Government Affairs Home > Teaching Hospitals > Medicare Inpatient PPS > Historical Regulations & AAMC Summaries

Prospective Payment System: FY 1998 Final Rule

			AAMC Documents Issue Brief - Final Rule

AAMC Summary and Analysis

IME and DGME: Final Regulations Implementing Provisions of the Balanced Budget Act of 1997

On May 12, 1998 the Health Care Financing Administration (HCFA) published the long-awaited rule that contains the final regulations for many of the graduate medical education (GME) provisions included in the Balanced Budget Act of 1997 (BBA). See Medicare Program: Changes to the Hospital Inpatient Prospective Payment Systems and Fiscal Year 1998 Rates; Final Rule. 63 Fed. Reg. 26,317 (May 12, 1998). The May 12 Federal Register contains changes HCFA made to the interim final rule it published on August 29, 1997 (62 Fed. Reg. 52,034). In addition, the preamble of the regulations contains replies to the comments the Agency received in response to the August 29 regulations. If no changes were made to the interim final rule, the text of the actual regulations is published in the August 29 regulations. Only changes to the August 29 regulations are included in the May 12 final rule. Persons wishing to obtain copies of the regulations may contact the AAMC Division of Health Care Affairs at (202) 828-0490.

While many of the BBA provisions apply to both direct graduate medical education (DGME) payments and indirect medical education (IME) payments, for organizational purposes, this document summarizes those provisions first as they apply to DGME payments, then as they apply to IME payments. Because many of the GME regulations were discussed and published in the August 29, 1997 interim final rule, this document summarizes the August 29 regulations for each distinct GME provision, and then highlights any changes or clarifications included in the May 12 final rule.

You should be aware that the terms "affiliation," "affiliated group," and "affiliation agreement," are repeated throughout this document. This is because these terms are included in the BBA and corresponding regulations. It is important that you recognize that these terms have a specific meaning under the BBA and GME regulations which is related solely to implementing the resident limits. These terms should not be read to place any restrictions on other types of affiliations or affiliation agreements.

Direct Graduate Medical Education (DGME) Payments

A major provision of the BBA states that for cost reports beginning on or after October 1, 1997, DGME payments will be limited if the number of full-time equivalent (FTE) allopathic and osteopathic residents exceeds the number of residents reported on the hospital's cost report for the period ending on or before December 31, 1996 ("Base Year Limit"). The BBA also required HCFA to prescribe rules that would permit an adjustment to the limits for "residency training programs established on or after January 1, 1995." In establishing the rules for new programs, the BBA required HCFA to take into account "facilities that meet the needs of underserved rural areas." HCFA was also given permission, but was not required, to establish rules so that hospitals that choose to be part of an "affiliated group" may have their limits applied on an aggregate basis.

The BBA also provided that beginning with cost reports on or after October 1, 1997, DGME payments will be based on the average of the weighted number of residents in three years--the current cost reporting year and the two previous years--after accounting for the resident limit (this is known as a "rolling average" payment provision). For the first cost reporting period on or after October 1, 1997, the average will based on two years--the current and immediately preceding cost reporting years.

The regulations implementing the DGME provisions of the BBA can be divided into four components: resident limits, new programs, affiliated groups, and the three-year rolling average.

I. Resident Limits

August 29 Regulations

• The resident limit applies before application of the initial residency period weighting factors.
• Limit excludes dental and podiatry residents.
• Limit takes effect with cost reporting periods beginning on or after Oct. 1, 1997.
• Intermediaries will make adjustments for hospitals whose base years are shorter than 12 months.

May 12 Regulations

• Emphasized that resident limits would not be modified to accommodate expansions of existing programs, even if those expansions occurred before the BBA law was enacted but after the base year.
• Added an adjustment to resident limits for hospitals that took on additional residents between July, 1996 and June 1997 because another hospital closed. After these residents finish training, the limit will revert to the base year. This provision requires that the closed hospital not seek reimbursement for the residents.
• The limit for a hospital that results from a merger will be the combined base year limits of the hospitals participating in the merger.
• HCFA plans to use the intern and resident information system (IRIS) to obtain base year counts.

Analysis--In the May 12 Final Rule, HCFA acknowledged that some hospitals may have expanded programs since their base year but prior to passage of the BBA (generally 1996 through August 5, 1997). The Agency stated that the limits could not be increased since that would result in creating limits based on a year that is contrary to the base year specified in the BBA.

HCFA agreed with the AAMC and other commenters that it is appropriate for hospitals to receive Medicare teaching payments for residents they train temporarily because another hospital is closing even if those residents exceed the base year limit. This is necessary to allow those residents to finish their training. However, the preamble indicates that a temporary adjustment will be allowed only if this situation occurred during the June 1996-July 1997 period. As an initial matter, this temporary adjustment is not codified in the regulatory language. In addition, the AAMC believes the stated time frame is too narrow. The residency training issues related to hospital closings occur any time a hospital closes; the time frame when the closure occurs should be irrelevant. It also appears that HCFA neglected to modify the regulations to address the aggregation of resident limits under a merger, despite the discussion in the regulatory preamble.  AAMC staff plan to pursue these issues with HCFA officials.

COTH members should take special note that HCFA plans to use IRIS to obtain the base year counts. Currently, IRIS is used to identify duplicate resident counts. This is the first time it will be used for payment purposes.  Accordingly, the AAMC advises COTH members to review immediately their IRIS resident counts for their base year for accuracy.

II. New Residency Training Program Regulations

A. General Requirements

August 29 Regulations

• For purposes of adjusting the resident limits, "new" training programs are defined as those that received accreditation for the first time, including provisional accreditation, on or after January 1, 1995.

May 12 Regulations

• The new program definition includes both programs that received initial accreditation or began training residents on or after January 1, 1995.
• Resident limits will not be adjusted for training programs that began before January 1, 1995 but that did not reach their full complement as of their 1996 cost report count.

Analysis--HCFA adopted the AAMC's comment to broaden the definition of new programs so that hospitals may have their limits adjusted either if they received an accreditation letter or began training residents on or after January 1, 1995. This is especially important for urban hospitals because, as discussed next, the new program exception for these hospitals ends as of August 5, 1997 (the date the BBA was signed into law). Despite the AAMC's and others' comments, however, HCFA did not further broaden the definition of new programs to include the date when an accreditation application was submitted.

It is important to remember that resident limits will not be adjusted for expansions of existing training programs which result in resident counts that are higher than the base year limit. It is still possible for hospitals to receive reimbursement for these expansions, however, so long as residents in other programs are reduced such that the resident limit is not exceeded.

B.   Specific Rules for New Teaching Hospitals, Urban Teaching Hospitals and Rural Teaching Hospitals

The BBA required that in creating rules for new programs, special consideration should be given to "facilities that meet the needs of underserved rural areas."  HCFA implemented this provision by creating different rules for urban and rural teaching hospitals. HCFA also relied on this provision to establish rules for nonteaching hospitals that begin training residents. Consequently, the adjustments to resident limits for new programs differ depending upon whether, as of January 1, 1995, the hospital: a) had no residents ("new teaching hospital"), b) had residents and is located in an urban area ("urban teaching hospital"), or c) had residents and is in a rural area ("rural teaching hospital").

1) New Teaching Hospitals

August 29 Regulations

• The resident limit will be determined by the number of PGY-1 residents in the third year after the hospital begins training residents. It will be based only on the first program (or programs if established simultaneously) beginning on or after January 1, 1995.

May 12 Regulations

• The resident limit will be based on the highest number of residents in any program year for all programs that exist in the third year of training existence and the minimum accreditation length of each program, respectively.
• Existing training programs that are transferred to hospitals that previously had no training programs are not considered "new" for purposes of establishing resident limits.

Analysis--In response to the AAMC's comments that hospitals need time to establish complementary training programs, HCFA is allowing the resident limit for new teaching hospitals to reflect the resident count of all programs started within the first three years of training (rather than counting only those programs that began on the same day). The limit will be determined by the highest number of residents in any program year, rather than looking solely at the PGY-1 year.

Consequently, nonteaching hospitals can become teaching hospitals at any time and have three years to establish training programs before their resident limit is established.

While HCFA did not adopt commenters' suggestions to permit nonteaching hospitals to take on existing programs under the new program exception, it noted that these hospitals could train residents of existing programs by affiliating with existing teaching hospitals.

2) Urban Teaching Hospitals

August 29 Regulations

• Resident limits will be adjusted only for those new programs established between January 1, 1995 and August 5, 1997.
• The resident limit will be adjusted in the third year of the new program and will reflect the number of PGY-1 residents multiplied by the number of training years required for the initial residency period.

May 12 Regulations

• Changed the limit adjustment to reflect the highest number of residents in any program year in the third year of the new program's existence.
• Clarified that to the extent the residents are not included in the 1996 base year limit, the limit will be adjusted incrementally, until it reflects the product of a full complement of residents and the minimum accreditation length.

Analysis--It must be emphasized that urban teaching hospitals will not have their resident limits adjusted for new programs established after August 5, 1997. And while urban non-teaching hospitals that begin training residents have three years to establish a resident count (see new hospital discussion, above), after that period, no adjustments will be made for new programs. HCFA rejected comments suggesting that urban teaching hospitals be permitted to add new programs after August 5, 1997 if the programs are in primary care. HCFA adopted the AAMC's suggestion to determine the upward adjustment for new programs based on the program year which has the highest number of residents (PGY-1, 2, or 3) in the third year of the program's existence, rather than looking solely at the number in the PGY-1 year.

3) Rural Teaching Hospitals

August 29 Regulations

• Resident limits are adjusted for all new programs, even those established after August 5, 1997.
• Resident limit is adjusted upward based on the number of PGY-1 residents in the new program's third year of existence.

May 12 Regulations

• Changed the limit adjustment to reflect the highest number of residents in any program year in the third year of the new program's existence.
• Clarified that to the extent the residents are not included in the 1996 base year limit, the limit will be adjusted incrementally, until it reflects the product of a full complement of residents and the minimum accreditation length.

Analysis--Allowing the limit for rural teaching hospitals to increase for any and all new programs into perpetuity is HCFA's response to implementing the BBA provision requiring the Secretary to take into account facilities that meet the needs of underserved rural areas. HCFA rejected comments to broaden its interpretation of this BBA provision to permit resident limit adjustments for urban program expansions in rural areas, or urban hospital programs that include training rotations in rural areas.

Like the provision for urban hospitals, HCFA adopted the AAMC's suggestion to determine the upward adjustment for new programs based on the program year which has the highest number of residents (PGY-1, 2, or 3)  in the third year of the program's existence, rather than looking solely at the number in the PGY-1 year.

III. Aggregate Limits for Affiliated Groups

The BBA permits, but does not require, HCFA to establish rules for hospitals that choose to apply their resident limits on an aggregate basis. This type of provision permits hospitals flexibility in structuring rotations within a combined cap when they share residents.

1. Hospital-Level Affiliations

August 29 Regulations

• Two or more hospitals in the same urban area that rotate residents during the course of the training program may affiliate.
• Two or more hospitals in the same rural area that rotate residents during the course of the training program may affiliate.
• Two or more hospitals not located in the same geographic area but listed in common as "major participating institutions" in the Graduate Medical Education Directory for one or more programs may affiliate.

May 12 Regulations

• Broadened the geographic area criteria so that urban hospitals may affiliate with hospitals in contiguous urban areas. Rural hospitals may also affiliate with hospitals in contiguous areas.
• Clarifies that an urban area is a metropolitan statistical area (MSA) or New England County Metropolitan Area (NECMA); a rural area is any area outside an urban area.
• Clarifies that hospitals not located in the same or contiguous areas, but that are jointly listed for one or more training programs in the GME Directory as a sponsor, primary clinical site, or major participating institution, may affiliate.
• Emphasizes that VA and DOD facilities cannot be part of affiliated groups for purposes of aggregating resident limits.
• Permits resident affiliations of hospitals under common ownership, regardless of geographic boundaries or rotation relationships.
• Creates special rules for hospitals that establish residency programs for the first time after August 5, 1997:
• Urban hospitals that begin training residents after August 5, 1997 are not permitted to affiliate with other urban teaching hospitals.
• Rural hospitals that begin training residents may affiliate with existing urban hospitals only if the rural hospital provides training for the FTE equivalent of at least one-third of the residents participating in the joint programs of the affiliated hospitals.

Analysis--The final rule adopted the AAMC's comment and broadened the geographic criteria for affiliated groups based on geographic proximity to include contiguous areas, athough it is not completely clear whether urban hospitals may aggregate their resident limits with hospitals located in statewide rural areas if they are contiguous. HCFA did not adopt suggestions to permit VA and DOD hospitals to participate in affiliations. The Agency stated that VA and DOD institutions do not participate in Medicare, so that HCFA is unable to monitor their resident counts and, therefore, cannot establish an aggregate limit for groups which include these governmental hospitals.

The special rules that HCFA discusses for affiliations between hospitals that begin training residents after August 5, 1997 and existing teaching hospitals were not part of the August 29 regulations. (It also does not appear that these rules were codified in the regulatory language.) HCFA stated in the May 12 final rule that it established the rule prohibiting these new hospitals from affiliating with existing urban teaching hospitals because it did not want an existing urban teaching hospital to be able to increase its resident limits de facto after August 5, 1997 by encouraging a nonteaching hospital to establish a new training program, affiliate with it, move the residents to its own hospital, and receive an upward adjustment to its resident limit pursuant to the affiliation agreement. The AAMC is concerned that this provision disregards new urban teaching hospitals that have a legitimate need to rotate residents to other hospitals. This will be constrained if they are not permitted to establish affiliation agreements with nearby urban teaching hospitals. HCFA stated that the one-third training requirement for affiliations between new rural teaching hospitals and existing urban teaching hospitals was intended to recognize that rural hospitals may not have the sufficient patient care utilization to meet accreditation standards. It is unclear whether this one-third training requirement applies to all joint programs or just new joint programs.

2. Affiliations at the Program Level

August 29 Regulations

• No provisions for affiliations at the program level.

May 12 Regulations

• Permits the affiliated group provisions to be used to address program-to-program rotations.
• Affiliation agreements must indicate resident limit changes effective July 1 of each year, generally the start date of rotation years.
• For Medicare cost reporting periods that do not begin on July 1, the resident adjustments indicated in the agreements will be used to adjust payments for resident counts provided on the hospital's cost report.
• Affiliation agreements must be given to all intermediaries in the case of hospitals that are part of affiliation agreements and have different intermediaries.

Analysis--The August 29 interim final rule did not include a provision to address program level affiliations. In its comments, the AAMC urged HCFA to allow this type of affiliation because hospitals may have rotation relationships with a number of hospitals, some of which do not have rotation relationships with each other. HCFA agreed and the May 12 provision is the Agency's effort to accommodate affiliations at the program level.

Understanding how program affiliations will work under the regulations is not easy. For ease of explanation, it may be best to begin with the situation of a hospital that has program rotations with only one other hospital. If these two hospitals meet the relevant criteria (for example, geographic proximity) they can form an affiliated group, and the redistribution of the aggregated cap among the two hospitals is relatively straightforward, reflecting the relevant program rotations of each of the hospitals.

If, however, a hospital's resident count fluctuates because it rotates residents with multiple hospitals, the affiliation group must include all of the involved hospitals, even if the other hospitals have no rotation relationship with each other. This is necessary because HCFA must verify that a hospital is not seeking reimbursement for a resident count that was not part of a base year limit.

Example: Hospital A rotates residents with Hospital B, and Hospital B rotates residents with Hospital C, but Hospitals A and C have no rotation relationship. All three hospitals have a base year limit of 100 residents.

In FY 1998, 10 of Hospital B's residents train at Hospital A, and 5 of Hospital C's residents train at Hospital B. Consequently, Hospital A will have 110 residents, Hospital B will have 95 and Hospital C will have 95. The total equals the aggregate count for all three hospitals -- 300 residents.

Example Analysis--In order for Hospital A to be reimbursed for the 10 additional residents, it must show an affiliation agreement with Hospital B. If this were the only agreement provided to HCFA, however, it would appear that only 5 of Hospital B's residents trained at Hospital A (because Hospital B's resident count is 95 rather than 100), so that Hospital A would be limited to reimbursement of 105 rather than 110 or, alternatively, Hospital B's reimbursement would be limited to 90 residents rather than the 95 it actually has.

In order for HCFA to understand that Hospital A's resident count is not reflecting new residents, but rather is within an aggregate limit, Hospital C's resident count must be included (as well as the agreement between hospitals B and C), even though Hospitals A and C might not otherwise meet the hospital-level criteria for becoming an affiliated group. When all three hospitals' resident counts are viewed simultaneously, it demonstrates that none of the hospitals is seeking reimbursement for a resident count that is outside the base year limits.

The May 12 final rule also includes the following table which illustrates the same example of a program affiliation group and its relationship to changes to cost reports that do not have a July 1 start date:

Source: 56 Fed. Reg. at 26339.

3. Distributing the Aggregate Resident Limit Among Affiliated Group Members

August 29 Regulations

• No explicit provisions

May 12 Regulations

• The aggregated resident limit for the affiliated group is the sum of the individual resident limits for all hospitals that are part of the group.
• Affiliated hospitals can redistribute the aggregate limit among themselves, so long as the aggregate limit is not exceeded. DGME payments to the individual hospitals will be based on the hospital specific resident counts determined under the affiliation agreements.
• If a hospital's resident count exceeds the limit agreed to in the affiliation agreement, Medicare DGME payments will be reduced proportionately (see next section on payments under the rolling average).
• Hospitals must provide HCFA and the relevant Medicare intermediaries with a minimum l-year agreement (it may be more than one year), signed by all hospital participants that:
• informs intermediaries and HCFA by July 1 of each year (the annual start date for most residency programs) the planned changes to each individual hospital's resident limits for the current or subsequent residency training years,
• specifies how each hospital's base year limit should be adjusted at the end of the agreement, or in the event the agreement dissolves,
• specifies that any upward adjustment for one hospital's limit is offset by an equal downward adjustment by the other hospital(s).

Analysis--The advantage of being in an affiliated group is that a hospital's resident limit can change, subject to the agreement of the other hospitals in the group. Note that if the affiliation agreement does not indicate how the resident limits will be set when an agreement ends, each hospital's limit will revert to its 1996 base year limit.

Hospitals that wish to pursue program level affiliations must jump through another hoop in addition to entering into signed agreements with each hospital where they are rotating residents. These hospitals involved must submit to HCFA and the appropriate intermediaries their own agreements as well as any agreements entered into by their affiliate hospital partners with other hospitals. They do not have to be a party to these other agreements. The submission of all of the agreements is necessary, however, so that HCFA and the relevant fiscal intermediaries can verify that the aggregate resident limits are not being exceeded.

To illustrate using the example described above under program affiliations: Hospital A must submit to its intermediary two agreements--its agreement with Hospital B and Hospital B's agreement with Hospital C, even though there is no rotation relationship, and hence no agreement, between Hospitals A and C. This is necessary to verify that Hospital A's resident count does not inappropriately exceed its resident limit. Hospitals B and C must submit the same information. If all three hospitals are serviced by the same fiscal intermediary, the submission is fairly straightforward. If one or more of the hospitals are serviced by a different fiscal intermediary, however, the two agreements must be submitted to both fiscal intermediaries.

IV. Three Year Rolling Average Resident Count for Computing DGME Payments

August 29 Regulations

• DGME payments will be based on the average of the weighted number of residents in the current and two preceding cost reporting years, subject to the impact of the resident limit. For the first  cost reporting period on or after October 1, 1997, the average will based on two years--the current and immediately preceding cost report.
• If the unweighted resident count is over the resident limit, the weighted count will be reduced proportionately.

May 12 Regulations

• Dental and podiatry residents are included in the three-year rolling average calculations even though they are excluded from the resident limits.
• Additional residents due to new programs will be added to the resident count after the three year rolling average is computed.

Analysis--The effect of the rolling average is to reduce the financial impact on hospitals that reduce their residency counts because a hospital's payment in a current year will reflect higher resident counts during previous years. For hospitals that have unweighted resident counts above the limit, the weighted count that is part of the  rolling average will be reduced proportionately. The Final Rule contains an appendix which provides detailed examples of computing DGME payments using the three-year rolling average and reflecting the resident limits. This appendix is reproduced at the end of this document. Even though dental and podiatry residents are excluded from the resident limits, they are not excluded from the rolling average calculations. To the extent a hospital's resident count is higher in the current year due to additional dental or podiatry residents, DGME payments will be lower because of the averaging.

The May 12 rule also modified the rolling average regulations so that hospitals with higher resident counts due to new programs are not penalized. In these situations, the rolling average resident count will first be computed excluding the new program residents. These residents will then be added to the rolling average count to determine the ultimate resident count used for payment purposes. New program residents will not be excluded from the rolling average calculations forever, however. They will be included in the calculations once the minimum length of time for the initial residency period has passed.

It should be noted that the same modification does not apply for hospitals that have higher resident counts because of affiliations. In this case, the additional residents are included in the rolling average calculations. Hospitals are disadvantaged under this scenario because rather than being paid based solely on a higher current year count, the rolling average calculation will reflect lower resident counts in the two previous years.

Indirect Medical Education Payments (IME) Payments

The BBA made several changes to IME payment policies, relating to payments for residents in nonhospital settings, limits on resident counts and intern-resident to bed (IRB) ratios, and reductions to the IME payment formula. The resident limits and new program and affiliation provisions are similar to the provisions applicable to DGME payments.

I. IME Payments for Residents in Nonhospital Provider Settings

The BBA provides that hospitals may receive IME payments for residents training in nonhospital sites so long as they incur all or substantially all of the costs for the training program in that setting.

August 29 Regulations

• Beginning with October 1, 1997 discharges, teaching hospitals can now receive IME payments for residents training in nonhospital sites subject to the IME resident limit (see below) so long as they incur "all or substantially all of the costs of training."

May 12 Regulations

• Emphasized that the base year IME resident limits do not include residents training in nonhospital sites (see discussion below under resident limits).

Analysis--HCFA states that it will rely on the DGME regulations to identify the nonhospital sites eligible for IME payment. The DGME regulations do not definitively define nonhospital sites, but indicate examples of such sites to include freestanding clinics, nursing homes, and physician offices (see 42 C.F.R. §413.86(f)(iii)). In addition, hospitals currently are deemed to incur "all or substantially all of the costs of  training" if they pay the resident's stipends and benefits. While this is currently the law, it is important to note that the fiscal year 1999 hospital inpatient PPS proposed rule (published May 8, 1999) proposes to broaden this definition to including supervisory physician costs. (See 63 Fed. Reg. at 25,597). The AAMC will be submitting comments to HCFA on this proposal.

II. Limits on Resident Counts and Intern-Resident to Bed (IRB) Ratios

IME payments are determined based on the ratio of interns and residents in a hospital to the number of staffed beds (IRB ratio). The higher the IRB ratio, the more IME payments a hospital receives. The BBA limits IME payments by providing for two limits which affect the IRB ratio: a) a limit on residents--which affects the numerator of the ratio, and b) a limit on the ratio itself.

A.IME Resident Limits

August 29 Regulations

• The IME resident limit is established based on the number of allopathic and osteopathic residents on a hospital's cost report for the period ending on or before December 31, 1996 that were counted for purposes of IME payments.
• Like DGME, dental and podiatry residents are excluded from the IME resident limit.
• Unlike DGME, the IME resident limit does not include residents working in nonhospital sites during the 1996 cost report period because hospitals could not count these residents towards IME payments in 1996.
• The IME resident limits are effective with discharges occurring on or after October 1, 1997.

May 12 Regulations

• Clarified that IME resident limit also applies to capital IME payments.
• Clarified that resident limit can be modified to reflect additional residents permitted under new program and affiliated group exceptions.

Analysis--The resident limit affects IME payments because it impacts the numerator of the IRB ratio (i.e, the hospital's resident count). To the extent the numerator is constrained and the denominator (staffed beds) is unchanged, the IRB limit cannot increase.

It also should be noted that if a hospital had residents training in nonhospital sites in its base year, its IME resident limit and DGME limit will be different. This is because during the base year (generally 1996) hospitals could receive DGME payments for residents in nonhospitals, but could not receive IME payments. Thus, the number of reimbursable residents in that year was higher for DGME than for IME.

B. IRB Ratio Limit

August 29 Regulations

• IRB ratio is limited to its computed value in the hospital's most recent prior cost reporting year (after accounting for the limit on the number of allopathic and osteopathic residents).

May 12 Regulations

• Emphasized that the IRB limit reflects all residents, including dental and podiatry residents.
• The IRB limit will be modified to reflect additional residents permitted under new program and affiliated group exceptions to IME resident limit (see discussion, below, on new programs and affiliated groups).
• Clarified that the IRB limit does not affect capital payments because capital payments are based on the ratio of hospital FTE's to average daily census.
• The calculated IRB ratio must reflect resident limits as well as the three-year rolling average resident count (see discussion below).

Analysis--The BBA imposed a limit on the IRB ratio in addition to a limit on residents. Without an IRB limit, hospitals could continue to increase their IME payments, even with a resident limit, by reducing their number of staffed beds. This is because staffed beds are reflected in the denominator of the IRB ratio; therefore, reducing this number will result in a higher ratio, which will trigger higher IME payments.

Because it is based on the prior cost reporting year's ratio, the IRB ratio limit is essentially a one-year lagged cap. Thus, reductions in the number of beds which increase the ratio will not be recognized in the current year, but will be recognized in the following year. The May 12 final rule also stated that HCFA considers dental and podiatry residents to be included under the IRB cap, even though they are excluded from the resident limit. Consequently, if the numbers of these residents increase in the current year, the computed IRB ratio would increase, but the additional IME payments would not occur until the following cost report year. By contrast, since there is a limit on number of allopathic and osteopathic residents, increasing their numbers would not result in an increase of the IRB ratio. (See payment example below.)

III. Exceptions to Resident Limits: New Programs and Affiliated Groups

August 29 Interim Final Rule 

• These provisions were not included or referenced in the August 29 regulatory language.

May 12 Final Rule

• Clarified that the new program and affiliated group exceptions to limits apply to IME payments and included the regulatory references.
• IRB ratio limit will be adjusted to reflect additional residents associated with new programs or resulting from resident redistributions under affiliation agreements.
• IRB ratios, in and of themselves, cannot be aggregated.

Analysis--If a hospital's current IRB ratio is greater than its ratio in the prior cost reporting year due to additional residents resulting from permitted new programs or resident affiliation agreements, the prior cost reporting year's ratio will be adjusted to reflect these additional residents. This is computed by adding the additional residents to the resident FTE count used in the prior cost reporting period's IRB ratio. The resultant higher ratio limit will enable these residents to be reimbursed in the current year. In contrast, the prior IRB ratio will not be adjusted to reflect current year increases in dental or podiatry residents since these residents are included in the count making up the IRB limit. These residents, however, will count towards IME payments in the following year when the IRB limit is higher because of the presence of the additional dental or podiatry residents.

It is important to note that the May 12 rule states that the affiliation provision in the BBA refers only to the IME resident limit, not to the IRB ratio. In other words, if two hospitals with different IRB ratios affiliate, they cannot agree to redistribute their IRB ratios. Their respective IRB ratios can change only to the extent that the two hospitals redistribute residents such that the numerator of the IRB ratio (which reflects the resident count) is different compared to the base year.

IV.IME Payments and Three-Year Rolling Average

August 29 Regulations

• The number of residents used to determine IME payments will be based on a three-year rolling average resident count, after taking into account the effect of the resident limit.
• The rolling average provision takes effect with cost reporting periods beginning on or after October 1, 1997 (compared to the resident and IRB limits which take effect with discharges on or after October 1, 1997).

May 12 Regulations

• Clarified that the rolling average applies to capital IME payments.
• Dental and Podiatry residents are included in the three-year rolling average calculation even though they are excludedn from the resident limits.

Analysis--While the IRB ratio used to determine IME payments in a particular year can be no higher than the IRB cap in the previous year, the numerator of the ratio (i.e., number of  residents) will be determined based on a three-year rolling average count of allopathic and osteopathic residents AND the number of dental and podiatry residents (two year average in FY 1998). The allopathic and osteopathic counts used in each of the years contributing to the average are capped by the resident limit number. The denominator will be the number of staff beds in the current year. If this ratio is higher than the previous year's ratio, IME payments would be based on the prior year ratio.

Example: Hospital A has a resident limit of 90 residents. Last year, it had 90 allopathic and osteopathic ("AO") residents, 10 dental residents, and 300 beds.  Accordingly, its IRB ratio was 0.33 (100/300).

A. In FY 1998 hospital A reduced its AO count to 80: The numerator of the IRB ratio would be the average of the AO and dental counts for both years, (100 + 90) ÷ 2 which would equal 95. Its calculated IRB ratio would be 95/300 or 0.32. Since this ratio is less than last year's ratio of 0.33, IME payments would be based on the IRB ratio of 0.32. Note that without the rolling average, the IRB ratio would be 0.30

B. In FY 1998, Hospital A increased its AO count to 95:  For purposes of the rolling average, the 95 resident count would be capped at 90. Therefore, the numerator of the IRB ratio would be the average of (100+100) ÷ 2, or 100 (90 AO count + 10 dental residents). Its calculated IRB ratio would be 100/300 or 0.33. Since this ratio is the same as last year's ratio of 0.33, IME payments would be based on the IRB ratio of 0.33.

C. In FY 1998, Hospital A reduced its AO count to 80, and reduced its staffed beds to 250. The calculated IRB ratio would be the rolling average count of 95, divided by 250 beds which equals 0.38. Since this ratio is above last year's ratio of 0.33, IME payments would be based on an IRB ratio of 0.33. But next year, the IRB cap will be this year's calculated IRB ratio, or 0.38.

V. Reductions to the IME Payment Formula

In fiscal year 1997, teaching hospitals were paid an additional 7.7 percent for every 10 percent increment in a hospital's resident-to-bed ratio. The BBA reduced this level over a four year period.

August 29 Regulations

May 12 Regulations

• No changes

Analysis--To calculate the hospital's specific IME percentage add-on adjustment at the new level, replace 1.89 used in the IME formula for FY 1997 with the multipliers shown in the table. The remainder of the IME formula is unchanged.

The Appendix, published in the Federal Register, is available Portable Document Format (PDF).

For more information contact:

Karen Fisher, Senior Associate Vice President
AAMC Health Care Affairs
kfisher@aamc.org
(202) 862-6140

This page contains documents in Portable Document Format (PDF). The Adobe Acrobat® Reader® is required to view PDF documents. Download Acrobat® Reader®.