[Federal Register: August 1, 2003 (Volume 68, Number 148)]
[Rules and Regulations]               
[Page 45445-45494]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr01au03-11]                         
 
[[pp. 45445-45494]] Medicare Program; Changes to the Hospital Inpatient Prospective 
Payment Systems and Fiscal Year 2004 Rates

[[Continued from page 45444]]

[[Page 45445]]

inappropriate counting), or to residents training inside the hospital--
inpatient or outpatient. Thus, it is technically possible to have a 
redistribution of direct GME costs for the training of residents inside 
the hospital setting (as well as in the nonhospital setting). 
Therefore, we are not adopting the commenter's suggestion to limit 
application of the principles to Sec.  413.86(f)(4) (the nonhospital 
site provision). However, we note that we believe a redistribution of 
all of the direct GME costs for training that occurs in a hospital 
setting would be rare. All of the direct costs of the program--resident 
salaries, teaching physician salaries, overhead expenses, etc., would 
need to be redistributed to an outside entity in order for there to be 
a disallowance of direct GME FTE residents for training inside the 
hospital due to redistribution of costs or community support.
    We contrast this application of the principles of redistribution of 
costs and community support in the current prospective payment system 
that depends upon PRA and FTE resident counts to application of the 
principles in the previous reasonable cost payment methodology that was 
based on cost finding and cost allocations. Under the former reasonable 
cost methodology, a hospital was eligible to receive direct GME payment 
for those direct GME costs that it incurred; however, any direct GME 
costs that were redistributed to the hospital were not allowable. We 
note that the instructions that CMS (then HCFA) gave to its Regional 
Offices in the 1990 audit instructions for purposes of calculating the 
direct GME base period PRA specifically addressed redistribution of 
costs and community support in the GME context:

    Where costs for services related to medical education activities 
have historically been borne by the university, it is assumed the 
community has undertaken to support these activities, and subsequent 
allocation of these costs to a hospital constitutes a redistribution 
of costs from an educational institution to a patient care 
institution. In such a situation, these costs are not allowable 
under the Medicare program. (See 42 CFR 413.85(c) and HCFA Pub. 15-
1, Sec.  406). For example, if in the past the hospital did not 
identify and claim costs attributable to the time teaching 
physicians spent supervising I&Rs [interns and residents] working at 
the hospital, it is assumed that these costs were borne by the 
university. Therefore, the hospital may not claim these costs in 
subsequent cost reports. (Instructions for Implementing Program 
Payments for Graduate Medical Education to ARAs for Medicare, 
Director of Office of Financial Operations of the Health Care 
Financing Administration, BPO-F12, February 12, 1990.)

    Thus, under the previous cost payment scheme, the principles of 
redistribution of costs and community support were applied to direct 
GME reasonable cost payment using a cost finding methodology. In 
contrast, in the current context where payment is no longer based 
solely on reasonable costs incurred, but on PRA and FTE resident 
counts, if the hospital can demonstrate that it has continuously 
incurred some of the direct GME costs of training the residents since 
the inception of the residency program at a training site, then no 
redistribution of costs or community support has taken place. As noted, 
current direct GME payments are no longer based on detailed cost 
finding of allowable costs of hospitals. Therefore, we believe it is 
appropriate to require that a hospital demonstrate that there has been 
no redistribution of costs or community support by proving that the 
hospital has incurred some of the direct GME costs of the program 
continuously since the inception of the program. Finally, contrary to 
the commenter's assertion, we believe we have been consistent with the 
other Medicare policies on counting residents, including the policy 
cited by the commenter concerning the prohibition on counting residents 
training at other hospitals. (See the August 1, 2002 final rule (67 FR 
60077). As stated above, there would be no redistribution of costs or 
community support if a hospital counts a resident when another hospital 
incurs the resident's salary, as long as the first hospital still 
incurs other direct GME costs associated with the training of that 
resident. In any case, as we explained above and also in the proposed 
rule, the principles of redistribution of costs and community support 
are not applicable to cost shifted between the hospitals, only costs 
shifted between a hospital and educational institutions or other 
organizations that are not Medicare providers.
    Comment: One commenter stated that a hospital was ``required'' to 
include in the calculation of its average per resident amount, time 
spent in the hospital by residents who were paid by ``other entities.'' 
This commenter quoted the September 29, 1989 final rule: ``the 1989 GME 
rule was modified after publication of the proposed rule in order `to 
require Medicare hospitals to count residents who are working in their 
facility even if the residents' salaries are fully paid by other 
entities, either Federal or non-Federal. This revised policy will apply 
to both GME base period and cost reporting periods subject to the new 
payment methodology.' 54 FR 40299 (emphasis added).''
    Response: We believe the language quoted above by the commenter 
from the 1989 final rule has been taken out of context. In essence, the 
commenter has generalized from the language selectively quoted above to 
support an argument that Medicare would have required a hospital to 
count resident time when the residents were ``paid by other entities,'' 
thereby supporting the commenter's argument that Medicare not only 
condones redistribution of costs but, in fact, would seem to 
``require'' them. However, we believe the language quoted by the 
commenter from a particular comment and response in the 1989 rule, if 
quoted in its full context, actually supports the CMS policy on the 
application of the principles of redistribution of costs and community 
support that as long as the hospital has continuously incurred at least 
some of the direct GME cost of the residency program since the 
inception of the program, there has been no redistribution of costs or 
community support and the hospital may count the FTE residents. 
Specifically, the commenters in that rule at 54 FR 40298 asked in 
relevant part: ``A particular problem referred to was the treatment of 
residents who are paid by medical schools, faculty practice plans, and 
others rather than by hospitals that participate in Medicare. It was 
pointed out that teaching hospitals incur other costs such as teaching 
physicians' salaries and overhead costs in connection with these 
residents, and it would be unfair not to count these residents for 
payment purposes.'' In our response to this comment, we stated, also in 
relevant part on 54 FR 40299: ``we note that some of the comments have 
led us to believe that, in addition to Federally-employed residents 
(for example, residents in Veterans Administration or Department of 
Defense programs), a significant number of residents are paid a salary 
by non-Federal, nonprovider entities (for example, medical schools or 
philanthropic agencies). As noted by the commenters, although no 
hospital participating in Medicare incurs salary costs for these 
residents, hospitals do incur other substantial GME costs associated 
with these residents. Therefore, we are modifying our proposed rule to 
require Medicare hospitals to count residents who are working in their 
facility even if the residents' salaries are fully paid by other 
entities, either Federal or nonfederal.'' (Emphasis added). It becomes 
apparent when the language quoted by the

[[Page 45446]]

commenter on this final rule is read in context that, even as far as 
back as the 1989 final rule, we acknowledged that hospitals may count 
the FTE residents where other entities may have incurred the residents' 
salaries, but where the hospitals still ``incur other substantial GME 
costs associated with these residents.'' This view is entirely 
consistent with the CMS application of redistribution of costs and 
community support. In a scenario where a nonhospital entity, such as a 
medical school, incurs the residents' salaries, we continue to believe 
that the hospital may count the FTE residents if the hospital can 
demonstrate that it has incurred other direct GME costs, such as the 
supervisory physician salaries, since the inception of the program.
    Comment: One commenter argued that when we explained our policy in 
the July 31, 1998 Federal Register (63 FR 40954) to require a written 
agreement indicating that the hospital must provide reasonable 
compensation for physicians' supervision of residents' training in the 
nonhospital setting, ``nothing was said about an additional requirement 
that a hospital must have continuously incurred this additional cost, 
as well as the residents' compensation required under the prior 
regulations, since the inception of the training program.'' This 
commenter further makes the point that in the final rule at 63 FR 
40986, in response to a comment 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, CMS responded:

    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 arrangements 
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 that 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 the costs for that resident may be lower during the time 
when 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 the nonhospital site and is consistent with 
the statutory requirement that the hospital must incur ``all or 
substantially all'' of the costs.

(63 FR 40995 (emphasis added by commenter).)
    The commenter believed that this explanation of the changes to the 
GME and IME rules, effective January 1, 1999, ``belies CMS' current 
assertion of a longstanding policy of applying the redistribution of 
costs and community support principles in the determination of the 
resident counts used to compute payment for GME and IME.''
    Response: The commenter has used the language quoted above from the 
1998 final rule to argue that when CMS (then HCFA) described the policy 
on counting residents in nonhospital sites for IME, ``nothing was said 
about an additional requirement that a hospital must have continuously 
incurred this additional cost * * * since the inception of the training 
program.'' The commenter has inferred from the language quoted above 
that CMS has not had a longstanding policy of applying the 
redistribution of costs and community support principles. However, we 
believe the language actually supports the longstanding existence of 
our policy in two ways. First, the quoted language demonstrates the 
agency's view that the nonhospital site policy was written from the 
standpoint of addressing the counting of residents when hospitals 
rotate residents from the hospital to the nonhospital site. Second, the 
quoted language is also indicative of the Agency's policy that as long 
as the hospital has continuously incurred at least some of the direct 
GME cost of the residency program since the inception of the program, 
there has been no redistribution of costs or community support and the 
hospital may count the FTE residents (assuming that other requirements 
are met).
    Specifically, the comment relating to the portion of the 1998 final 
rule quoted above stated at 63 FR 40994, in relevant part: ``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.'' (Emphasis 
added.) In addition, as quoted above in the comment, we responded, in 
relevant part at 63 FR 40995 (with different emphasis):

    * * * 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 arrangements 
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 that 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 the costs for that resident may be lower during the time 
when 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 the nonhospital site and is consistent with 
the statutory requirement that the hospital must incur ``all or 
substantially all'' of the costs. Ibid.

    We believe the quoted comment and response from the 1998 rule paint 
a picture of a hospital that has had a pre-existing relationship with a 
nonhospital site involving rotation of residents from the hospital to 
the nonhospital site for a period of time during the residency program. 
The language we emphasized in the response--that the hospital may 
``continue to count residents'' when they train in the nonhospital 
sites, and that the hospital ``may count the resident even though the 
costs for the resident may be lower during the time when the resident 
trains outside the hospital''--clearly refers to a rotational 
arrangement between the hospital and the nonhospital site. In addition, 
according to the circumstances described by the commenter in the 1998 
rule, the hospitals had been incurring the residents' salaries, a 
direct GME cost, because they had formerly complied with the earlier 
regulation requiring that hospitals incur residents' salaries for 
purposes of meeting ``all or substantially all of the costs'' under 
Sec.  413.86(f)(3). We had no reason to believe that the hospitals had 
not incurred at least the residents' salaries since the inception of 
the training program (the commenters state that the arrangements 
``predate the GME base year''). In that event, the counting of 
residents in the nonhospital sites would not result in a redistribution 
of costs if, as of January 1, 1999, the hospital was required to incur 
the additional direct GME cost for supervisory physician costs while 
the residents rotate to the

[[Page 45447]]

nonhospital site. We believe that the commenter in the 1998 rule simply 
did not agree with the additional regulatory requirement finalized in 
the 1998 final rule that the hospital must also incur the supervisory 
physician costs for purposes of incurring ``all or substantially all of 
the costs,'' and hoped to label this new regulatory requirement as a 
``cost shift'' in order to avoid it. As we have explained, it appears 
that there has been no redistribution in the case described by the 1998 
final rule commenter because it can be inferred that the hospital had 
incurred at least some of the direct GME costs (the residents' 
salaries) since the inception of the program.
    Therefore, we believe the language the commenter quotes from the 
1998 rule is consistent with our clarifications in this final rule on 
redistribution of costs and community support. In addition, the 
language cited by the commenter supports our interpretation of the 
policy on counting residents in nonhospital sites that it was intended 
to address the situation when hospitals rotate residents from the 
hospital to the nonhospital site.
    Comment: Some commenters disputed the CMS interpretation of 
Congressional intent as discussed in the preamble of the proposed rule 
(see 68 FR 27213). One commenter stated: ``there is no support in the 
legislative history of the non-provider setting amendments [the 1986 
and 1997 amendments of the Act] for the Secretary's view that these 
changes were not intended to shift new costs to hospitals in support of 
on-going training in non-provider settings * * * it can be reasonably 
inferred that Congress was aware, and even intended, that some costs of 
existing residency training programs in non-provider settings would be 
shifted to hospitals in order for the hospitals to qualify for direct 
GME and IME funding under the 1986 and 1997 amendments of the Act.'' 
Similarly, another commenter stated that the Secretary ``must look 
elsewhere to the statute [other than section 1886(h)(4) of the Act] for 
support for his proposed rule; he cannot simply create out of whole 
cloth an interpretation that is inconsistent with the amendment's other 
provisions.''
    Response: The commenters would have us interpret and implement 
policy in a statutory vacuum. We believe we have reasonably discerned 
Congressional intent by interpreting the plain language of the statute 
at sections 1886(d)(5)(B) and 1886(h) of the Act in conjunction with 
the accompanying legislative history of these sections.
    As we stated in the preamble to the proposed rule, Congress has 
delegated broad authority to the Secretary to implement a policy on the 
count of FTE residents for purposes of calculating direct GME and IME 
payments. In section 1886(d)(5)(B) of the Act (IME), the statute does 
not specify at all how FTE counts should be determined, and the plain 
language in the statute under section 1886 (h)(4) of the Act (direct 
GME) indicates that the Secretary ``shall establish rules'' for direct 
GME consistent with the statute. We also considered the deference 
expressed in the conference agreement that accompanied Pub. L. 105-33, 
which established a cap on the number of allopathic and osteopathic 
residents a hospital may count--``[T]he Conferees recognize that such 
limits raise complex issues, and provide for specific authority for the 
Secretary to promulgate regulations to address the implementation of 
this provision.''(H.R. Conf. Rep. No. 105-217, 105th Cong., 1st Sess., 
821 (1997).
    Thus, in the absence of statutory specificity on determining FTE 
counts and the declared Congressional delegation of authority to the 
Secretary on the subject are clear indications that Congress has given 
the Secretary broad discretion to promulgate reasonable regulations in 
order to implement the policy on the counting of residents for direct 
GME and IME payments.
    In addition, we have not, as the second commenter suggests, 
``created out of whole cloth'' an interpretation of the policy 
concerning counting residents in nonhospital settings that is 
``inconsistent with the amendment's other provisions,'' nor do we at 
all believe that ``it can be reasonably inferred that Congress was 
aware, and even intended, that some costs of existing residency 
training programs in non-provider settings would be shifted to 
hospitals in order for the hospitals to qualify for direct GME and IME 
funding under the 1986 and 1997 amendments of the Act,'' as the first 
commenter suggests. Rather, as we have stated, we believe that when 
Congress created the provisions on counting resident FTEs in 
nonhospital settings, it was creating a monetary incentive for 
hospitals to rotate residents from the hospital to nonhospital 
settings. We have drawn this conclusion, as we explained, from the 
legislative history of both the direct GME and IME provisions 
authorizing payments to hospitals for training in nonhospital settings. 
First, legislative history associated with passage of the direct GME 
provision (as part of Pub. L. 99-509) indicates that Congress intended 
to broaden the scope of settings in which a hospital could train its 
residents and still receive separate direct GME cost reimbursement, and 
to provide incentives to hospitals for training residents in primary 
care programs. The Conference committee report indicates that ``[s]ince 
it is difficult to find sufficient other sources of funding [than 
hospitals and Medicare] for the costs of such training, [that is, 
training in freestanding primary care settings such as family practice 
clinics or ambulatory surgery centers] assignments to these settings 
are discouraged. It is the Committee's view that training in these 
settings is desirable, because of the growing trend to treat more 
patients out of the inpatient hospital setting and because of the 
encouragement it gives to primary care.'' (Emphasis added.) (H.R. Rep. 
No. 99-727, 99th Cong., 1st Sess., 70 (1986).)
    Thus, from the inception of the policy allowing payment for 
training in nonprovider sites, we believe Congress intended to create a 
monetary incentive for hospitals to rotate residents from the hospital 
to the nonhospital settings. We do not believe Congress intended for 
hospitals to be paid for residents who had previously been training at 
nonhospital sites without hospital funding.
    Further, in the Conference committee report accompanying the 
provision of Pub. L. 105-33 that authorizes IME payment for training in 
nonhospital settings, Congress stated that ``[t]he conference agreement 
includes new permission for hospitals to rotate residents through 
nonhospital settings, without reduction in indirect medical education 
funds.'' (Emphasis added.) (H.R. Conf. Rep. No. 105-217, 105th Cong., 
1st Sess., 817 (1997).)
    We note that, prior to enactment of Pub. L. 105-33, if a hospital 
rotated a resident from the hospital to train at a nonhospital site, 
the hospital could not count the time the resident spent at the 
nonhospital site for purposes of Medicare IME payments. As a result, 
the ``loss'' of IME payments acted as a disincentive and discouraged 
hospitals from rotating residents out of the hospital. It appears from 
the legislative history that Congress authorized hospitals to count 
residents in nonhospital sites for IME purposes as a specific incentive 
to encourage hospitals to rotate their residents to nonhospital sites 
(and not to encourage hospitals to incur the costs of a program at a 
nonhospital site that had already been funded by other sources). This 
legislative intent becomes more apparent when the nature of the 

Medicare IME payment is considered.

[[Page 45448]]

The Medicare IME payment is inherently a payment that reflects the 
increased operating costs of treating inpatients as a result of the 
hospital having a residency program. For example, as explained in the 
September 29, 1989 final rule (54 FR 40286), the indirect costs of 
medical education might include added costs resulting from an increased 
number of tests ordered by residents as compared to the number of tests 
normally ordered by more experienced physicians.
    The IME payment is an ``add-on'' adjustment that is made for each 
Medicare discharge from the areas subject to the IPPS in a teaching 
hospital. The authorization by Congress for IME payments relating to 
nonhospital services while residents are training at nonhospital sites 
would be absurd if not viewed as an incentive to transfer existing 
residency training from the hospital to the nonhospital setting. We do 
not believe Congress intended to permit IME payments to be allowable to 
the hospital that is incurring ``all or substantially all the costs'' 
of residents training in nonhospital sites except in the situations 
where either the hospital rotated residents from the hospital to the 
nonhospital settings or where the hospital started new programs in the 
nonhospital settings (and incurred the direct GME costs from the 
programs' inception). The illustrative situations described above and 
in the proposed rule in which nonhospital sites, such as dental 
schools, are shifting the costs of existing programs to the hospitals 
are not consistent with the intent of Congress to encourage hospitals 
to rotate residents from the hospital setting to nonhospital sites.
    Thus, we believe Congress intended both cited provisions of the Act 
on counting residents in nonhospital sites for purposes of direct GME 
and IME payments to be limited to situations in which hospitals rotate 
residents from the hospital to the nonhospital settings, and not 
situations in which nonhospital sites transfer the costs of an existing 
program at a nonhospital site to the hospital.
    Comment: One commenter cited section 1886(h)(5)(J) of the Act to 
support the general argument that CMS lacks the authority under the 
statute to ``impose additional conditions'' on counting FTE residents 
training in nonhospital sites--that is, the principles of 
redistribution of costs and community support. The commenter stated:

    This conclusion is further supported by Congress' treatment of 
family practice residency programs. In 42 U.S.C. Sec.  
1395ww(h)(5)(J), Congress provided a special payment provision for 
family practice residency programs. Specifically, Congress 
authorized hospitals to claim costs related to such programs even 
if, during the GME prospective payment base year--a year reimbursed 
under the reasonable cost system and a year to which the community 
support principle applied--the cost of such programs had been paid 
by the United States, a State, a political subdivision of the State, 
or an instrumentality of the State or political subdivision. 
Congress also provided that, in the event that such program payments 
were part of the PRA calculation during the GME base year, the 
payment in future years would be reduced ``in an amount equal to the 
proportion of such program funds received during the cost reporting 
period involved * * *.'' Thus, Congress has spoken to the issue of 
whether hospitals may claim costs in the current year if those costs 
have been paid in the past by third parties, and it has allowed 
reduction in current-year payments only if: (1) During the GME PPS 
base year, a third party had paid for the cost of the hospital's 
family practice residency program; and (2) as a result, the hospital 
had received a PRA that included an ``estimate of the amount that 
would have been recognized as reasonable * * * if the hospital had 
not received such funds.'' 42 U.S.C. Sec.  1395ww(h)(5)(J)(i). In 
all other situations, I submit, Congress does not permit the 
Secretary to reduce payments in the current year simply because, in 
the past, some third party may have paid the cost.

    Response: We disagree with the commenter that section 1886(h)(5)(J) 
of the Act supports the assertion that ``Congress has spoken to the 
issue'' of whether a hospital may claim third party costs and has 
allowed reductions in direct GME reimbursement resulting from 
redistribution of costs or community support in only the very limited 
circumstance of that exception in the Act. Generally, section 
1886(h)(5)(J) of the Act did two things: first, in subparagraph 
(J)(i)(1), Congress specifically allowed a hospital that only has an 
approved training program in family medicine and received a PRA in the 
base year of less than $10,000 for its family practice program, to 
receive a revised PRA that reflects the inclusion of ``funds from the 
United States, a State, or a political subdivision of a State * * *'' 
for the hospital's family practice program. Thus, the provision 
recognizes that ordinarily such funds would not be included in the 
hospital's base year per resident amount (because they were not 
incurred by the hospital in the base year). However, Congress 
explicitly created a narrow exception to the ``cost finding'' 
principles to allow such a hospital to include Federal, State, or local 
government grants to be included in the hospital's PRA base year 
calculation. Second, subparagraph (J)(i)(2) requires that direct GME 
payment to such a hospital that received a revised PRA amount under 
subparagraph (J)(i)(1) must also be reduced in subsequent cost 
reporting periods by the proportionate amount of funding the hospital 
receives from Federal, State, or local government payments. In other 
words, what subparagraph (J)(i)(2) does is to prohibit this hospital 
from receiving duplicative payments for the same GME program--both 
through the adjusted PRA and through continued Federal, State, and 
local government funding.
    The commenter argues that subparagraph (J)(i)(2) is the ``only'' 
situation where Congress has ``spoken'' about reductions in current 
year payment because of third party reimbursement. However, as we 
stated above, we believe the effect of subparagraph (J)(i)(2) is to 
prevent of duplicative payments for the same program that could 
otherwise occur in the narrow circumstances of the exception provided 
by section 1886(h)(5)(J), and has nothing to do with the continued 
applicability of the principles of redistribution of costs and 
community support. To the contrary, as we have stated, we believe that 
subparagraph (J)(i)(1) addresses a limited theoretical ``retroactive 
redistribution'' of costs and community support to allow a very narrow 
exception of allowing costs to be included in direct GME payment. Thus, 
we believe section 1886(h)(5)(J) of the Act would support our assertion 
that Congress intends application of redistribution of costs and 
community support to direct GME payment (except in the narrow 
circumstance of the type of hospital described in that section), rather 
than support the commenter's contrary assertion that the section is 
inconsistent with our proposal on application of the principles.
    Comment: One commenter suggested that the redistribution of costs 
and community support principles at nonhospital sites should apply on a 
``year-by-year basis,'' such that if another entity funds a training 
program during a particular fiscal year, the hospital would not be 
allowed to include the residents in its count for that fiscal year.
    Response: We believe the commenter's suggestion of a ``year-by-year 
basis'' policy is, in effect, already in place under existing Medicare 
policy without reference to the redistribution of costs or community 
support principles. Under the existing policy, where another entity 
funds a training program in a particular year while the residents are 
training at a nonhospital site--that is, incurs the residents' salaries 
and fringes, and the supervisory

[[Page 45449]]

physician costs (``all or substantially all of the costs''), the 
hospital may not include the residents in its FTE count for that fiscal 
year. This requirement, of course, is independent of the redistribution 
of costs and community support policy. It is based on the statutory 
requirement that allows a hospital to count residents training at 
nonhospital sites only if the hospital has incurred for all or 
substantially all of the costs of the program at that site during the 
hospital's fiscal year.
    Comment: One commenter stated that the 1989 final rule made clear 
that a hospital's resident count may also include residents for whom 
``community support was received'' through a State or local grant. 
Similarly, another commenter stated ``certain family medicine training 
programs that may have received outside funds, for example, State 
dollars, at any time in the past will be prohibited [by the hospital we 
proposed] from receiving GME reimbursement.''
    Similarly, another commenter stated that ``it is axiomatic'' that 
State-supported and public teaching hospitals receive State 
appropriations to support their residency programs. The commenter urged 
CMS to clarify that the application of the redistribution of costs and 
community support principles would not apply to State or local 
appropriations to public hospitals, with respect to the counting of FTE 
residents in either the hospital or the nonhospital setting.
    Response: As we explained in the 1989 final rule (54 FR 40302), 
grants that were restricted (those grants that were designated by the 
donor to pay for certain specified provider costs) or unrestricted were 
considered allowable costs of the hospital (including direct GME costs) 
when Medicare paid hospitals on a reasonable cost basis. The policy 
allowing payment to hospitals for costs that had been funded by grants 
was authorized by section 901 of the Omnibus Budget Reconciliation Act 
(OBRA) of 1980 (Pub. L. 96-499), which added section 1134 of the Act. 
Section 1134 of the Act applies to ``the reasonable costs of services 
provided by nonprofit hospitals or critical access hospitals.'' Section 
1134(1) of the Act specifies that a ``grant, gift or endowment or 
income therefrom which is to or for such a hospital * * *'' may not be 
deducted from the operating costs of such hospitals that are paid on a 
reasonable cost basis. Therefore, when hospitals were paid on a 
reasonable cost basis for direct GME costs, the ``community support'' 
that came from ``grants, gifts, or endowments'' was allowable under 
Medicare. We are clarifying in this final rule, that under the direct 
GME prospective payment methodology under section 1886(h) of the Act, 
if a hospital had received a grant, gift or endowment to subsidize its 
residency programs at the hospital, and the hospital requested direct 
GME payment for training the residents, it would not be considered 
community support. Under section 1134 of the Act, it is as if the 
hospital had itself incurred the cost for which it had received the 
grant subsidy. For example, if in 2003 a hospital received a State 
grant to fund its family practice program at the hospital, the grant 
would not be considered community support under our regulation. This is 
because we would treat the hospital as if itself incurred the costs for 
the family practice program, instead of the State grant.
    However, we note that this policy would not include ordinary State 
and local appropriations. As we mentioned in the January 12, 2001 final 
rule at 66 FR 3367, ``In administrative, legal and policy matters, we 
have consistently maintained that State appropriations for the cost of 
medical education activities constitute community support that is to be 
offset from a provider's allowable costs.'' Therefore, if a program 
were entirely funded by State or local appropriations, an inappropriate 
redistribution of costs would occur if the hospital subsequently begin 
to incur the costs of the residency program--for training inside or 
outside the hospital. Although, for most hospitals that receive State 
and local appropriations for their residency programs, the hospitals 
continuously incur (since the inception of the programs) some direct 
GME costs, there would be no disallowance of FTEs due to community 
support.
    We contrast the situation of a grant to a hospital with the 
situation of a grant to a nonhospital site. If, hypothetically, 
nonhospital sites were reimbursed by Medicare on a reasonable cost 
basis, and the nonhospital site had received grants to subsidize all of 
the direct GME costs for the residency program there, under section 
1134 of the Act, we would treat the costs the grant subsidized as if 
they were costs of the nonhospital site. If a hospital then tried to 
incur the direct GME costs, this could be a redistribution of costs or 
community support issue, since the hospital would be claiming FTE 
residents who had historically trained at the nonhospital site for whom 
the community had assumed the cost of that training, as described in 
the scenarios at 68 FR 27213.
    Comment: Several commenters objected to the sentence in the 
preamble to the proposed rule that stated: ``* * * a hospital is 
required to assume financial responsibility for the full complement of 
residents training in a nonhospital site in a particular program in 
order to count any FTE residents training there for purposes of IME.'' 
One commenter explained that there are a number of situations where a 
hospital is truly incurring the cost of having a resident at a site, 
but the hospital is not incurring the cost of the entire complement of 
residents. ``For example, if two different hospital programs each elect 
to send residents to the same clinic, under the interpretation in the 
[proposed rule], neither of the two hospitals would be able to count 
any of the residents because neither of the two programs would incur 
the cost of the full complement of residents.'' Another commenter 
believed that ``this change'' runs contrary to other current Medicare 

policies that focus on the resident rather than the program. The 
commenter believed that both the direct GME and IME regulations ``are 
replete with references to `resident' rather than `program'.'' The 
commenter believed that ``residency program'' is referenced only in the 
context of the requirement that, for residents to be counted for direct 
GME and IME payments, they must be part of an ``approved program'' 
(Sec.  413.86(f)(1)).
    Response: We understand the concerns of the commenters about the 
requirement for a hospital to incur ``all or substantially all of the 
cost'' of training residents in a training program at a nonhospital 
site. However, we do not believe this is a change in policy. We believe 
that the policy that requires a hospital to incur the cost of ``the 
program'' in the nonhospital site has existed since the passage of the 
direct GME provisions, section 9314 of the Omnibus Budget 
Reconciliation Act of 1986 (Pub. L. 99-509), and the passage of the IME 
provision, section 4621(b)(2) of the Balanced Budget Act of 1997 (Pub. 
L. 105-33), that permitted hospitals to continue to count residents in 
nonhospital sites, for purposes of direct GME and IME payment, if the 
hospital incurred ``all or substantially all of the cost'' of residents 
training in the program.
    As we explained in the proposed rule, this policy is derived from 
the language of the IME and direct GME provisions of the statute on 
counting residents in nonhospital settings; both sections 
1886(d)(5)(B)(iv) and 1886(h)(4)(E) of the Act state that the hospital 
must incur ``all, or substantially all, of the costs for the training 
program in that setting.'' (Emphasis added.) Therefore, we believe a 
better reading of this language is that hospitals must incur all

[[Page 45450]]

or substantially all of the cost for the full complement of residents 
in the training program at the nonhospital site.
    We note that the policy that requires the hospital to incur the 
cost of the program does appear to be somewhat of a departure from 
other current Medicare policies on graduate medical education that 
focus on the resident rather than the program, as the commenter 
suggests. However, we believe the statutory provisions cited above 
require hospitals to assume the cost of the full complement of 
residents training in the program at the nonhospital sites in order to 
count any FTE residents training at that site.
    In addition, as we noted at 68 FR 27217 of the proposed rule, and 
also above, under policy on the application of the redistribution of 
costs and community support principles, it is permissible for the 
hospital to count FTE residents where the hospital incurs direct GME 
costs of FTE residents that are added to an existing program, even 
though the hospital is not permitted to count the existing FTE 
residents due to the application of the redistribution of costs or 
community support rules. In the nonhospital setting, as a result of the 
interaction of these two separate FTE-counting requirements--(1) that 
the hospital must not violate the redistribution of costs and the 
community support principles in order to count the resident FTEs in the 
nonhospital settings; and (2) that the hospital must incur ``all or 
substantially all'' of the costs for the training program in that 
setting--a hospital would be prohibited from counting FTE residents 
added to an existing program at a nonhospital site unless the hospital 
incurs all or substantially all of the costs of training all of the 
residents in that program at that setting. That is, even if the 
hospital incurs all or substantially of the costs for all of the 
training program at the nonhospital site, the hospital would only be 
able to count the additional FTE residents who were not excluded by 
application of the redistribution of costs or community support 
principles.
    Comment: Several comments cited a letter from CMS (then the Health 
Care Finance Administration, or ``HCFA'') dated March 30, 1999 to C. 
Scott Litch of the American Association of Dental Schools (now the 
American Dental Education Association). Specifically, these commenters 
cited a sentence in the letter to Mr. Litch which stated: ``If a 
hospital establishes a new relationship with a dental clinic and meets 
the conditions for counting residents training outside the hospital, 
the hospital may count more residents currently for indirect and direct 
graduate medical education than were counted in 1996 if those residents 
are dental residents.'' One commenter stated that the ``new 
relationship'' referred to in the letter from CMS presupposes the 
existence of an ongoing program whose costs presumably had been met by 
means other than the hospital before the affiliation with a nonhospital 
dental clinic began. This commenter believed that this letter provided 
assurance to many hospitals that new affiliations with preexisting 
dental programs were permissible.
    Response: We do not agree with the commenter that the sentence in 
the letter to Mr. Litch ``presupposes the existence of an ongoing 
program'' where the costs of such a program ``had been met by means 
other than the hospital''. Rather, we believe the ``new relationship'' 
between the hospital and the dental clinic could be reconciled with 
application of the principles of redistribution of costs and community 
support and characterized by two possible interpretations, both of 
which would allow for the counting of residents in nonhospital sites--
(1) where the hospital would rotate residents from the hospital to the 
nonhospital site; or (2) where the hospital would fund new training 
slots at the nonhospital site (the dental clinic referred to in the Mr. 
Litch's letter). Such assignments from the hospital to the dental 
clinic, or new residency training slots, would be the ``new 
relationship,'' but in either case, no redistribution would occur. 
Therefore, we do not believe the letter from 1999 is necessarily 
inconsistent with the principles of redistribution of costs and 
community support described in the proposed rule.
    Comment: Many commenters, while remaining generally opposed to 
application of redistribution of costs and community support 
principles, requested that if CMS were to finalize the proposed rule, 
CMS apply the principles prospectively. One commenter, a dental school, 
explained that it had just admitted a new class of residents, many of 
whom will not complete their programs until 2006. The commenter 
believed that, in the application of the principles, CMS seeks to 
remove all Medicare funding for these residents retroactively. Along a 
similar vein, another commenter pointed out in support of the 
suggestion to apply the principles only prospectively, that the 
implementation of the proposed regulation would result in ``substantial 
dislocation and hardship to hospitals, dental and other schools, and 
the residents themselves.'' Therefore, the commenter believed CMS 
should indicate specifically in the final rule that such changes will 
only be applied to a provider's cost reporting period beginning on or 
after October 1, 2003, and CMS should not apply its final GME policy on 
redistribution of costs and community support to any prior cost 
reporting periods that remain open or unsettled, or are settled but 
potentially subject to reopening under the Medicare rules.
    In addition, several commenters requested clarification regarding 
the effective date for the proposed application of the principles of 
redistribution of costs and community support to FTE counts. 
Specifically, the commenters point to the following language in the 
proposed rule:
    [sbull] ``A hospital must continuously incur direct GME costs of 
residents training in a particular program at a training site since the 
date the residents first began training in that site in order for the 
hospital to count the FTE residents.'' (68 FR 27215)
    [sbull] ``We propose * * * to identify January 1, 1999, as the date 
our fiscal intermediaries should use to determine whether a hospital or 
another entity has been incurring the costs of training in a particular 
program at a training setting.'' (68 FR 27216)
    [sbull] ``[i]f the fiscal intermediaries determine that there is a 
redistribution of costs or community support exists with respect to 
certain residents prior to January 1, 1999, a disallowance of direct 
GME and IME payment with respect to those FTE residents would certainly 
be required.'' (68 FR 27216)
    [sbull] ``We are proposing that, effective October 1, 2003, in 
order for a hospital to receive IME and direct GME payment, the 
hospital must have been continuously incurring the direct GME cost of 
residents training in a particular program since the date the residents 
first began training in the program in order for the hospital to count 
the FTE residents.'' (68 FR 27417)
    Response: We have stated that we believe the principles of 
redistribution of costs and community support are longstanding Medicare 
policy. While we have reminded the public of the continuing application 
of the principles in various regulations and program guidance, we also 
recognize that CMS has not had occasion to invoke them in Agency policy 
expressions relating specifically to direct GME payments since the 
direct GME PRA base year.
    As we have stated, we believe redistributions would occur only in 
rare circumstances for residency training inside the hospital. Between 
1987 and 1997 when hospitals could count FTE

[[Page 45451]]

residents training in nonhospital sites for purposes of direct GME 
payments, but not IME payments, we did not observe the kinds of 
inappropriate counting of FTE residents we described in our proposed 
rule. It is only since hospitals have been allowed to count FTE 
residents training in nonhospital sites for purposes of IME payment, 
that CMS has become aware that cost shifting has become prevalent in 
the hospital industry, which has implicated the principles of 
redistribution of costs and community support. Therefore, in general, 
we are implementing a prospective effective date of October 1, 2003, 
for purposes of payment. That is, for direct GME, effective for 
portions of cost reporting periods beginning with October 1, 2003, and 
for IME, effective for discharges occurring on or after October 1, 
2003, a hospital must have been continuously incurring direct GME costs 
of residents training in a particular program since the date the 
residents first began training in the program in order for the hospital 
to count the FTE residents. We note that the effective dates apply only 
as they relate to disallowances of FTEs and bear no relation to 
determinations of redistributions or community support. Therefore, in 
general, a fiscal intermediary that determines that a redistribution of 
costs has taken place for a particular hospital prior to October 1, 
2003, may disallow FTEs based on that determination beginning with 
October 1, 2003. For example, if a fiscal intermediary determines that 
a redistribution of costs has occurred that affected 10 FTEs for direct 
GME and IME during the hospital's cost report ending in fiscal year 
ending in 1999, the fiscal intermediary would take disallowances for 
those 10 FTEs, but not until October 1, 2003, for purposes of direct 
GME and IME payment.
    In addition, because we have received a large number of public 
comments expressing surprise and confusion regarding our policy on 
these principles, we are grandfathering residents who began training in 
a program on or before October 1, 2003. That is, an FTE resident who 
began training in a residency program on or before October 1, 2003 (the 
effective date of this final rule), and with respect to whom there has 
been a redistribution or community support, may continue to be counted 
by a hospital for purposes of direct GME and IME payments after October 
1, 2003, until the resident has completed training in that program, or 
until 3 years after the date the resident began training in that 
program, whichever comes first. We believe continued direct GME and IME 
payments to the hospital while the ``redistributed'' residents finish 
their training for up to 3 years is appropriate to address many 
situations in which nonhospital sites have made arrangements with 
hospitals to shift the costs of training those residents. We understand 
that, in nonhospital sites, virtually all dental residency programs are 
of a duration of 3 years in length or less. This policy addresses the 
situation pointed out by the dental school commenter and other 
commenters that a school may have just admitted a new class of 
residents, many of whom will not complete 3-year programs until 2006.
    We note that this prospective ``grandfather'' policy does not apply 
to resident FTEs with respect to whom there has been a redistribution 
of costs or community support, and who begin training after October 1, 
2003. In addition, those residents described above who began training 
in a program on or before October 1, 2003, may be counted until those 
particular residents finish their training in that program (or 3 years, 
whichever comes first). In order to count such residents, we are 
requiring that hospitals identify those residents (by social security 
number) to their fiscal intermediary and specify the length of time the 
hospital will be counting these FTE residents for direct GME and IME 
payment purposes.
    We note that the policy described above that effectively 
``grandfathers'' residents who began their training on or before 
October 1, 2003, applies only as it relates to payments to hospitals 
for those specified FTE residents, and bears no relation to 
determinations of whether a redistribution of costs or community 
support has taken place. Therefore, if a fiscal intermediary determines 
that a redistribution of costs has taken place with respect to 
residents counted by a particular hospital even prior to October 1, 
2003, the intermediary will disallow any FTEs based on that 
determination, beginning October 1, 2003, except for the 
``grandfathered'' residents. Hospitals that continue to count 
grandfathered FTE residents (where the costs of whom had been 
redistributed) may only do so until those residents finish their 
training in the specific program they were training in on or before or 
to October 1, 2003 (which would be no later than September 30, 2006, 3 
years after October 1, 2003).
    For example, a fiscal intermediary determines for a hospital's FYE 
December 31, 2003 cost report that a redistribution of costs has taken 
place with respect to certain FTEs the hospital counted for direct GME 
and IME (that is, the costs of training residents at a nonhospital site 
were incurred by a university from 1990 through 1999). Assume that 5 
FTEs began training in a 2-year orthodontics program in a dental school 
on July 1, 2003, and another 5 residents begin their training in the 
same program on July 1, 2004. The 5 FTEs who began training on July 1, 
2003, are ``grandfathered,'' and, therefore, the fiscal intermediary 
would not disallow these 5 FTEs as of October 1, 2003. The hospital may 
continue to count these 5 FTEs that began training on July 1, 2003 
through June 30, 2005, when they finish the 2-year orthodontics 
program. We note that subsequent to completion of the 2-year 
orthodontics program on June 30, 2005, if any of these 5 FTEs 
participate in additional GME training programs, the fiscal 
intermediary would disallow these FTEs because disallowances for 
redistribution of costs and community support relate to FTE slots and 
not specific residents.
    However, the 5 FTEs that began training in the 2-year orthodontics 
program on July 1, 2004 are not ``grandfathered,'' and, therefore, 
beginning July 1, 2004 of the hospital's December 31, 2004 cost report, 
the fiscal intermediary will disallow IME and direct GME payment 
associated with these 5 FTE slots.
    Comment: Commenters disputed the situations we cited in the 
preamble to the proposed rule that were supposed to be illustrative of 
what we believe to be inappropriate application of Medicare direct GME 
and IME policy at 68 FR 27213. One commenter, in particular, requested 
information on the identity of programs cited in the examples.
    Response: We do not believe it is appropriate to disclose the 
identities of those cited in the examples. Therefore, we are unable to 
respond to the commenters' points on the matter, except to state that 
the situations in the examples represent what we believed are the more 
``egregious'' scenarios involving redistribution of costs and community 
support principles and inappropriate counting of FTE residents, we note 
that the same issues arise, and the same principles apply, whether the 
counting of residents relates to training that is taking place in 
another country, another State, or on the same hospital campus, as the 
hospital.
    Comment: One commenter believed that CMS's policy on the 
application of the redistribution of costs and community support will 
lead to considerable, ``but needless,'' litigation over what it means 
to ``incur'' the costs of off-site training.
    Response: We disagree with the commenter and see no reason to be

[[Page 45452]]

concerned that these clarifications would result in any more litigation 
than other Medicare payment policies that are conditioned on whether a 
provider incurs costs. For example, for several decades, Medicare 

policy required that hospitals ``incur'' costs in order to receive 
payment from Medicare. The Medicare statute and regulations currently 
require that a hospital incur certain costs in order to count FTE 
residents training in nonhospital sites for purposes of direct GME and 
IME payments. We are unsure why the requirement under the policy on 
redistribution of costs and community support that a hospital ``incur'' 
the direct GME cost continuously for a residency program at a training 
site is any more complex than other cost requirements under Medicare.
    Comment: One commenter suggested that we craft a narrower solution 
to the issue of inappropriate counting of FTE residents in nonhospital 
sites by focusing the language on salary and benefits for residents. 
The commenter believed that CMS could state that, unless the hospital 
in 1999 had incurred the costs of salary and benefits for FTE residents 
who were training in offsite locations, the hospital may not receive 
direct GME and IME payment for training those FTE residents at the 
nonhospital sites today.
    Response: We do not believe a policy such as the one the commenter 
suggested--determining redistribution of costs based upon whether a 
hospital continuously incurs the residents' salaries and benefits 
during training in the nonhospital site-- is necessary or appropriate. 
This is because, under the policy on redistribution of costs and 
community support we describe in the proposed rule and in this final 
rule, a hospital that continuously incurs the residents' salaries and 
benefits (from 1999 or before) while the residents train in the 
nonhospital site, or even inside the hospital, would not be 
redistributing costs if the nonhospital site later incurs the other 
direct GME costs (such as supervisory physician salaries) in the 
nonhospital site. There would be no redistribution of costs because the 
hospital would have continuously incurred at least some of the direct 
GME costs (the residents' salaries and benefits) since the inception of 
the program. However, we note that even if there has not been a 
redistribution of costs or community support with FTE residents 
training in a nonhospital site in such a scenario, the hospital would 
still need to meet the requirements in the existing regulations (at 
Sec.  413.86(f) and Sec.  412.105(1)(ii)(c)) in order to count those 
FTE residents for purposes of direct GME and IME payment.
    For example, Hospital A has had a family practice program with 10 
FTE residents for about 20 years, for which the hospital has incurred 
the residents' salaries and fringes and some other (but not all) direct 
GME costs for the program. For the first time, in fiscal year ending 
2003, Hospital A rotates 2 FTE residents to an ambulatory clinic (a 
nonhospital site), and fulfills the requirements at Sec.  413.86(f)(4), 
including incurring ``all or substantially all of the costs'' of the 
training program in the nonhospital site. There is no redistribution of 
costs with respect to these 2 FTE residents because Hospital A has 
continuously incurred some of the direct GME costs of the program--the 
residents' salaries--and therefore it may count the 2 FTE residents 
training at the clinic (up to the hospital's FTE cap), since it also 
has complied with the requirements at Sec.  413.86(f)(4).
    Comment: Some commenters suggested that the application of 
redistribution of costs and community support principles would impose 
large administrative burdens on hospitals to demonstrate which entity 
has been ``continuously incurring'' the costs of the residency 
training. One commenter stated: ``[t]his burden would be additive to a 
policy that already is fraught with excessive administrative 
requirements.''
    One commenter asked if hospitals would be required to document 
responsibility for the costs of training residents prior to January 1, 
1999.
    Response: If the hospital has continuously been incurring at least 
some of the direct GME costs (for example, resident salaries or 
supervisory physician salaries) since the inception of the residency 
program, we do not believe any additional documentation is necessary 
beyond which hospitals are already required to maintain. If resident or 
supervisory physician salaries, for instance, are paid through the 
hospital payroll, the hospital will have kept documentation of such 
costs for Federal tax purposes.
    In response to the second comment, we stated in the proposed rule 
that January 1, 1999 should be used by our fiscal intermediaries as the 
date for determinations of whether a hospital or another entity has 
been incurring the costs of a training in a particular program at a 
training site for purposes of determining whether there has been a 
redistribution of costs or community support. This date was chosen as 
an administrative convenience because we believe it could otherwise be 
difficult for our fiscal intermediaries to obtain contemporaneous 
documentation that the hospitals have appropriately been incurring 
costs in earlier years. Therefore, we believe that, for purposes of 
determining redistribution of costs or community support, most 
hospitals would only be required to maintain appropriate documentation 
to demonstrate that they have continuously been incurring the direct 
GME costs from January 1, 1999 forward. However, as we mentioned in the 
proposed rule, if the fiscal intermediaries determine that there was a 
redistribution of costs or community support for a fiscal year ending 
for a cost report for a particular hospital prior to January 1, 1999, 
the hospital would be required to show contemporaneous documentation to 
prove otherwise.
    Comment: One commenter stated that it may be difficult to track 
residents that have been funded by some type of community support. The 
commenter described a scenario where a program at a hospital has four 
internal medicine residents and one is covered by some type of 
community support for a 3-year period. The commenter stated that it may 
be difficult to track that slot over the next 5, 10, or 20 years to 
avoid submitting it for future direct GME or IME payments.
    Response: As we stated above, we understand there may be 
administrative issues that hospitals must confront in their efforts to 
comply with the principles of redistribution of costs and community 
support. However, we do not believe it would very difficult to track 
the FTEs in a program that receives community support. Once the FTE 
residents for which community support is received have been identified, 
the hospital will know the number of FTE residents to remove from the 
count that is submitted in future cost reports (all of which will be 
subject to audit by our fiscal intermediaries). Using the commenter's 
example, if direct GME costs for one out of four FTEs in an internal 
medicine program is identified as being entirely subsidized by 
community support for three years (the duration of an internal medicine 
program), the hospital would know to refrain from counting one FTE in 
future cost reports, even after the 3 years of training for a 
particular resident has passed. This is because, as the commenter 
seemed to understand, the redistribution of costs and community support 
principles are applied to the FTE resident training slots of a 
hospital; the principles are not associated with a particular resident, 
to which the principles could apply differently from year to year.
    Comment: One commenter disagreed with the choice of words used in 
the

[[Page 45453]]

proposed definition of ``redistribution of costs'' at proposed Sec.  
413.86(b). As proposed, the definition states: ``Redistribution of 
costs means an attempt by a hospital to increase the amount it is 
allowed to receive from Medicare under this section by counting FTE 
residents who were in medical residency programs where the costs of the 
programs had previously been incurred by the educational institution.'' 
In particular, the commenter objected to the first part of the 
definition: ``an attempt by a hospital to increase the amount it is 
allowed to receive from Medicare.'' The commenter believed that the 
phrase was unnecessary to the definition and should be deleted.
    Response: We understand the concern of the commenter. However, we 
have used ``the attempt'' language at Sec.  413.86(b) for the proposed 
definition of ``redistribution of costs'' primarily because we have 
adopted the language of the existing regulation at Sec.  413.85(c) that 
defines ``redistribution of costs'' (now applicable to costs of 
approved nursing and allied health education activities). The language 
was not intended to be offensive. Rather, we meant it to be descriptive 
of a possible motive for a redistribution of costs. In light of the 
commenter's suggestion, we are revising the language to be purely 
descriptive of the scenario of the redistribution and not reflect a 
possible motive. Accordingly, we are revising the language at Sec.  
413.86(b) to state: ``Redistribution of costs'' occurs when a hospital 
counts FTE residents in medical residency programs and the costs of the 
programs had previously been incurred by an educational institution. In 
the future, we will consider conforming changes to the definition of 
``redistribution of costs'' at Sec.  413.85(c) as well.
    Comment: Some commenters believed that, through the enactment of 
the 1996 cap on the count of allopathic and osteopathic residents, 
Congress has already dealt with the problem that CMS is attempting to 
revisit with the proposed rule. The commenters believed that when 
Congress exempted the dental residents from the caps, it intended to 
create hospital incentives for dental training. The commenters believed 
that the CMS redistribution of costs and community support policy 
contradicts this Congressional intent.
    Response: We do not believe that when Congress instituted the caps 
on the count of residents with the Balanced Budget Act of 1997, it was 
aware that inappropriate counting of FTE residents could occur through 
redistribution of costs. CMS, itself, did not become aware that many 
hospitals were engaging in these cost shifting arrangements, very often 
involving dental residents since at least October 1, 1997, when 
hospitals were authorized to count FTE residents for purposes of IME 
payments, as well as direct GME payments, for training in nonhospital 
sites. As we stated above, it is only since the audits by our fiscal 
intermediaries of the fiscal year ending 1998 and 1998 cost reports 
that have occurred within the last 2 years that CMS became aware that 
significant cost shifting was taking place. Therefore, we do not 
believe Congress would have been in a position to consider whether to 
authorize cost shifting in its 1997 legislation. Thus, we do not 
believe, as the commenters do, that Congress expected, or tacitly 
condone, cost shifting to dental residents as a result of exempting the 
dental residents from the 1996 caps. Rather, we believe that when 
Congress exempted dental residents from the 1996 caps, it intended to 
allow more dental training to occur in the hospital, not to authorize 
cost shifting from dental schools to hospitals and to the Medicare 
program.
    Comment: One commenter asked what types of costs the hospital is 
required to incur for training in nonhospital sites in order for there 
to be no redistribution of costs or community support. Specifically, 
the commenter described a scenario under which a teaching hospital and 
a medical school are related parties and asked whether the teaching 
hospital is required to pay for the teaching physician services 
relating to offsite rotations at a medical school clinic before the FTE 
residents participating in the rotation can be counted for purposes of 
IME or direct GME payment.
    Response: We understand from the scenario described by the 
commenter that hospital-based residents are being rotated to the 
medical school clinic. As such, we assume that the hospital is already 
incurring at least the residents' salary and fringe benefits. 
Therefore, when rotating the residents to the clinic, the hospital is 
incurring at least some of the direct GME costs of training the 
residents. Under these circumstances, a redistribution of costs has not 
taken place. However, according to the requirements for counting FTE 
residents in nonhospital settings under Sec.  413.86(f)(4), among other 
requirements, the hospital is required to incur the portion of the 
teaching physicians' salaries and fringe benefits attributable to 
direct GME (by the term ``related party,'' we are assuming that the 
medical school clinic is not provider-based as specified under Sec.  
413.65, and therefore, is not considered part of the hospital). Thus, 
under the commenter's scenario, the hospital may be prohibited from 
counting the FTE residents, not because of redistribution of costs but 
because of failure to incur ``all or substantially all of the cost'' 
under Sec.  413.86(f)(4) if the hospital is not incurring the 
supervisory physician's salary attributable to direct GME.
    Comment: A number of commenters argued that the proposed 
application of the redistribution of costs and community support 
principles is bad public policy from the perspective of access, quality 
and cost-effectiveness of oral health care.
    Response: We understand that dental training programs provide much 
needed oral health care to the American public and did not 
intentionally target them with our policy on redistribution of costs 
and community support. However, we believe much of the inappropriate 
cost sifting to hospitals and to the Medicare program is related to 
dental residency programs--which is probably due to the fact that 
dental residents are exempted from the statutory 1996 FTE caps. 
Although we regret that publication of this rule may upset some newly 
formed relationships between hospitals and dental schools, we continue 
to believe that the Medicare program should not pay for nonhospital 
dental residency training that had previously been funded by other 
sources, without any sponsorship by hospitals or the Medicare program.
    Comment: One commenter stated that by establishing a PRA floor 
equal to 85 percent of the locality-adjusted national average PRA, 
Congress created an exception to the principles of community support 
and redistribution of costs. The commenter noted that this floor 
increased reimbursement to a number of teaching hospitals around the 
country whose own PRAs were low ``precisely'' because the community or 
another educational institution had been bearing the training costs in 
the GME PRA base year. Therefore, the commenter argued, the PRA floor 
``picked up'' some of those disallowed costs, and that Medicare is, in 
effect, currently paying for those costs in the PRAs that were raised 
to the floor.
    Response: The commenter is referring to section 311 of the Balanced 
Budget Refinement Act (BBRA) of 1999 (Pub. L. 106-113), which, for FY 
2001, established a floor PRA at 70 percent of the locality-adjusted 
national average PRA, and to section 511 of the Benefits Improvement 
and Protection Act (BIPA) of 2000 (Pub. L. 106-554), which, for FY 
2002, established a floor PRA at 85 percent of the locality-adjusted 
national average PRA. Regulations concerning

[[Page 45454]]

these provisions are implemented at Sec.  413.86(e)(4). These 
provisions were intended, in part, to narrow the disparities (both high 
and low) in direct GME payments to teaching hospitals across the 
country. One of the reasons a number of hospitals had low base year 
PRAs is because a significant amount of their GME costs in the PRA base 
year was incurred by another entity (that is, the ``community''). 
(Variations in base year PRAs were otherwise due to differences in 
hospital-specific accounting practices and differences in reimbursement 
methods for supervising physician and resident salaries.) By providing 
for increased GME payments to certain hospitals with low PRAs, we do 
not believe Congress implicitly condoned, or made an exception to, the 
redistribution of costs and community support principles. We note that 
Congress provided for an increase to the floor PRA for all hospitals 
that had PRAs below the floor, not just to hospitals that, in the base 
year, did not incur certain GME costs. Rather, we believe Congress 
intended to provide increased GME payments to hospitals with low PRAs, 
regardless of the reasons those particular hospitals may have had low 
PRAs, in an attempt to even out some of the disparity in PRAs, 
nationally.
    Comment: A commenter noted that the among the examples cited in the 
proposed rule at 68 FR 27213 as illustrative of inappropriate 
application of Medicare IME and direct GME policy, we described a 
situation where a hospital on the East Coast of the United States is 
counting dental residents training in nonhospital sites in Hawaii. The 
commenter believed that we have incorrect information regarding this 
program, and that there is, in fact, no redistribution of costs from 
the community to the Medicare program with respect to the program in 
Hawaii. Specifically, the commenter explained that in August 2002, a 
hospital in New York placed one dental resident in a clinic located in 
Honolulu. The New York hospital pays the costs of the resident's 
stipend and the supervising faculty's salary, and there is a written 
agreement between the hospital and the clinic. The commenter stated 
that in the future, the program anticipates placing additional 
residents at other nonhospital sites in Hawaii.
    Response: As we stated in the preambles to the proposed rule and 
this final rule, there would be no redistribution of costs or community 
support if, from the outset of the program, a hospital incurs direct 
GME costs. Therefore, if, in fact, a hospital in New York has been 
incurring direct GME costs for a training program located in a clinic 
in Hawaii since the program's inception, then there would be no 
redistribution of costs or community support. The hospital in New York 
could count FTE residents training in the nonhospital site as long as 
the applicable requirements are met.
    Comment: One commenter that described a scenario in which a 
university funded a family practice program for many years. However, in 
2000, a Federally Qualified Health Center (FQHC) entered into a written 
agreement with the university and began reimbursing the university for 
``all or substantially all'' of the costs of the program. The FQHC has 
been receiving Medicare direct GME payments since that time. The 
commenter stated that under the terms of the proposed rule, this FQHC 
would be ineligible for receipt of GME payments, since, prior to 2000, 
the program was funded exclusively by the university.
    Response: The commenter raised the point that the redistribution of 
costs and community support principles are applicable to providers 
other than hospitals that may receive Medicare payments for residency 
training. Specifically, FQHCs and RHCs under Sec.  405.2468, CAHs under 
Sec.  413.70, and Medicare+Choice organizations (MCO) under Sec.  
422.270 may qualify to receive payments for direct GME costs. We note 
that the existing regulations at Sec.  405.2468(f)(6) for FQHCs and 
RHCs, and at Sec.  422.270(c) for MCOs, already clearly state that the 
allowable direct GME costs of these entities are subject to the 
redistribution of costs and community support principles in Sec.  
413.85(c). We agree with the commenter and are also clarifying the 
regulations at Sec.  413.86(i) to clearly state that the principles of 
redistribution of costs and community support apply equally to 
hospitals, FQHCs, RHCs, CAHs, and MCOs. Therefore, we agree that, in 
the situation described by the commenter the FQHC would not be eligible 
for Medicare direct GME payments since the family practice program 
represents a redistribution of costs from the community (that is, the 
university) to the Medicare program (that is, the FQHC through direct 
GME payments).
3. Rural Track FTE Limitation for Purposes of Direct GME and IME for 
Urban Hospitals That Establish Separately Accredited Approved Medical 
Programs in a Rural Area (Sec. Sec.  412.105(f)(1)(x) and 
413.86(g)(12))
a. Change in the Amount of Rural Training Time Required for an Urban 
Hospital To Qualify for an Increase in the Rural Track FTE Limitation
    To encourage the training of physicians in rural areas, section 
407(c) of Pub. L. 106-113 amended sections 1886(d)(5)(B) and 
1886(h)(4)(H) of the Act to add a provision that, in the case of an 
urban hospital that establishes separately accredited approved medical 
residency training programs (or rural tracks) in a rural area or has an 
accredited training program with an integrated rural track, an 
adjustment shall be made to the urban hospital's cap on the number of 
residents. For direct GME, the amendment applies to payments to 
hospitals for cost reporting periods beginning on or after April 1, 
2000; for IME, the amendment applies to discharges occurring on or 
after April 1, 2000.
    Section 407(c) of Pub. L. 106-113 did not define a ``rural track'' 
or an ``integrated rural track,'' nor are these terms defined elsewhere 
in the Act or in any applicable regulations.
    Currently, there are a number of accredited 3-year primary care 
residency programs in which residents train for 1 year of the program 
at an urban hospital and are then rotated for training for the other 2 
years of the 3-year program to a rural facility(ies). These separately 
accredited ``rural track'' programs are recognized by the Accreditation 
Council of Graduate Medical Education (ACGME) as ``1-2'' rural track 
programs. As far as CMS is able to determine, ACGME is the only 
accrediting body to ``separately accredit'' rural track residency 
programs, a requirement specified in Pub. L. 106-113.
    We implemented the rural track program provisions of section 
1886(d)(5)(B) and 1886(h)(4)(H) of the Act to address these ``1-2'' 
programs and to account for other programs that are not specifically 
``1-2'' programs but that include rural training components. As stated 
above, since there is no existing definition of ``rural track'' or 
``integrated rural track,'' we define at Sec.  413.86(b) a ``rural 
track'' and an ``integrated rural track'' as an approved medical 
residency training program established by an urban hospital in which 
residents train for a portion of the program at the urban hospital and 
then rotate for a portion of the program to a

[[Page 45455]]

rural hospital(s) or to a rural nonhospital site(s). We have previously 
noted that the terms ``rural track'' and ``integrated rural track,'' 
for purposes of this definition, are synonymous.
    To implement these provisions, we revised Sec.  413.86 to add 
paragraph (g)(11) (since redesignated as (g)(12)), and Sec.  412.105 to 
add paragraph (f)(1)(x) to specify that, for direct GME, for cost 
reporting periods beginning on or after April 1, 2000, or, for IME, for 
discharges occurring on or after April 1, 2000, an urban hospital that 
establishes a new residency program, or has an existing residency 
program, with a rural track (or an integrated rural track) may, under 
certain circumstances, include in its FTE count residents in those 
rural tracks, in addition to the residents subject to the FTE cap at 
Sec.  413.86(g)(4). (See the August 1, 2000 interim final rule with 
comment period (65 FR 47033) and the August 1, 2001 IPPS final rule (66 
FR 39902)). These regulations specify that an urban hospital may count 
the residents in the rural track in excess of the hospital's FTE cap up 
to a ``rural track FTE limitation'' for that hospital. We defined this 
rural track FTE limitation at Sec.  413.86(b) as the maximum number of 
residents (as specified in Sec.  413.86(g)(12)) training in a rural 
track residency program that an urban hospital may include in its FTE 
count, in addition to the number of FTE residents already included in 
the hospital's FTE cap.
    Generally, the rural track policy is divided into two categories: 
Rural track programs in which residents are rotated to a rural area for 
at least two-thirds of the duration of the program; and rural track 
programs in which residents are rotated to a rural area for less than 
two-thirds of the duration of the program. Currently, family practice 
is the only specialty that has separately accredited rural track 
programs. As previously noted, to account for other specialties that 
have program lengths greater than or less than 3 years, or that are not 
``1-2'' programs, but may establish separately accredited rural track 
residency programs that are longer than 3 years, our regulations 
specify that residents must train in the rural area for ``two-thirds of 
the duration of the program,'' rather than ``2 out of 3 program 
years,'' in order for the urban hospital to count FTEs in the rural 
track (up to the rural track FTE limitation) in addition to the 
residents included in the hospital's FTE limitation. Thus, for example, 
under current policy, if a surgery program, which is a 5-year program, 
were to establish a separately accredited rural track, the urban 
hospital must rotate the surgery residents to the rural area for at 
least two-thirds of the duration of the 5-year program in order to 
qualify to count those FTEs in excess of the hospital's FTE cap, as 
provided in Sec.  413.86(g)(12) and Sec.  412.105(f)(1)(x).
    Accordingly, our policy for determining whether an urban hospital 
qualifies for an adjustment to the FTE cap for training residents in 
rural areas is dependent upon the proportion of time the residents 
spend training in the rural areas. If the time spent training in rural 
areas (either at a rural hospital or a rural nonhospital site) 
constitutes at least two-thirds of the duration of the program, then 
the urban hospital may include the time the residents train at that 
urban hospital in determining GME payments. However, if the urban 
hospital rotates residents to rural areas for a period of time that is 
less than two-thirds of the duration of the program, although the rural 
hospital may count the time the residents train at the rural hospital 
if the program is new, the urban hospital may not include the time the 
residents train at the urban hospital for GME payment purposes (unless 
it can do so within the hospital's FTE cap).
    When we first implemented this policy on rural tracks, it was 
consistent with our understanding of how the ACGME accredits rural 
track ``1-2'' programs, in which residents train for 1 year of the 
program at an urban hospital and are then rotated for training years 2 
and 3 to a rural facility. We believed that the ACGME did not 
separately accredit an approved program as a rural track program unless 
it met this ``1-2'' condition; that is, the residents were spending 
one-third of program training in the urban area and two-thirds of the 
program training in the rural area. However, we have recently learned 
that there are a few rural track programs that are separately 
accredited by the ACGME as ``1-2'' rural track programs, but the 
residents in these programs are not training in rural areas for at 
least two-thirds of the duration of the program. We understand that in 
certain instances in which the case-mix of the rural facilities might 
not be sufficiently broad to provide the residents with an acceptable 
range of training opportunities, the ACGME allows the residents in 
program years 2 and 3 to return to the urban hospital for some training 
in both years. However, because the training in years 2 and 3 is 
predominantly occurring at the rural locations, the ACGME still 
separately accredits the urban and rural portions as a ``1-2'' program.
    The existing regulations at Sec. Sec.  412.105(f)(1)(x) and 
413.86(g)(12) specify two main criteria for an urban hospital to count 
the time spent by residents training in a rural track while at the 
urban hospital in excess of the hospital's FTE limitation: (1) the 
program must be separately accredited by the ACGME; and (2) the time 
spent training in rural areas (either at a rural hospital or a rural 
nonhospital site) must constitute at least two-thirds of the duration 
of the program.
    We believe that an urban hospital that operates a program that is 
separately accredited by the ACGME as a ``1-2'' program, but in which 
residents train in rural areas for more than half but less than two-
thirds of the duration of the program, should still be allowed to count 
those FTE residents for GME payment purposes. Therefore, to be 
consistent with the ACGME accreditation practices, in the May 19, 2003 
proposed rule, we proposed to revise our regulations. Proposed Sec.  
413.86(g)(12) still addressed our policy that an urban hospital 
qualifies for an adjustment to the FTE cap for training in rural areas 
based upon the proportion of time the residents spend training in the 
rural areas. However, instead of using ``two-thirds'' as the criterion 
to specify the amount of time residents training in the rural areas 
under regulations at Sec. Sec.  413.86(g)(12)(i) through (iv) and 
412.105(f)(1)(x), as under current policy, the proposal would use 
``one-half'' as the criterion. This proposal addressed the limited 
cases where ACGME separately accredits programs as ``1-2'' rural tracks 
but residents in those programs train in the rural areas less than two-
thirds of the time, although greater than one-half of the time. 
Specifically, we proposed at Sec.  413.86(g)(12) to state:
    [sbull] If an urban hospital rotates residents to a separately 
accredited rural track program at a rural hospital(s) for at least two-
thirds of the duration of the program for cost reporting periods 
beginning on or after April 1, 2000 and before October 1, 2003, or for 
more than one-half of the duration of the program for cost reporting 
periods beginning on or after October 1, 2003, the urban hospital may 
include those residents in its FTE count for the time the rural track 
residents spend at the urban hospital.
    [sbull] If an urban hospital rotates residents to a separately 
accredited rural track program at a rural nonhospital site(s) for at 
least two-thirds of the duration of the program for cost reporting 
periods beginning on or after April 1, 2000, and before October 1, 
2003, or for more than one-half of the duration of the program for cost

[[Page 45456]]

reporting periods beginning on or after October 1, 2003, the urban 
hospital may include those residents in its FTE count, subject to the 
requirements under Sec.  413.86(f)(4).
    [sbull] If an urban hospital rotates residents in the rural track 
program to a rural hospital(s) for less than two-thirds of the duration 
of the program for cost reporting periods beginning on or after April 
1, 2002, and before October 1, 2003, or for one-half or less than one-
half of the duration of the program for cost reporting periods 
beginning on or after October 1, 2003, the rural hospital may not 
include those residents in its FTE count (if the rural track is not a 
new program under Sec.  413.86(g)(6)(iii), or if the rural hospital's 
FTE count exceeds that hospital's FTE cap), nor may the urban hospital 
include those residents when calculating its rural track FTE 
limitation.
    [sbull] If an urban hospital rotates residents in the rural track 
program to a rural nonhospital site(s) for a period of time that is 
less than two-thirds of the duration of the program for cost reporting 
periods beginning on or after April 1, 2002, and before October 1, 
2003, or for one-half or less than one-half of the duration of the 
program for cost reporting periods beginning on or after October 1, 
2003, the urban hospital may include those residents in its FTE count, 
subject to the requirements under Sec.  413.86(f)(4).
    We also proposed to make a conforming change to Sec.  
412.105(f)(1)(x) to make these proposed provisions applicable to IME 
payments for discharges occurring on or after October 1, 2003.
    We believe the proposal produces a more equitable result than the 
existing policy; the proposal encompasses what we believe to be all 
situations in which the ACGME separately accredits rural track programs 
and in which residents in the programs spend a majority of the time 
training in rural settings, fulfilling the intent of Congress for 

Medicare to provide GME payments for significant rural residency 
training.
    Comment: Several commenters supported our proposal that, effective 
for cost reporting periods beginning on or after October 1, 2003, an 
urban hospital would be allowed to include residents in its FTE count 
above its FTE cap for the time that the residents train at the urban 
hospital, if the residents rotate to a separately accredited rural 
track program in a rural area for more than one-half of the duration of 
the program. The commenters believed that this proposed policy better 
reflects Congressional intent to encourage training in rural areas, 
while allowing residency programs the flexibility to rotate residents 
back to urban areas for needed clinical experiences that are not 
available in the rural setting.
    One commenter recommended that the proposal should reduce the 
required rural training time even further, since research suggests that 
more than 50 percent of family practice residents who spend as little 
as 3 months training in rural areas end up practicing in rural 
settings.
    Response: We agree with the commenters that an urban hospital that 
operates a program that is separately accredited by the ACGME as a ``1-
2'' program, but in which residents train in rural areas for more than 
half but less than two-thirds of the duration of the program, should 
still be allowed to count those FTE residents for GME payment purposes. 
However, we do not agree that urban hospitals should be allowed to 
receive an increase in their FTE caps to include residents in its FTE 
count for the time that the residents train at the urban hospital, if 
the residents rotate to a rural area for one-half or less than one-half 
of the duration of the program. As we stated in the August 1, 2001 
Federal Register (66 FR 39904-39905), we interpret section 
1886(h)(4)(H)(iv) of the Act as only allowing for an urban hospital to 
receive an adjustment under the rural track provision if the rural 
track program is ``separately accredited.'' In order to be separately 
accredited as a rural track, the program must meet the ACGME's ``1-2'' 
criteria; that is, the residents are typically spending approximately 
two-thirds of the duration of the program in the rural area. We also 
explained that while we agree that post-residency retention in rural 
areas is important, we also believe it is important to prevent 
hospitals from receiving adjustments to their FTE caps in situations 
when only a nominal amount of training occurs in the rural area. 
Therefore, we are not adopting the commenter's request to allow an 
urban hospital to receive an increase in its FTE caps to include 
residents in its FTE count for the time that the residents train at the 
urban hospital, if the residents rotate to a rural area for one-half or 
less than one-half of the duration of the program.
    Comment: One commenter that works for a community health center 
(CHC) that treats a high percentage of patients below the poverty line 
expressed concern about the detrimental effects that shrinking hospital 
revenues are having on the training of family practice residents at the 
CHC and at other rural and community-based settings. The commenter 
noted that doubling the number of CHCs is a goal of the President, and 
urged that, if there should be further ``restraint'' on teaching 
programs, programs that expand into CHCs should be exempt from such 
restrictions.
    Response: We appreciate the comment. However, we note that since we 
did not specifically make any proposals related to residency training 
in community health centers, this comment is outside the scope of this 
final rule. Therefore, we are not responding to it at this time.
b. Inclusion of Rural Track FTE Residents in the Rolling Average 
Calculation
    Section 1886(h)(4)(G) of the Act, as added by section 4623 of Pub. 
L. 105-33, provides that, for a hospital's first cost reporting period 
beginning on or after October 1, 1997, the hospital's FTE resident 
count for direct GME payment purposes equals the average of the actual 
FTE resident count for that cost reporting period and the preceding 
cost reporting period. Section 1886(h)(4)(G) of the Act requires that, 
for cost reporting periods beginning on or after October 1, 1998, a 
hospital's FTE resident count for direct GME payment purposes equals 
the average of the actual FTE resident count for the cost reporting 
period and the preceding two cost reporting periods (that is, a 3-year 
rolling average). This provision phases in over a 3-year period any 
reduction in direct GME payments to hospitals that results from a 
reduction in the number of FTE residents below the number allowed by 
the FTE cap. We first implemented this provision in the August 29, 1997 
final rule with comment period (62 FR 46004) and revised Sec.  
413.86(g)(5) accordingly. Because hospitals may have two PRAs, one for 
residents in primary care and obstetrics and gynecology (the ``primary 
care PRA''), and a lower PRA for nonprimary care residents, we revised 
our policy for computing the rolling average for direct GME payment 
purposes (not for IME) in the August 1, 2001 final rule (66 FR 39893) 
to create two separate rolling averages, one for primary care and 
obstetrics and gynecology residents (the ``primary care rolling 
average''), and one for nonprimary care residents. Effective for cost 
reporting periods beginning on or after October 1, 2001, direct GME 
payments are calculated based on the sum of: (1) the product of the 
primary care PRA and the primary care rolling average; and (2) the 
product of the nonprimary care PRA and the nonprimary care FTE rolling 
average. (This sum is then multiplied by the

[[Page 45457]]

Medicare patient load to determine Medicare direct GME payments).
    Section 407(c) of Pub. L. 106-113, which amended sections 
1886(d)(5)(B) and 1886(h)(4)(H) of the Act to create the rural track 
provision, provided that, in the case of an urban hospital that 
establishes a separately accredited rural track, ``* * * the Secretary 
shall adjust the limitation under subparagraph (F) in an appropriate 
manner insofar as it applies to such programs in such rural areas in 
order to encourage the training of physicians in rural areas'' 
(emphasis added). Subparagraph (F) of the Act is the provision that 
establishes a cap on the number of allopathic and osteopathic FTE 
residents that may be counted at each hospital for Medicare direct GME 
payment purposes. Thus, the provision authorizes the Secretary to allow 
for an increase to an urban hospital's FTE cap on allopathic and 
osteopathic residents in certain instances when an urban hospital 
establishes a rural track program. Although the rural track provision 
effectively allows an increase to the urban hospital's FTE cap by 
adjusting the FTE limitation under subparagraph (F), the statute makes 
no reference to subparagraph (G), the provision concerning the rolling 
average count of residents. That is, the statute does not provide for 
an exclusion from the rolling average for the urban hospital for those 
FTE residents training in a rural track.
    Since we implemented this rural track provision in the August 1, 
2000 interim final rule with comment period (65 FR 47033), we have 
interpreted this provision to mean that, except for new rural track 
programs begun by urban teaching hospitals that are establishing an FTE 
cap for the first time under Sec.  413.86(g)(6)(i), when an urban 
hospital establishes a new rural track program or expands an existing 
rural track program, FTE residents in the rural track that are counted 
by the urban hospital are included in the hospital's rolling average 
calculation immediately. Although we have not specified in the 
regulations that rural track FTE residents counted by an urban hospital 
are included in the hospital's rolling average FTE resident count, this 
has been our policy. The Medicare cost report, Form CMS-2552-96 (line 
3.05 on Worksheet E, Part A, for IME payments, and on line 3.02 on 
Worksheet E-3, Part IV, for direct GME payments), reflects this policy. 
Accordingly, FTE residents in a rural track program are to be included 
in the urban hospital's rolling average count for IME and direct GME 
for cost reporting periods beginning on or after April 1, 2000.
    In the May 19, 2003 proposed rule, we proposed to revise the 
regulations at Sec.  413.86(g)(5) to add a new paragraph (vii) to 
clarify that, subject to regulations at Sec.  413.86(g)(12), except for 
new rural track programs begun by urban hospitals that are first 
establishing an FTE cap under Sec.  413.86(g)(6)(i), when an urban 
hospital with an existing FTE cap establishes a new program with a 
rural track (or an integrated rural track), or expands an existing 
rural track (or an integrated rural track) program, the FTE residents 
in that program that are counted by the urban hospital are included in 
the urban hospital's rolling average FTE resident count immediately. We 
also proposed to revise Sec. Sec.  413.86(g)(12)(i)(A), (g)(12)(ii)(B), 
and (g)(12)(iv)(A) to indicate that for the first 3 years of the rural 
track's existence, the rural track FTE limitation for each urban 
hospital will be the actual number of FTE residents, subject to the 
rolling average, training in the rural track at the urban hospital.
    Comment: Commenters supported our proposal to revise Sec.  
413.86(g)(5) to clarify that the FTE residents in that program that are 
counted by the urban hospital are included in the urban hospital's 
rolling average FTE resident count immediately. The commenters stated 
that allowing immediate inclusion of rural track resident counts will 
serve to assist urban hospitals in their development of educational 
partnerships with rural hospitals.
    Response: We appreciate the commenters support and, as explained 
below, are adopting revisions to the regulations concerning inclusion 
of rural track residents in the rolling average count of urban 
hospitals as final.
    Except for new rural track programs begun by urban hospitals that 
are first establishing an FTE cap under Sec.  413.86(g)(6)(i), or for 
rural hospitals that are establishing new rural track programs under 
Sec.  413.86(g)(6)(iii), we are implementing sections 1886(d)(5)(B) and 
1886(h)(4)(H) of the Act to require that FTE residents that are counted 
by an urban hospital based on the residents' participation in a rural 
track are included in the rolling average calculation. Accordingly, for 
IME and direct GME purposes, unless the rural track program is a new 
program under Sec.  413.86(g)(13) and qualifies for a cap adjustment 
under Sec.  413.86(g)(6)(i) or (g)(6)(iii), in instances where an urban 
hospital increases the number of residents it trains due to the 
establishment of a new or an expansion of an existing rural track 
program, the additional FTE residents in the rural track program are 
only gradually included (over a 3-year period) in the urban hospital's 
FTE count, since they are immediately included in the rolling average 
calculation of the urban hospital.
    The following is an example of how residents in a rural track would 
be included in the rolling average calculation:
    Assume that urban Hospital A, with a fiscal year end (FYE) date of 
June 30, had 10 unweighted FTE residents training in its cost reporting 
period ending June 30, 1996, thereby establishing an FTE cap of 10. 
Hospital A only trains primary care residents. In its cost reporting 
periods ending on June 30, 2002, and June 30, 2001, Hospital A again 
trained 10 FTE residents. However, in July 2002, Hospital A starts a 
rural training track program, adding 2 FTE residents. Since the 
additional rural track residents are included immediately in the 
rolling average, in FYE June 30, 2003, Hospital A's FTE residents for 
payment purposes equal 10.67 FTEs (12 + 10 + 10 / 3) and not 12 FTEs 
[(10 + 10 + 10 / 3) + 2], which would be the FTE count if FTEs in a 
rural track program were not subject to the rolling average 
calculation.
    We are finalizing our proposed revision of Sec.  413.86(g)(5) to 
add a new paragraph (vii) as explained above. In addition, we are 
finalizing our revision of Sec. Sec.  413.86(g)(12)(i)(A), 
(g)(12)(ii)(B), and (g)(12)(iv)(A) to indicate that for the first 3 
years of the rural track's existence, the rural track FTE limitation 
for the urban hospital will be the actual number of FTE residents, 
subject to the rolling average, training in the rural track at the 
urban hospital.
4. Technical Change Relating to Affiliated Groups and Affiliation 
Agreements
    Section 1886(h)(4)(H)(ii) of the Act permits, but does not require, 
the Secretary to prescribe rules that allow institutions that are 
members of the same affiliated group (as defined by the Secretary) to 
elect to apply the FTE resident limit on an aggregate basis. This 
provision allows the Secretary to give hospitals flexibility in 
structuring rotations within a combined cap when they share a 
resident's time. Consistent with the broad authority conferred by the 
statute, we established criteria for defining an ``affiliated group'' 
and an ``affiliation agreement'' in both the August 29, 1997 final rule 
(62 FR 45965) and the May 12, 1998 final rule (63 FR 26317). We further 
clarified our policy concerning affiliation agreements in the August 1, 
2002 final rule (67 FR 50069).

[[Page 45458]]

    We are aware that there has been some confusion at times among 
members of the provider community when using the term ``affiliation 
agreement,'' since the term is used in contexts other than for Medicare 

GME payment purposes. For example, an ``affiliation agreement'' is a 
term historically used in the academic community that generally relates 
to agreements made between hospitals and medical schools or among 
sponsors of medical residency education programs. To help prevent 
further confusion, in the May 19, 2003 proposed rule, we proposed to 
change the term in the regulations to ``Medicare GME affiliation 
agreement.'' We believe this will help to distinguish these agreements 
used for purposes of GME payments from agreements used for other 
purposes in the provider community. We proposed to revise the 
regulations at Sec.  413.86(b) to state ``Medicare GME affiliated 
group,'' and ``Medicare GME affiliation agreement''. We proposed to 
make similar revisions to Sec.  413.86(g)(4)(iv), (g)(7)(i) through 
(v), and Sec.  412.105(f)(1)(vi) for IME payment purposes.
    Comment: Commenters supported our proposal to change the terms 
``affiliated group'' and ``affiliation agreement'', as defined in Sec.  
413.86(b), to ``Medicare GME affiliated group'' and ``Medicare GME 
affiliation agreement'', respectively. The commenters believed that the 
changes in terminology will help distinguish these terms from other 
affiliation agreements that are entered into by hospitals, medical 
schools, and other institutions that sponsor residency training.
    Response: We agree with the commenters and are adopting as final 
the proposed changes throughout Sec.  412.105 for IME and Sec.  413.86 
for direct GME.

Out of Scope Comments Relating to GME

    Comment: Several comments addressed miscellaneous IME and direct 
GME issues, including the initial residency period (IRP) and volunteer 
physicians.
    Response: Because we did not propose any changes in policy 
concerning these issues, we are unable to respond to these comments at 
this time. We will consider them for purposes of future rulemaking.

G. Updates to the Reasonable Compensation Equivalent (RCE) Limits 
(Sec.  415.70)

1. Background
    Under the Medicare program, payment for services furnished by a 
physician is made under either the Hospital Insurance Program (Part A) 
or the Supplementary Medical Insurance Program (Part B), depending on 
the type of services furnished. In accordance with section 1848 of the 
Act, physicians' charges for medical or surgical services to individual 

Medicare patients generally are covered under Part B on a fee-for-
service basis under the Medicare physician fee schedule. The 
compensation that physicians receive from or through a provider for 
services that benefit patients generally (for example, administrative 
services, committee work, teaching, and supervision) can be covered 
under Part A or Part B, depending on the provider's setting.
    As required by section 1887(a)(2)(B) of the Act, allowable 
compensation for services furnished by physicians to providers that are 
paid by Medicare on a reasonable cost basis is subject to reasonable 
compensation equivalent (RCE) limits. Under these limits, payment is 
determined based on the lower of the actual cost of the services to the 
provider (that is, any form of compensation to the physician) or a 
reasonable compensation equivalent. For purposes of applying the RCE 
limits, physician compensation costs means monetary payments, fringe 
benefits, deferred compensation and any other items of value (excluding 
office space or billing and collection services) that a provider or 
other organization furnishes a physician in return for the physician's 
services.
    The RCE limits do not apply to the costs of physician compensation 
that are attributable to furnishing inpatient hospital services paid 
under the IPPS or as GME costs. In addition, RCE limits do not apply to 
the costs CAHs incur in compensating physicians for services. 
Furthermore, compensation that a physician receives for activities that 
may not be paid under either Part A or Part B is not considered in 
applying the RCE limits.
    The limits apply equally to all physician services to providers 
that are payable on a reasonable cost basis under Medicare. If a 
physician receives any compensation from a provider for his or her 
physician services to the provider (that is, those services that 
benefit patients generally), payment to those affected providers for 
the costs of such compensation is subject to the RCE limits. The RCE 
limits are not applied to payment for services that are identifiable 
medical or surgical services to individual patients and paid under the 
physician fee schedule, even if the physician agrees to accept 
compensation (for example, from a hospital) for those services. 
(However, payments to teaching hospitals that have elected to be paid 
for these services on a reasonable cost basis in accordance with 
section 1861(b)(7) of the Act are subject to the limits.)
    Section 415.70(b) of the regulations specifies the methodology for 
determining annual RCE limits, considering average physician incomes by 
specialty and type of location, to the extent possible using the best 
available data. On October 31, 1997, the revised RCE limits update 
methodology was published in the Federal Register (62 FR 59075). For 
cost reporting periods beginning on or after January 1, 1998, updates 
to the RCE limits are calculated using the Medicare Economic Index 
(MEI). The inflation factor used to develop the initial RCE limits and, 
subsequently, to update those limits to reflect increases in net 
physician compensation was the Consumer Price Index for All Urban 
Consumers (CPI-U). In 1998, we revised the update methodology for the 
RCE limits by replacing the CPI-U with the inflation factor for the 
physician fee schedule (the MEI) to achieve a measure of consistency in 
the methodologies employed to determine reasonable payments to 
physicians for direct medical and surgical services furnished to 
individual patients and reasonable compensation levels for physicians' 
services that benefit provider patients generally.
2. Updated RCE Limits
    In the May 19, 2003 proposed rule, we indicated our intent to 
publish updated payment limits on the amount of allowable compensation 
for services furnished by physicians to providers in this FY 2004 IPPS 
final rule. These revised RCE limits are based on updated economic 
index data and replace the limits that were published in the Federal 
Register on May 5, 1997 (62 FR 24483). We calculated the revised RCE 
limits by using the methodology published in the Federal Register on 
October 31, 1997 (62 FR 59075). These limits are specified in the chart 
below and are effective for cost reporting periods beginning on or 
after January 1, 2004.
    The revised RCE limits are mere updates that have been calculated 
by applying the most recent economic index data. In the proposed rule, 
we did

[[Page 45459]]

not propose to change the methodology used to determine the limits. We 
indicated that, in accordance with Sec.  415.70(f), we are allowed to 
publish the revised RCE limits in a final rule without prior 
publication of a proposed rule for public comment. Furthermore, 
indicated our belief that publication of the revised RCE limits in a 
proposed rule with opportunity for public comment was unnecessary, and 
that we found good cause to waive the procedure.
    Comment: One commenter was encouraged to learn of our proposal to 
publish updated RCE limits and suggested that these updates occur on an 
annual basis.
    Response: We will continue to review the RCE limits on a regular 
basis by applying the most recent economic index data and publish 
updates as necessary.
3. Application of RCE Limits
    This section, as well as the two following sections, is not 
describing new policy, but rather is simply a discussion of a 
continuation of the existing policies with respect to the application 
of and exceptions to the RCE limits and the geographic area 
classifications used for purposes of establishing the RCE limits. We 
will continue to use the RCE limits to compute Medicare payments when a 
physician is compensated by a provider that is subject to the RCE 
limits in some or all of its areas. We also will use these limits when 
the physician is compensated by any other related organization for 
physician administrative, supervisory, and other provider services paid 
under Medicare. In applying the RCE limits, the intermediary will 
assign each compensated physician to the most appropriate specialty 
category. If no specialty category is appropriate (for example, in 
determining the reasonable cost for an emergency room physician), the 
fiscal intermediary will use the RCE level for the ``Total'' category, 
which is based on income data for all physicians. The fiscal 
intermediary will determine the appropriate geographic area 
classification given in Table 9 of the addendum of this final rule.
    If the physician's contractual compensation covers all duties, 
activities, and services furnished to the provider and to its patients 
and the physician is employed full-time, the appropriate specialty 
compensation limit will be used and adjusted by the physician's 
allocation agreement to arrive at the program's share of allowable 
costs as physician compensation costs. In the absence of an allocation 
agreement, we generally will assume that 100 percent of the 
compensation was related to services paid under the physician fee 
schedule and that there are no allowable costs for the physician's 
services to the provider.
    If a physician's compensation from the provider represents payment 
only for services that benefit patients generally (that is, the 
physician bills fees for all services furnished to individual 
patients), the appropriate specialty compensation limit will be used. 
If a physician is employed by a provider to furnish services of general 
benefit to patients on other than a full-time basis, the RCE amount 
will be adjusted upward or downward to reflect the percentage of time 
his or her actual hours related to a full work year of 2,080 hours.
4. Exceptions to the RCE Limits
    Some providers, particularly but not exclusively small or rural 
hospitals, may be unable to recruit or maintain an adequate number of 
physicians at a compensation level within the prescribed limits. In 
accordance with section 1887(a)(2)(C) of the Act, if a provider is able 
to demonstrate to the intermediary its inability to recruit or maintain 
physicians at a compensation level allowable under the RCE limits (as 
documented, for example, by unsuccessful advertising through national 
medical or health care publications), the intermediary may grant an 
exception to the RCE limits established under these rules.
5. Geographic Area Classifications for RCE Limits
    We adjust the RCE limits to account for differences in salary 
levels by location as well as by specialty. Under our methodology for 
establishing limits, and in the limits set forth below, we have 
classified geographic areas into three types: nonmetropolitan areas, 
metropolitan areas less than 1 million, and metropolitan areas greater 
than 1 million.
    As we do for purposes of the IPPS and the physician fee schedule, 
we use the most current MSA designations for purposes of establishing 
the RCE limits. In New England, we use the NECMAs for this purpose. 
Tables 4A and 4B of the Addendum to this final rule includes 
information that identifies, by type of location (urban and rural), the 
geographic areas affected; that is, they list all MSAs and their 
constituent counties and identifies whether their population are 
classified as large urban. Any county not listed in the tables and all 
other affected U.S. possessions and territories not part of a State are 
considered rural areas. This information will enable providers, 
physicians, Medicare fiscal intermediaries, and other members of the 
public to determine which RCE limit level will apply in specific areas.

Estimates of FTE Annual Average Net Compensation Levels for Cost Reporting Periods Beginning on or After January
                                                    1, 2004 *
----------------------------------------------------------------------------------------------------------------
                                                                          Metropolitan areas  Metropolitan areas
                      Specialty                         Nonmetropolitan      less than one     greater than one
                                                             areas              million             million
----------------------------------------------------------------------------------------------------------------
Total...............................................             159,800             171,400             177,200
General/Family Practice.............................             142,500             136,700             138,700
Internal Medicine...................................             150,200             154,100             165,600
Surgery.............................................             182,900             204,100             208,000
Pediatrics..........................................             130,900             152,100             140,600
OB/GYN..............................................             200,300             194,500             196,400
Radiology...........................................             217,600             231,100             225,300
Psychiatry..........................................             138,700             142,500             154,100
Anesthesiology......................................             167,500             200,300             200,300
Pathology...........................................             208,000             219,500            215,700
----------------------------------------------------------------------------------------------------------------
*All figures are rounded to the nearest $100.


[[Page 45460]]



V. PPS for Capital-Related Costs

    In the May 19, 2003 proposed rule, we did not propose any changes 
in the policies governing the determination of the payment rates for 
capital-related costs for short-term acute care hospitals under the 
IPPS. However, for the readers' benefit, in this section of this final 
rule, we are providing a summary of the statutory basis for the PPS for 
hospital capital-related costs, the methodology used to determine 
capital-related payments to hospitals, and a brief description of the 
payment policies under the PPS for capital-related costs for new 
hospitals, extraordinary circumstances, and exception (regular and 
special) payments. (Refer to the August 1, 2001 IPPS final rule (66 FR 
39910) for a more detailed discussion of the statutory basis for the 
system, the development and evolution of the system, the methodology 
used to determine capital-related payments to hospitals both during and 
after the transition period, and the policy for providing regular and 
special exceptions payments.)
    Section 1886(g) of the Act requires the Secretary to pay for the 
capital-related costs of inpatient hospital services ``in accordance 
with a PPS established by the Secretary.'' Under the statute, the 
Secretary has broad authority in establishing and implementing the PPS 
for capital related costs. We initially implemented the capital PPS in 
the August 30, 1991 IPPS final rule (56 FR 43358), in which we 
established a 10-year transition period to change the payment 
methodology for Medicare hospital inpatient capital-related costs from 
a reasonable cost-based methodology to a prospective methodology (based 
fully on the Federal rate).
    Federal fiscal year (FY) 2001 was the last year of the 10-year 
transition period established to phase in the PPS for hospital 
inpatient capital-related costs. Beginning in FY 2002, capital PPS 
payments are based solely on the Federal rate for the vast majority of 
hospitals. The basic methodology for determining capital prospective 
payments based on the Federal rate is set forth in Sec.  412.312. For 
the purpose of calculating payments for each discharge, the standard 
Federal rate is adjusted as follows: (Standard Federal Rate) x (DRG 
Weight) x (Geographic Adjustment Factor (GAF)) x (Large Urban Add-on, 
if applicable) x (COLA Adjustment for hospitals located in Alaska and 
Hawaii) x (1 + DSH Adjustment Factor + IME Adjustment Factor, if 
applicable) Hospitals also may receive outlier payments for those cases 
that qualify under the thresholds established for each fiscal year that 
are specified in Sec.  412.312(c) of existing regulations.
    During the 10-year transition period, a new hospital (as defined at 
412.300(b)) was exempt from the capital PPS for its first 2 years of 
operation and was paid 85 percent of its reasonable costs during that 
period. Originally, this provision was effective only through the 
transition period and, therefore, ended with cost reporting periods 
beginning in FY 2002. As we discussed in the August 1, 2002 final rule 
(67 FR 50101), this payment provision was implemented to provide 
special protection to new hospitals during the transition period in 
response to concerns that prospective payments under a DRG system may 
not be adequate initially to cover the capital costs of newly built 
hospitals. Therefore, we believe that the rationale for this policy 
applies to new hospitals after the transition period as well, and in 
that same final rule, we established regulations under Sec.  
412.304(c)(2) that provide the same special payment to new hospitals 
for cost reporting periods beginning on or after October 1, 2002. 
Therefore, a new hospital, defined under Sec.  412.300(b), is paid 85 
percent of its allowable Medicare inpatient hospital capital-related 
costs through its first 2 years of operation unless the new hospital 
elects to receive fully prospective payment based on 100 percent of the 
Federal rate. (For more detailed information regarding this policy, see 
the August 1, 2002 IPPS final rule (67 FR 50101).)
    Regulations at Sec.  412.348(f) provide that a hospital may request 
an additional payment if the hospital incurs unanticipated capital 
expenditures in excess of $5 million due to extraordinary circumstances 
beyond the hospital's control. This policy was established for 
hospitals during the 10-year transition period, but we established 
regulations at Sec.  412.312(e) to specify that payments for 
extraordinary circumstances are also made for cost reporting periods 
after the transition period (that is, cost reporting periods beginning 
on or after October 1, 2001). (For more detailed information regarding 
this policy, refer to the August 1, 2002 Federal Register (67 FR 
50102).)
    During the transition period, under Sec. Sec.  412.348(b) through 
(e), eligible hospitals could receive regular exception payments. These 
exception payments guaranteed a hospital a minimum payment of a 
percentage of its Medicare allowable capital-related costs depending on 
the class of hospital (Sec.  412.348(c)). However, after the end of the 
transition period, eligible hospitals can receive additional payments 
under the special exceptions provisions at Sec.  412.348(g), which 
guarantees an eligible hospital a minimum payment of 70 percent of its 
Medicare allowable capital-related costs. Special exceptions payments 
may be made only for the 10 years after the cost reporting year in 
which the hospital completes its qualifying project, which can be no 
later than the hospital's cost reporting period beginning before 
October 1, 2001. Thus, an eligible hospital may receive special 
exceptions payments for up to 10 years beyond the end of the capital 
PPS transition period. Hospitals eligible for special exceptions 
payments were required to submit documentation to the intermediary 
indicating the completion date of their project. (For more detailed 
information regarding the special exceptions policy under Sec.  
412.348(g), refer to the August 1, 2001 IPPS final rule (66 FR 39911 
through 39914) and the August 1, 2002 IPPS final rule (67 FR 50102).)



VI. Changes for Hospitals and Hospital Units Excluded From the IPPS

A. Payments to Excluded Hospitals and Hospital Units (Sec. Sec.  
413.40(c), (d), and (f))

1. Payments to Existing Excluded Hospitals and Hospital Units
    Section 1886(b)(3)(H) of the Act (as amended by section 4414 of 
Pub. L. 105-33) established caps on the target amounts for certain 
existing hospitals and hospital units excluded from the IPPS for cost 
reporting periods beginning on or after October 1, 1997 through 
September 30, 2002. For this period, the caps on the target amounts 
apply to the following three classes of excluded hospitals or units: 
psychiatric hospitals and units, rehabilitation hospitals and units, 
and LTCHs.
    In accordance with section 1886(b)(3)(H)(i) of the Act and 
effective for cost reporting periods beginning on or after October 1, 
2002, payments to these classes of existing excluded hospitals or 
hospital units are no longer subject to caps on the target amounts. In 
accordance with existing Sec. Sec.  413.40(c)(4)(ii) and (d)(1)(i) and 
(ii), where applicable, excluded psychiatric hospitals and units 
continue to be paid on a reasonable cost basis, and payments are based 
on their Medicare inpatient operating costs, not to exceed the ceiling. 
The ceiling would be computed using the hospital's or unit's target 
amount from the previous cost reporting period, updated by the rate-of-
increase specified in Sec.  413.40(c)(3)(viii) of the regulations, and 
then multiplying this figure by the number of Medicare discharges. 
Effective for cost reporting

[[Page 45461]]

periods beginning on or after October 1, 2002, rehabilitation hospitals 
and units are paid 100 percent of the Federal rate. Effective for cost 
reporting periods beginning on or after October 1, 2002, LTCHs also are 
no longer paid on a reasonable cost basis but are paid under a DRG-
based PPS. As part of the PPS for LTCHs, we established a 5-year 
transition period from reasonable cost-based reimbursement to a fully 
Federal PPS. However, a LTCH, subject to the blend methodology, may 
elect to be paid based on a 100 percent of the Federal prospective 
rate. (Sections VI.A.3. and 4. of this preamble contain a more detailed 
discussion of the IRF PPS and the LTCH PPS.)
2. Updated Caps for New Excluded Hospitals