenter, residency training differs from other types of services because it involves complicated transactions with nongovernmental entities such as medical schools that may sponsor a hospital's programs and compensate physicians directly, and accreditation bodies that may require a certain content and curriculum in training programs. Response: We did not propose the use of national average per resident amounts in the nonhospital setting but will consider whether a national average per resident amount is appropriate after we have experience with the provision and have reliable data on the costs of training in the nonhospital setting. f. Technical errors concerning GME policy published in the May 12, 1998 final rule. In the May 12, 1998 final rule for the FY 1998 inpatient hospital prospective payment system, we set forth certain policies on GME. The portion of the May 12, 1998 final rule concerning counting residents for direct medical education (beginning at (63 FR 26327)) contained the following technical errors: Merged Hospitals--On page 26329, third column, we stated that the FTE cap of merged hospitals would be the aggregation of the FTE cap for each hospital participating in the merger. We stated that Sec. 413.86 would be modified to reflect this policy, but we did not modify the regulations text. We do not believe a change to the regulations text is necessary. Application of the FTE Cap--There is a discrepancy between the methodologies described in the August 29, 1997 final rule with comment period (62 FR 46005) and the May 12, 1998 final rule (63 FR 26330) for application of the FTE cap in situations where a hospital has more residents than the cap. The methodology described in the May 12, 1998 final rule is incorrect. The correct methodology is described in the August 29, 1997 final rule with comment period. New Medical Residency Training Program--On page 26332, in the first column, we stated, ``for these reasons, we believe it is appropriate to consider a medical residency training program to be newly established if the program received initial accreditation or began training residents on or after January 1, 1995.'' We are clarifying that, for hospitals that trained residents prior to January 1, 1995, we will adjust the FTE caps for programs were accredited or began training residents on or after January 1, 1995 and prior to August 5, 1997. Application of the FTE Cap to an Affiliated Group--On page 26341, in the third column, we stated, ``If the combined FTE counts for the individual hospitals do not exceed the aggregate cap, we will pay each hospital based on its FTE cap as adjusted per agreements.'' That sentence should have read as follows: ``If the combined FTE counts for the individual hospitals exceed the aggregate cap, we will pay each hospital based on its FTE cap as adjusted per agreements.'' V. Changes to the Prospective Payment System for Capital-Related Costs A. Cap on the Capital Indirect Medical Education Adjustment Ratio (Sec. 417.322) Under section 1886(g) of the Act, the Secretary has broad discretion in implementing the capital prospective payment system. Section 412.322 of the regulations specifies the formula for the capital indirect medical education (IME) adjustment factor. The capital IME adjustment is intended to pay the Medicare capital prospective payment system share of the indirect costs of medical education to teaching hospitals. The formula was incorporated in the August 30, 1991 final rule for the capital prospective payment system (56 FR 43380), and uses the ratio of interns and residents to average daily census (defined as total inpatient days divided by the number of days in the cost reporting period). Section 1886(d)(5)(B) of the Act requires the use of the ratio of residents-to-beds to calculate the IME adjustment for the operating prospective payment system. However, pursuant to our authority under section 1886(g) of the Act, we adopted the resident to average daily census ratio for the capital prospective payment system because we believed it was a more appropriate method for measuring teaching intensity, and because we believed it was less subject to manipulation. The IME adjustment factor increases by approximately 2.8 percentage points for each 0.10 increase in the hospital's ratio of residents to average daily census. The IME adjustment for inpatient capital-related costs for hospitals paid under the prospective payment system takes the form of [e raised to the power (.2822 x ratio of interns and residents to average daily census)-1] where e is the natural antilog of 1, based on the total cost regression results. In order to determine the Federal rate portion of the hospital's payment, the IME adjustment factor is multiplied by the standard Federal rate, the DRG weight, the geographic adjustment factor, and any other relevant payment adjustments such as the DSH adjustment or the large urban add-on. The formula is as follows: (Standard Federal Rate) x (DRG weight) x (GAF) x (Large Urban Add-on, if applicable) x (COLA adjustment for hospitals located in Alaska and Hawaii) x (1 + Disproportionate Share Adjustment Factor + IME Adjustment Factor, if applicable). [[Page 40998]] In the May 8, 1998 proposed rule (63 FR 25600) we indicated that it had come to our attention that because of the application of the capital IME adjustment, one hospital would receive a capital IME payment greater than its total hospital costs. We also stated that of the approximately 1,200 teaching hospitals in the United States, based on December 1997 data, 8 hospitals had a resident to average daily census ratio of more than 1.5. A resident to average daily census ratio of 1.5 results in a capital IME adjustment factor of 0.53, which increases the Federal rate portion of the hospital's capital payment by 53 percent. To address this unintended effect of the capital IME methodology, we proposed capping the capital IME ratio at 1.5. A ratio greater than 1.5 means a hospital has, on average, considerably more residents than inpatients. Capping the ratio at 1.5 would allow for one resident per patient on the inpatient side plus some outpatient training, and would keep capital IME payments more consistent with the costs incurred. Because the operating IME ratio is based on the number of beds, it has only slightly exceeded 1.0 in two cases. This change would ensure that the capital IME adjustment is more in line with hospital costs. We received no comments on our proposed change. We have decided to implement this policy as proposed. Effective October 1, 1998, the capital IME ratio will be capped at 1.5. B. Payment Methodology for Mergers Involving New Hospitals (Sec. 412.331) The August 30, 1991 final rule (56 FR 43418), which implemented the capital prospective payment system, established special payment provisions for new hospitals. Under Sec. 412.324(b), a new hospital is paid 85 percent of its allowable Medicare capital-related costs through its first cost reporting period ending at least 2 years after the hospital accepts its first patient. The first cost reporting period beginning at least 1 year after the hospital accepts its first patient is the hospital's base year for purposes of determining its hospital- specific rate. Section 412.302(b) defines a new hospital's old capital costs as allowable capital-related costs for land and depreciable assets that were put in use for patient care on or before the last day of the hospital's base year cost reporting period. Beginning with the third year, the hospital is paid under the fully prospective or hold- harmless payment methodology, as appropriate. If the hospital is paid under the hold-harmless payment methodology, the hospital's hold- harmless payments for its old capital costs can continue for up to 8 years. In the August 30, 1991 final rule, we defined a new hospital as one that had operated (under previous or present ownership) for less than 2 years and did not have a 12-month cost reporting period that ended on or before December 31, 1990. In the September 1, 1992 final rule (57 FR 39789), as a result of situations brought to our attention after publication of the original prospective payment system final rule, we clarified that the new hospital exemption would not apply in situations where the facility was not truly a new hospital. In the May 8, 1998 proposed rule (63 FR 25600), we indicated that questions had arisen regarding application of our rules for payment of new hospitals in merger situations. We stated that consistent with our previously stated policy, we were proposing to further clarify the new hospital payment provisions. We proposed that, if during the period it is eligible for payment as a new hospital (as defined at Sec. 412.300(b) and Sec. 412.328(b)), a new hospital merges with one or more existing hospitals, and the merger meets the existing capital- related reasonable cost rules regarding the criteria for recognizing a merger at Sec. 413.134 and the new hospital is the surviving corporation (as defined in Sec. 413.134(l)(2)), we would treat as old capital only those assets of the existing hospital that met the definition of old capital (as defined in Sec. 412.302(b)) prior to the merger, for purposes of determining payments after the merger. Any assets of the existing hospital that were considered new capital prior to the merger would still be considered new capital after the merger. However, the merger cannot be used to convert the existing hospital's new capital into old capital. After the merger, the discharges of each campus of the merged entity would maintain their pre-merger payment methodology until the end of the 2-year period that the new hospital campus is eligible for reasonable cost reimbursement as defined at Sec. 412.324(b). That is, the discharges at the new hospital would be paid based on 85 percent of its allowable Medicare hospital capital-related costs, while discharges from the existing hospital would continue to be paid under that hospital's methodology, that is, fully prospective or hold-harmless. At the end of this period, the intermediary would calculate a hospital specific rate for the ``new'' campus of the merged hospital. Finally, the calculation methodology for hospital mergers at new Sec. 412.331(a)(1) and (2) would be performed and a combined hospital-specific rate would be determined and a payment methodology selected for the merged hospital as a whole. The calculation at Sec. 412.331(a)(1) and (2) uses each hospital's base year old capital costs. Any new capital of the previously existing hospital would not be used in the determination. If the merged entity qualifies for the hold-harmless payment methodology, only the capital which meets the definition of old capital at Sec. 412.302(b) would be eligible for hold-harmless payments. We received one comment on our proposal. Comment: One hospital association commented on the policy that only the assets of the existing hospital that met the definition of old capital prior to the merger would be treated as old capital after the merger, even if all of the capital had been acquired and put into use during the new hospital's base year. They also stated that the proposal changes the regulatory definition of a new hospital's old capital, revises its payment methodology determination, and creates special payment rules for new hospitals that merge with existing hospitals. The commenter also states that a hospital in a situation similar to that described in our example was told that after a merger between a new hospital and an existing hospital, all assets acquired by the new hospital in the base year would become old capital costs. The commenter suggests that if HCFA will not reconsider the proposed change, at least it should not be applied retroactively. Response: As indicated in the proposed rule, we addressed this issue because questions have arisen regarding application of our rules for payment for new hospitals in merger situations. Accordingly, we proposed to clarify the application of our rules in merger situations. Before the proposed rule, we had not specifically addressed in the Federal Register the issue of mergers between an ``existing'' hospital and a ``new'' hospital, but our clarification is consistent with existing rules; the clarification does not reflect new policy or a change in policy that can only be applied prospectively. The commenter is correct that with regard to the capital of the existing hospital that merges with a new hospital, our proposal would treat as old capital only capital that qualified as old capital prior to the merger. Any capital that was new capital of the existing hospital prior to the merger would remain new capital after the merger. The new hospital will be paid 85 percent of its allowable Medicare inpatient hospital capital-related costs through its [[Page 40999]] cost reporting period ending at least two years after the hospital accepts its first patient. In our September 1, 1992 final rule (57 FR 39789), we clarified that the new hospital exemption under the capital prospective payment system would not apply to a facility that opened as an acute care hospital if that hospital had previously operated under current or prior ownership and had a historic asset base. We also clarified that even a hospital that replaced its entire facility (with or without a change of ownership) would not qualify for a new hospital exemption and that a previously existing PPS-excluded hospital (paid under section 1886(b) of the Act) that became an acute care hospital (paid under section 1886(d)) of the Act would not qualify as a new hospital. With this current proposal we are clarifying our rules as they apply to a new hospital which merges with an existing hospital. When a new hospital merges with an existing hospital that has already had the benefit of reasonable cost reimbursement prior to the inception of capital PPS, on October 1, 1991, we believe it would be inappropriate for all of the capital assets of a previously existing hospital to be eligible for payment as old capital simply because it merged with a new hospital. As with the other situations that we clarified in 1992, this current clarification of the regulation at Sec. 412.331(a)(3) is consistent with the principle that the new hospital exemption should only be available to those hospitals that had not received reasonable cost payments in the past and needed special payment protection during their initial period of operation. Our policy seeks to ensure that when a new hospital acquires the assets of an existing hospital through a merger, any assets of the existing hospital that were previously considered new capital prior to the merger are not transformed to old capital, as a result of the merger. The new hospital will still be paid 85 percent of its allowable Medicare capital-related costs for all other assets it acquires through the end of its base period. The commenter fails to note that our current payment rules at Sec. 412.331(a)(3) for merger situations already provide that only the existing capital-related costs related to the assets of each merged or consolidated hospital as of December 31, 1990 are recognized as old capital costs during the transition period. If the merged hospital is paid under the hold-harmless methodology after merger or consolidation, only that original base year old capital is eligible for hold-harmless payments. These rules mean that in cases of a merger between two existing hospitals, only the capital assets which were recognized as old capital prior to December 31, 1990 are eligible for payment as old capital after the merger. We are clarifying that this principle would also apply to the situation of merger between an existing hospital and a new hospital. The regulation that defines a new hospital's old capital was not intended to apply to capital acquired through merger with an existing hospital subject to capital PPS. Finally, the commenter is mistaken that HCFA has previously ruled that the new capital assets of an existing hospital could be paid as old capital after a merger with a new hospital. In fact, our policy is consistent with our regulation at Sec. 412.331(a)(3) cited above, in that only the existing capital-related costs related to the assets of each merged or consolidated hospital as of December 31, 1990 are recognized as old capital costs during the transition period. We are implementing this clarification as proposed. For an example of how our policy works, see the May 8, 1998 proposed rule (63 FR 25601). C. Special Exceptions Process As described in Sec. 412.348(g) of the regulations, an additional payment may be made for up to 10 years beyond the end of the capital PPS transition period for eligible hospitals that meet: (1) a project need requirement, (2) a project size requirement, and, (3) in the case of certain urban hospitals, an excess capacity test. The regulation establishing this special exceptions provision, and describing the criteria by which eligible hospitals qualify, was published on September 1, 1994 (59 FR 45385). At that time we described the purpose of the special exceptions process as ``* * * narrowly defined, focusing on a small group of hospitals who found themselves in a disadvantaged position. The target hospitals were those who had an immediate and imperative need to begin major renovations or replacements just after the beginning of the capital prospective payment system. These hospitals would not be eligible for protection under the old capital and obligated capital provisions, and would not have been allowed any time to accrue excess capital prospective payments to fund these projects.'' The special exceptions process is available to certain classes of hospitals that meet the eligibility criteria described at Sec. 412.348(g)(1). The eligible classes of hospitals are sole community hospitals; urban hospitals with at least 100 beds that either have a disproportionate share percentage of 20.2 percent or receive at least 30 percent of their revenue from State or local funds for indigent care; and hospitals with a combined inpatient Medicare and Medicaid utilization of at least 70 percent. Eligible hospitals must satisfy a project need requirement as described at Sec. 412.348(g)(2) and a project size requirement as described at Sec. 412.348(g)(5). For hospitals in States with Certificate of Need (CON) requirements, the project need requirement is satisfied by obtaining CON approval. For other hospitals, the project need requirement is satisfied by meeting an age of assets test. The project size requirement is satisfied if the hospital completes the qualifying project during the period beginning on or after its first cost reporting period beginning on or after October 1, 1991 to the end of its last cost reporting period beginning before October 1, 2001, and the project meets certain cost thresholds specified in the regulations. The minimum payment level for qualifying hospitals is 70 percent of allowable capital-related costs. A qualifying hospital may receive payments for up to ten years from the year which it completes a qualifying project. Finally, the regulations at Sec. 412.348(g)(8) describe the cumulative payment comparison and offsetting amounts which are used to determine a qualifying hospital's exception payment. A few hospitals have expressed concern with the required completion date of October 1, 2001, and other qualifying criteria for the special exceptions. When we established the special exceptions process, we selected the hospital's cost reporting period beginning before October 1, 2001 as the project completion date, because hospitals are eligible to receive special exceptions payments for up to ten years from the year in which they complete their project. If a project is completed by September 30, 2001, then exceptions payments could continue up to October 30, 2011. We intended to limit cost-based exceptions payments to the period not more than ten years beyond the end of the transition to fully prospective payment for capital. When we adopted the criteria for the special exceptions process, we selected the project completion date with the goal of not extending this transition unnecessarily. In addition, we believed that eligible hospitals will not have had the opportunity to reserve prior year capital PPS payments for financing projects begun in the early years of PPS. [[Page 41000]] In order for us to analyze the impact of potential changes in the special exceptions policies, we are soliciting the following information on major capital construction projects as defined at Sec. 412.348(g)(5) that will be put to use for patient care on or after October 1, 1996: (1) Name, address, phone number and provider number of hospital; (2) Cost of capital project; (3) Date of CON approval, if required; (4) Start date of project; and (5) Anticipated completion date. Please forward this information by September 30, 1998 to the Division of Acute Care, Attention: Cassandra Black at the following address: HCFA, C4-01-26, 7500 Security Blvd., Baltimore, Md. 21244- 1850. We will analyze the data to determine whether any changes in the special exceptions policies are necessary. Any changes, if necessary, would be included in next year's FY 2000 proposed rule for hospital PPS. VI. Changes for Hospitals and Units Excluded From the Prospective Payment System Limits on and Adjustments to the Target Amounts for Excluded Hospitals and Units (Sec. 413.40(g)) 1. Updated Caps Section 1886(b)(3) of the Act as amended by section 4414 of the BBA established caps on the target amounts for excluded hospitals and units for cost reporting periods beginning on or after October 1, 1997, through September 30, 2002. The caps on the target amounts apply to the following three categories of excluded hospitals: psychiatric hospitals and units, rehabilitation hospitals and units, and long-term care hospitals. For purposes of calculating the caps, the statute requires the Secretary to first calculate the 75th percentile of the target amounts for each class of hospital (psychiatric, rehabilitation, or long-term care) for cost reporting periods ending during FY 1996. The resulting amounts are updated by the market basket percentage to the applicable fiscal year. A discussion of how the caps on the target amounts were calculated for cost reporting periods beginning during FY 1998 can be found in the August 29, 1997, final rule with comment period (62 FR 46018). On March 6, 1998, we published a correction notice correcting the caps for FY 1998 (63 FR 11148). In the May 8 proposed rule for FY 1999, we published proposed caps for cost reporting periods beginning during FY 1999 (63 FR 25601); however, the caps that we published inadvertently reflected updates to the amounts published on August 29, 1997, rather than the corrected amounts published on March 6, 1998 (see May 13, 1998 correction notice, 63 FR 26565). Thus, as corrected, the proposed caps for FY 1999 were as follows: (1) Psychiatric hospitals and units: $10,797 (2) Rehabilitation hospitals and units: $19,582 (3) Long-term care hospitals: $38,630 These proposed caps reflected an update of 2.5 percent, the projected market basket percentage increase at the time we developed the proposed rule. The final projection of the market basket percentage for excluded hospitals and units for FY 1999, based on the most recent data available, is 2.4 percent. Accordingly, the final caps on the target amounts for existing hospitals for cost reporting periods beginning during FY 1999 are as follows: (1) Psychiatric hospitals and units: $10,787 (2) Rehabilitation hospitals and units: $19,562 (3) Long-term care hospitals: $38,593 2. New Excluded Hospitals and Units (Sec. 413.40(f)) Section 1886(b)(7) of the Act establishes a new statutory payment methodology for new psychiatric hospitals and units, rehabilitation hospitals and units, and long-term care hospitals. Under the statutory methodology, for a hospital that is within a class of hospitals specified in the statute and which first receives payments on or after October 1, 1997, the amount of payment will be determined as follows. For each of the first two cost reporting periods, the amount of payment is lesser of (1) the operating costs per case, or (2) 110 percent of the national median of target amounts for the same class of hospitals for cost reporting periods ending during FY 1996, updated and adjusted for differences in area wage levels. In the August 29, 1997 final rule with comment period, we published the figures for 110 percent of the national median of target amounts for each class of hospital (62 FR 46020). In the May 12, 1998 final rule for FY 1998, we revised the figure for long-term care hospitals to $21,494 (63 FR 26347). The table below lists 110 percent of the wage neutral national median target amounts for each class of excluded hospitals for cost reporting periods beginning during FY 1999. These figures reflect updates to the final FY 1998 figures by the projected market basket increase of 2.4 percent. For a new provider, the labor-related share of the target amount should be multiplied by the appropriate geographic area wage index and added to the nonlabor-related share in order to determine the limit on payment under the statutory payment methodology for new providers. ------------------------------------------------------------------------ Labor- Nonlabor- Total related related share share ------------------------------------------------------------------------ (1) Psychiatric................................... $6,214 $2,472 (2) Rehabilitation................................ 12,219 4,858 (3) Long-Term Care................................ 15,749 6,261 ------------------------------------------------------------------------ 3. Classification of Hospitals and Units (Sec. 413.40(c)) In the May 8 proposed rule, we stated that, after publication of the August 29, 1997 final rule with comment period, some excluded facilities had suggested that if they are currently excluded as one class of hospital or unit but also qualify for exclusion as another class of hospital, they should be permitted to choose which classification applies for purposes of applying the cap on target amounts. For example, some hospitals that participate in Medicare as psychiatric hospitals (defined under section 1861(f) of the Act, and the special conditions of participation in 42 CFR part 482 subpart E) have noted that they have average lengths of stay greater than 25 days. Those hospitals have asked to be ``reclassified'' as long-term care hospitals and given the benefit of the higher cap on target amounts applicable to that hospital class. In the proposed rule, we indicated that we had considered these hospitals' suggestions but, for reasons explained in that document, believed it would not be appropriate to adopt them. Accordingly, in the May 8 proposed rule, we proposed to revise Sec. 413.40(c)(4)(iii) to specify that, for purposes of that paragraph, the classification of a hospital that was excluded from the prospective payment system for its cost reporting period ending in FY 1996 would be determined by its classification (that is, the basis on which it was excluded) in FY 1996. If a hospital or unit was not excluded for a cost reporting period ending in FY 1996, but could be excluded on more than one basis (for example, as either a rehabilitation or long-term care hospital) in a given cost reporting period, it would be assigned to the classification group with the lowest limit. Comment: One commenter agreed that psychiatric hospitals should not be allowed the higher cap on target amounts that is applicable to long- term care hospitals, even if they also have average lengths of inpatient stay greater than 25 days. The commenter pointed out that psychiatric hospitals participate in Medicare under a provision of the law (section 1861(f) of the Act) that is separate from the provision applicable [[Page 41001]] to other excluded hospitals (section 1861(e) of the Act), and that the exclusion criteria for psychiatric hospitals differ from those for other hospitals. The commenter stated that because of these differences, a psychiatric hospital could not qualify for exclusion as another type of hospital or be eligible for the cap that applies to another type of hospital. The commenter suggested that it is unnecessary to specify that a psychiatric hospital cannot qualify for the cap on target amounts applicable to long-term care or other types of excluded hospitals. Response: If a hospital qualifies under more than one of the exclusion criteria pursuant to section 1886(d)(1)(B) of the Act, we would apply the lowest applicable cap to the hospital. For example, where a hospital qualifies as both a rehabilitation and long-term care hospital, we will apply the lower rehabilitation hospital cap to the hospital. Since this rule applies to all PPS-excluded hospitals, whether a psychiatric hospital can qualify as another type of hospital or not, the policy of applying the lowest cap is still needed. Comment: One commenter pointed out that some non-psychiatric (section 1861(e) of the Act) hospitals might be able to qualify for exclusion either as rehabilitation or as long-term care hospitals. The commenter stated that in many cases such facilities are excluded as long-term care hospitals. Therefore, the commenter recommended that any hospital in this category be given the benefit of the long-term care hospital cap. Response: We understand that some hospitals may simultaneously be able to qualify for exclusion on more than one basis. If a hospital is excluded from PPS as a certain type of hospital, we believe the hospital should be subject to the cap applicable for that class of hospital, even if it qualifies for exclusion on another basis. Thus, if a hospital qualifies for exclusion on more than one basis, then it is subject to all applicable caps, which in turn means the hospital's target amount cannot exceed the lowest of the applicable caps. We believe this policy not only is appropriate, but also provides greater incentives for efficient and cost-effective operation. Comment: Two commenters stated that if a hospital is classified as one type of hospital in any period to which the limits apply, and does not simultaneously qualify for exclusion on any other basis, the law (section 1886(b)(3) of the Act) does not authorize application of any cap other than the one applicable to the exclusion category to which the hospital is assigned. One commenter stated that this is the case even if the basis for the hospital's exclusion in a given cost reporting period is different than the basis for its exclusion for the cost reporting period ending during FY 1996 (for example, a hospital may have been excluded as a rehabilitation hospital during that period and later qualified for exclusion as a long-term care hospital). Response: We agree with the commenter that, if the basis for a hospital's exclusion for a given cost reporting period is different than the basis for the hospital's exclusion for the cost reporting period ending during FY 1996, the earlier basis of exclusion should not control which cap applies. We are revising Sec. 413.40(c)(4)(iv) accordingly. Thus, in applying the caps to excluded hospitals (or units), we will consider only the current basis (or bases) for exclusion. As stated above, if a hospital qualifies for more than one type of exclusion, its target amount may not exceed the lowest of the applicable caps. We note that, for the reasons explained in the proposed rule, we continue to be concerned that hospitals and units may seek changes in their basis of exclusion solely to take advantage of a higher cap, and that the resulting changes could compromise the effectiveness of the caps. We will monitor this situation carefully and may seek further legislative changes to the extent necessary to preserve the effectiveness of the caps. Comment: One commenter recommended that the regulations be revised to state that where two hospitals who are subject to different caps on TEFRA limits merge, the TEFRA cap that applies is the cap of the surviving hospital. Response: If two hospitals merge, the cap that applies depends on the status of the surviving entity. However, we do not believe that the regulations as described above, can be interpreted in any other way. Therefore, we do not agree that the regulations need to be revised to specifically address this situation. Comment: One commenter suggested that if a new hospital subject to the limits revised under Sec. 413.40(f)(2)(ii) changes the basis on which it is excluded from the PPS (for example, from being a rehabilitation hospital to a long-term care hospital), the cap applied for purposes of the comparison should be the cap applicable to the hospital's ``current'' exclusion category, not the hospital's previous exclusion category. Response: We agree that the cap applied should be based on the exclusion category for which the hospital currently qualifies. In light of the changes made in response to comments described above, we do not believe the regulations need to be further revised. 4. Exceptions The August 29, 1997 final rule with comment period (62 FR 46018) specified that a hospital that has a target amount that is capped at the 75th percentile, would not be granted an adjustment payment to the target amount (also referred to as an exception payment) as governed by Sec. 413.40(g)(3) based solely on a comparison of its costs or patient mix in its base year to its costs or patient mix in the payment year. Since the hospital's target amount would not be determined based on its own experience in a base year, any comparison of costs or patient mix in its base year to costs or patient mix in the payment year would be irrelevant. In addition, in the May 8, 1998 proposed rule, we proposed to clarify that, to the extent we grant an exception in accordance with Sec. 413.40(g)(3) to a hospital not affected by the cap, the amount of the exception would be limited to the cap on the hospital's target amount. By establishing caps on TEFRA target amounts, Congress has limited payments to individual hospitals based on amounts that reflect the cost experience of other hospitals. Therefore, in determining the extent of any adjustment paid to a hospital as an exception under our regulations at Sec. 413.40(g)(3), we believe it is consistent with Congressional intent to limit the extent of the adjustment to the hospital's cap on its target amount. We proposed to revise Sec. 413.40(g)(1) in order to set forth the limitation on the adjustment payments. Comment: One commenter stated that the proposed rule conflicts with section 1886(b)(4)(A)(i) of the Act, which requires HCFA to provide for adjustments to providers who exceed their TEFRA ceiling. The commenter also believed that our proposed provision limiting the TEFRA exception to the TEFRA cap is inconsistent with HCFA's past TEFRA adjustment processing practices. The commenter also stated that the proposed rule would adversely affect beneficiaries by limiting the scope and extent of services that hospitals in high wage areas are financially able to deliver. For these reasons, the commenter requested that HCFA modify the proposed rule to permit the granting of exceptions to the TEFRA cap. Response: Section 1886(b)(4)(A)(i) of the Act provides that the Secretary [[Page 41002]] ``shall provide'' for exceptions and adjustments ``where events beyond the hospital's control or extraordinary circumstances, including changes in the case mix of such hospital, create a distortion in the increase in costs for a cost reporting period.'' Prior to the enactment of Public Law 105-33, the payment for each excluded hospital was limited by a hospital-specific target amount, which was updated each year. The exceptions and adjustments provision provided for payments above the hospital's target amount if the hospital experienced ``a distortion in the increase in costs'' for a given period. Thus, a hospital could receive an exception based on its cost experience. The BBA enacted a system of caps which significantly changed the TEFRA payment system. Under the new system of TEFRA caps, a hospital's payments are not based solely on its own cost experience; instead, a hospital is now subject to a cap based on the cost experience of other hospitals. We believe our policies harmonize the exceptions provision and the cap provision. Under our policies, a hospital whose target amount is below the cap may receive an exception up to the cap. Thus, consistent with the mandate of section 1886(b)(4) of the Act, we continue to provide for exceptions, contrary to the assertion of the commenter. However, by establishing caps on TEFRA target amounts, Congress has limited payments to individual hospitals based on amounts that reflect the cost experience of other hospitals. Therefore, in determining the extent of any adjustment paid to a hospital as an exception under our regulations, we believe it is consistent with Congressional intent to limit the extent of the adjustment to the hospital's cap on its target amount. If a hospital's otherwise applicable target amount is above the cap, it cannot receive an exception based solely on a comparison of its current year costs or patient mix to base year costs or patient mix. VII. MedPAC Recommendations As required by law, we have reviewed the March 1998 report submitted by MedPAC to Congress and gave its recommendations careful consideration in conjunction with the proposals set forth in the proposed rule. We also responded to the individual recommendations in the proposed rule. The comments we received on the treatment of the MedPAC recommendations are set forth below along with our responses to those comments. However, if we received no comments from the public concerning a MedPAC recommendation or our response to that recommendation, we have not repeated the recommendation and response in the discussion below. Recommendations concerning the update factors for inpatient operating costs and for hospitals and hospital distinct-part units excluded from the prospective payment system are discussed in Appendix C, of this final rule. Potential Effects of Target Amount Caps Recommendation: The wage-related portion of the excluded hospital target amount caps should be adjusted by the appropriate hospital wage index to account for geographic differences in wages. (For more information see Volume 1, chapter 7, page 71 of the March 1998 report.) Response in the Proposed Rule: As MedPAC indicated in its recommendation, legislation would be required to adjust the target amount caps in such a substantial manner as to adjust for differences in area labor costs. Comment: Several commenters believed that the caps on the target amounts should be wage adjusted in order to recognize the different labor markets. They believe to do otherwise would be unfair and inequitable and may cause hospitals to cut back on services they provide to their Medicare beneficiaries. Response: We previously addressed this issue in the final rule published in the Federal Register on May 12, 1998 (63 FR 26345). Our decision, as expressed in our response in that final rule, remains unchanged. VIII. Other Required Information Requests for Data From the Public In order to respond promptly to public requests for data related to the prospective payment system, we have set up a process under which commenters can gain access to the raw data on an expedited basis. Generally, the data are available in computer tape format or cartridges; however, some files are available on diskette, and on the Internet at HTTP://WWW.HCFA.GOV/STATS/PUBFILES.HTML. In our May 8 proposed rule, we published a list of data files that are available for purchase (63 FR 25603). List of Subjects 42 CFR Part 405 Administrative practice and procedure, Health facilities, Health professions, Kidney diseases, Medicare, Reporting and recordkeeping requirements, Rural areas, X-rays. 42 CFR Part 412 Administrative practice and procedure, Health facilities, Medicare, Puerto Rico, Reporting and recordkeeping requirements. 42 CFR Part 413 Health facilities, Kidney diseases, Medicare, Puerto Rico, Reporting and recordkeeping requirements. 42 CFR Chapter IV is amended as set forth below: A. Part 405 is amended as follows: PART 405--FEDERAL HEALTH INSURANCE FOR THE AGED AND DISABLED 1. The authority citation for part 405 is revised to read as follows: Authority: Secs. 1102, 1861, 1862(a), 1871, 1874, 1881, and 1886(k) of the Social Security Act (42 U.S.C. 1302, 1395x, 1395y(a), 1395hh, 1395kk, 1395rr and 1395ww(k)), and sec. 353 of the Public Health Service Act (42 U.S.C. 263a), unless otherwise noted. Subpart X--Rural Health Clinic and Federally Qualified Health Center Services 2. In Sec. 405.2468, a new paragraph (f) is added to read as follows: Sec. 405.2468 Allowable costs * * * * * (f) Graduate medical education. (1) Effective for that portion of cost reporting periods occurring on or after January 1, 1999, if an RHC or an FQHC incurs ``all or substantially all'' of the costs for the training program in the nonhospital setting as defined in Sec. 413.86(b) of this chapter, the RHC or FQHC may receive direct graduate medical education payment for those residents. (2) Direct graduate medical education costs are not included as allowable cost under Sec. 405.2466(b)(1)(i); and therefore, are not subject to the limit on the all-inclusive rate for allowable costs. (3) Allowable graduate medical education costs must be reported on the RHC's or the FQHC's cost report under a separate cost center. (4) Allowable graduate medical education costs are non-reimbursable if payment for these costs are received from a hospital or a Medicare+Choice organization. (5) Allowable direct graduate medical education costs under paragraphs (f)(6) and (f)(7)(i) of this section, are subject to reasonable cost principles under part 413 and the reasonable compensation equivalency limits in Secs. 415.60 and 415.70 of this chapter. (6) The allowable direct graduate medical education costs are those costs [[Continued on page 41003]]