Introduction

CMS requires facilities with IME/GME teaching arrangements to maintain written agreements between hospitals and non-hospital sites stating that the hospital will incur “all or substantially all” of the costs of training in that non-hospital site.  The agreement must also indicate the amount of compensation provided by the hospital to the non-hospital site for supervisory teaching activities.  CMS in 72 Fed. Reg. 26950, dated May 11, 2007, modified the definition of “All or Substantially All of the Costs for the Training Program in the Non-Hospital Setting.”

Revised Definition of “All or Substantially All”

To address providers’ concerns over documentation requirements to support the “all or substantially all” requirement, CMS has established an alternative methodology that hospitals may choose to utilize in determining and paying for the teaching physicians costs attributable to non-hospital sites.  Hospitals are no longer required to pay 100% of the residents’ salaries and fringe benefits (including travel and lodging where applicable) plus the portion of the teaching physicians’ costs at the non-hospital site.  Instead effective for cost reporting periods beginning on or after July 1, 2007, a hospital is required to pay at least 90% of the total costs of the residents’ salaries and fringe benefits (including travel and lodging where applicable) and the portion of the teaching physicians’ salaries attributable to direct GME at the non-hospital site.

Exceptions for Solo Practitioners

A solo practitioner typically has no costs associated for teaching time if their compensation is based solely on the number of patients treated and which he or she individually bills and is reimbursed.  There is a presumption in a clinic or group practice however that the physician receives a predetermined payment amount, such as a salary, for their work at the non-hospital site.  This predetermined payment amount implicitly compensates the physician for all of their responsibilities including supervision of residents.

If a teaching physician in a group practice or clinic is receiving a predetermined salary for his or her activities, and included in his or her activities is the supervision of residents at the non-hospital site, then there is a cost associated with those activities.  If teaching physicians that are members of a group practice or clinic can document that their circumstances are similar to that of a solo practitioner in that they receive no predetermined salary and receive income solely from the patients they treat and the services for which they bill, the hospital may supply this documentation to the Medicare contractor during audit to rebut the presumption.

Variables in the Calculation

In revising the definition of  “all or substantially all” of the costs of the program at a non-hospital site, and in establishing a 90% threshold, there are several variables that are important in the methodology for determining the minimum amount of training program costs that a hospital must pay in order to count FTE residents training in a non-hospital site.

These variables are:

• teaching physicians’ salaries;
• residents’ salaries and fringe benefits (including travel and lodging where applicable);
• the number of hours per week that the teaching physician spends in direct GME (not billable patient care) activities in the non-hospital site; and
• the number of hours that a non-hospital site is open each week.

Alternative Methodology

Rather than utilizing actual costs, CMS has adopted an alternative methodology that hospitals may choose to use to calculate teaching physician costs in non-hospital sites.  Using this alternative methodology, to facilitate a less burdensome way for a hospital to calculate the teaching physician costs associated with GME training at the non-hospital site, CMS is permitting hospitals to use 3 hours per week as a presumptive standard number of hours that a teaching physician spends in non-patient care direct GME activities at a particular non-hospital site.  To determine the percentage of the average salary associated with the 3 hours the teaching physician is presumed to spend in nonpatient care direct GME activities, a hospital would divide 3 hours by the number of hours the non-hospital site is open each week.  Next, the hospital would multiply this percentage of time spent in nonpatient care direct GME activities by the national average salary of that teaching physician’s specialty to calculate the cost of the teaching physician’s direct GME time. The cost of the teaching physician’s direct GME time would then be added to the costs of the salaries and fringe benefits (including travel and lodging expenses, where applicable) of the FTE resident(s) rotating in that program to that non-hospital site to determine the GME costs for that program at that site.   Where non-hospital rotations are not for a full FTE (a whole year rotation), then the national average salary of the teaching physician would be prorated accordingly.  The cost of the residents’ salaries and fringe benefits (including travel and lodging where applicable) would already be reflective of an FTE count.

The hospital must pay at least 90% of these total GME costs for the program at that non-hospital site to count the resident(s) training there for direct GME and IME purposes.  If the hospital is already paying all, or even a portion of the residents’ salaries and fringe benefits (including travel and lodging where applicable), and if the amount that the hospital is paying for the residents’ salaries and fringe benefits (including travel and lodging where applicable) is equal to at least 90% of the GME costs at the non-hospital site (that is, the 90 % threshold), then the hospital would be considered to be incurring “all or substantially all” of the costs, and need not incur an additional amount for teaching physician compensation to be permitted to include the FTE residents training in the non-hospital site in its FTE count for purposes of direct GME and IME payments.

However, if the costs of the residents’ salaries and fringe benefits (including travel and lodging where applicable) does not equal at least 90% of the GME costs of the training program at the non-hospital site, then the hospital must incur an additional amount for teaching physician costs based on the national average salary information until it is incurring at least 90% of the GME costs for that non-hospital site program.  If the hospital does not meet the 90% threshold, the resident cannot be counted for IME/GME purposes.

Two Examples of the Alternative Methodology

Example 1: Assume one teaching physician is supervising one FTE resident in a non-hospital site for one residency year.  The national average published salary amount for that teaching physician’s specialty is $120,000, and he works in a clinic that is open 60 hours per week. Using the standard of 3 hours spent in nonpatient care direct GME activities per week, the teaching physician spends 5% of his time in GME activities (that is, 3/60 = 5%). To determine the cost of the teaching physician’s time, the hospital may make the following calculation: $120,000 x 0.05 = $6,000. This teaching physician’s cost is added to the resident’s salary and fringe benefits to calculate the cost of the training at the non-hospital site in the following manner: $6,000 [cost of one teaching physician] + $60,000 [actual cost of the FTE residents’ salary & fringe benefits] = $66,000.  To meet the new definition of “all or substantially all,” the hospital would be required to pay at least 90% of the costs of the training program at the non-hospital site, which in this example equals $59,400 (0.90 x $66,000).  Since in this case the cost of one FTE resident’s salary and fringe benefits is $60,000, the hospital could reach the 90% cost threshold by simply incurring the resident’s salary and fringe benefits during training at the non-hospital site.

Example 2: Assume one teaching physician is supervising one FTE resident in a non-hospital site for an entire residency year.  The national average published salary amount for that teaching physician’s specialty is $200,000, and she works in a clinic that is open 40 hours per week. Using the standard of 3 hours spent in nonpatient care direct GME activities per week, the teaching physician spends 7.5 % of her time in GME activities (that is, 3/40 =7.5%).  To determine the cost of the teaching physician’s time, the hospital may make the following calculation: $200,000 x 0.075 = $15,000.  This teaching physician’s cost is added to the resident’s salary and fringe benefits to calculate the cost of the training at the non-hospital site in the following manner: $15,000 [cost of one teaching physician] + $60,000 [actual cost of the FTE residents’ salary and fringe benefits] = $75,000. To meet the new definition of “all or substantially all,” the hospital would be required to incur at least 90% of the costs of the training at the non-hospital site, which in this example equals $67,500 (0.90 x $75,000). Since in this case the cost of one FTE resident’s salary and fringe benefits is $60,000, the hospital has not met the 90% threshold by only incurring the resident’s salary and fringe benefits. The hospital would have to incur at least an additional $7,500 of the cost ($67,500 - $60,000) to reach the 90% threshold to be permitted to count the FTE resident for IME and direct GME purposes.  Alternatively, the hospital could document the actual teaching physician cost using time or salary information specific to that teaching physician at that site, and use that amount to calculate 90% of the actual training program costs.

Explanation of Variables

In the following section, we discuss each variable in the methodology for determining the cost that a hospital must incur to count FTE residents training in non-hospital sites. 

1. National Average Physician Salary Data by Specialty
   
   CMS is adopting the American Medical Group Association’s  (AMGA) Medical Group Compensation and Financial Survey as the basis for the national average physician salary.  The data can be obtained via their website www.amga.org.  Providers should utilize the most recent survey data available as of the beginning of the hospital’s particular cost reporting year.  Additionally, CMS has instructed hospitals to use the single national average or median salary amount for each specialty, rather than consider geographic variations.  RCE amounts or other surveys are not acceptable.  In cases where no subspecialty salary amount is available in the AMGA data, hospitals should use the physician salary amount for the closest less-specialized form of that specialty.
   CMS has made available the physician specialty salary survey data on their website free of charge.  It can be obtained by going to www.cms.hhs.gov, and selecting the following:
   
   Medicare
   Acute Inpatient PPS
   Direct Graduate Medical Education (DGME)
   Downloads
   
   
   
   1. Determining Teaching Physicians’ Cost
      
      In determining the teaching physicians’ cost, the specialty of the teaching physician is the relevant criterion, not the specialty of the residents that the teaching physician is training in the non-hospital site.  Although it may not be a common occurrence, it is possible that residents could be receiving training in a non-hospital site from a teaching physician that is board certified in more than one specialty, but the residents are only receiving training in one of the specialties in which the physician is board certified.  In this case, the national average salary that should be used to determine the teaching physician’s cost should be the one for the specialty in which the teaching physician is training the residents.
      
   2. Multiple Teaching Physicians and Residents: 1:1 Resident to Teaching
      
   3. Physician Ratio
      
      To maintain administrative simplicity, CMS is allowing hospitals to apply a maximum of a 1:1 resident-to-teaching physician ratio “limit” in determining the total GME costs applicable to a program at a non-hospital site.  For example, if at the non-hospital site there are two teaching physicians and one FTE resident, the hospital may determine 90% of the total costs of the program using a 1:1 resident-to-teaching physician ratio, not a 1:2 resident-to-teaching physician ratio.  The 90% threshold would be based on the total cost of the one FTE resident (salary and fringe benefits, and travel and lodging where applicable) and one teaching physician (national average salary for the specialty multiplied by the percentage of time spent in nonpatient care direct GME activities). Similarly, if a hospital rotated 3 FTE residents in the same program to a particular non-hospital site with 7 physicians, unless the hospital documents otherwise, we would assume that all 7 physicians supervise the residents at some point during the training, but, for purposes of determining the 90% threshold, we assume that there are only 3 FTE residents being supervised by 3 teaching physicians. Accordingly, the 90% threshold would be based on the total cost of the 3 FTE residents’ salaries and fringe benefits (including travel and lodging where applicable) and 3 teaching physicians (national average salaries for the specialties multiplied by the percentage of time spent in nonpatient care direct GME activities).  In addition, the 1:1 limit may be applied to FTE fractions, as well. That is, if in the preceding example, 3.5 FTE residents were being supervised by 7 physicians, the 90% threshold would be determined based on the costs associated with a resident-to-teaching physician ratio of 3.5:3.5.
      
      In the case of multiple teaching physicians, hospitals must also consider that a particular non-hospital site may be staffed by physicians in different specialties.  In this case, we would still maintain the 1:1 resident-to-teaching physician limit, even if the teaching physicians are in different specialties, unless the hospital can document that the number of physicians actually teaching the residents is less than the number of FTE residents training at that non-hospital site.  Once the number of teaching physicians is established, the hospital would determine the national average salary for each of those teaching physicians from the national survey data, and then calculate the average national salary of the mix of physician specialties in the practice to be used in computing the 90% threshold.  For example, assume that 3 FTE residents are rotating to an orthopedic surgery practice staffed by a total of 7 physicians; 4 are orthopedic surgeons, and 3 are diagnostic radiologists.
      
      Again, unless the hospital documents otherwise, we would assume that all 7 physicians supervise the residents at some point during their rotation to this practice. First, the hospital would access the national average salary for orthopedic surgeons (assume $400,000), and the national average salaries for diagnostic radiologists (assume $412,000). Then, the hospital would calculate the average salary for these physicians as follows: [($400,000 x 4) + ($412,000 x 3)]/7 = $405,143. Next, the 1:1 resident-to-teaching physician ratio would be applied, such that for purposes of determining the 90 % threshold, there would be 3 FTE residents and 3 teaching physicians. Since the 3 teaching physicians are not in the same specialty, the hospital would multiply the average salary cost of $405,143 by 3 to get the total teaching physician salaries for the training program at that site ($405,143 x 3 = $1,215,429). The hospital would then multiply $1,215,429 by the percentage of time spent by the teaching physicians in nonpatient care direct GME activities (that percentage is 3 hours divided by the number of hours the practice is open during a week) to determine the teaching physician GME cost for the training program at that site. This teaching physician cost is then added to the salaries and fringe benefits (including travel and lodging where applicable) of the 3 FTE residents to determine the GME cost of the program at that practice, and the hospital must ensure that it incurs at least 90 % of that GME cost to count the 3 FTE residents training at the non-hospital site.
      
      If there are several physicians in a non-hospital site, we would assume that they all supervise the residents at some point during the residents’ training.  However, it may be that in fact only some of the physicians actually supervise the residents, while other physicians are not involved in the training program at all. The hospital may wish to document that only certain physicians are involved in the training program (to more accurately represent the structure and costs of the training program in a particular non-hospital site). Such documentation would increase the number of residents relative to teaching physicians that is used to calculate the teaching physician costs. That is, using the example above where the resident-to-teaching physician limit was presumed to be 3:3, since there were actually 3 FTE residents and 7 physicians, if the hospital can document that only 2 physicians supervised the residents (and the other 5 physicians were not involved in the GME program at all), then the resident-to-teaching physician ratio would be 3:2.
      
2. Residents’ Salaries and Fringe Benefits
   
   The second variable in the methodology for determining the costs of a program at a non-hospital site is the salaries and fringe benefits (including travel and lodging where applicable) of the FTE residents that are rotating to a particular non-hospital site.  The hospital must use the actual cost of each FTE resident’s salary and fringe benefits (including travel and lodging where applicable).  Malpractice costs should not be included as a fringe benefit in the calculation of the 90 % threshold.  The only travel and lodging costs that are applicable are the additional travel and lodging costs that a hospital incurs due to the fact that a resident is training at a non-hospital site. For example, if a resident needs to travel long distance to another part of the state, and is staying in a hotel for the duration of the non-hospital site training, the costs of the traveling and accommodations would be costs that the hospital must incur and include in the determination of the 90% threshold. However, expenses that are normally incurred when the resident trains at or nearby the hospital, such as commuting and living expenses, would not be applicable.
   
   In addition, the cost of the residents will vary by specialty and by program year. Furthermore, as with current policy, the total residents’ costs will be based on the FTE number rotating to a particular non-hospital site in a cost reporting period, not the number of individuals actually training in a non-hospital site. 
   
3. The Number of Hours Spent in Nonpatient Care Direct GME Activities in a Week
   
   The third variable used in the determination of the costs of a training program at a non-hospital site is the amount of time that the teaching physician(s) spends on direct GME nonpatient care activities in a week.  Determination of the teaching physician costs to the non-hospital site is dependent upon the teaching physician’s salary and the percentage of time he or she devotes to activities related to non-billable direct GME activities at the non-hospital site (such as conferences, practice management, lectures, and administrative activities like resident evaluations).  In an effort to eliminate the documentation burden on physicians of keeping track of the amount of time they spend in nonpatient care direct GME activities in the non-hospital site, an alternative option is being made available.
   
   Under the proxy methodology, a presumed standard number of three hours spent by teaching physicians in nonpatient care direct GME activities in every non-hospital site will be applied. The 3 hour standard would be used in all cases in the formula for determining the teaching physician costs at all non-hospital sites, regardless of the specialty of the residents or the number of teaching physicians or residents training at that non-hospital site.
   
   However, a hospital always has the option of documenting and paying for at least 90% of the costs of a program at a non-hospital site using the teaching physician’s actual salary and information on the time spent in nonpatient care direct GME activities, rather than utilize the proxy.
   
4. The Number of Hours the Non-Hospital Site is Open per Week
   
   The fourth variable will vary depending on the specific non-hospital site. This variable is the number of hours that the non-hospital site is open each week.  As indicated above the numerator will always be 3 hours, and the denominator will vary depending on the non-hospital site.  In determining the number of hours a clinic is open per week, it is not the actual hours the non-hospital site is open per week, but instead, you are to utilize the “posted” or advertised hours. Therefore, the fact that a non-hospital site might be closed several days in a year on legal holidays, for example, would not affect the denominator. That is, if a non-hospital site’s posted hours are 9 a.m. to 5 p.m. from Monday through Friday, then the denominator would be 40 hours, even if that site was closed for a day(s) for a holiday or some other reason. The hospital may obtain the non-hospital site’s posted or advertised hours of operation as documentation to support the number of hours used in the denominator of the teaching time proxy.
   
   CMS has instituted a cap of 7.5% on the teaching physician cost ratio, such that a hospital need not employ more than 7.5% of the teaching physician cost in calculating the amount of payment necessary to meet the 90% threshold.  However, in adopting this policy, application of the 7.5% cap must always be after a hospital prorates the teaching physician cost to reflect the amount of FTE time that the residents are in the particular non-hospital site per year. 
   
   For example, if FTE residents rotate throughout the year to a non-hospital site that is open 40 hours per week, then the percentage of time spent by the teaching physician(s) in nonpatient care direct GME activities throughout the year at that site is 3/40 = 7.5%. If FTE residents rotate to that non-hospital site for only a portion of a year, then the ratio of 3/40 would be further multiplied by the percentage of the year that the FTE residents train there. For example, if the FTE residents only rotate to this non-hospital site for 3 months of the year, then the percentage of time that the teaching physician(s) spends on nonpatient care direct GME activities at that site equals (3/40 x 0.25 = 1.9%)
   
   If a hospital chooses not to use the proxies, then the hospital should use actual data specific to the teaching physician in the particular non-hospital site, and not an arbitrary amount or information from local surveys or broader samples.  However, it would be acceptable for the physician to provide to the hospital a signed document specifying, based on actual records kept, the amount of such time spent with the residents.  Similar to the documentation that was historically required of hospitals to allocate teaching physician costs between Part A and Part B and between operating costs and direct medical education costs, if the physician is supervising residents in the non-hospital site throughout the academic year, the physician may complete a 2-week time study at two different points during the academic year (that is, two separate 2-week time studies).  If a physician only supervises residents in the non-hospital site for the equivalent of a month or less in an academic year, then the physician may complete a 1 week time study. The percentage of time a teaching physician spends with or on behalf of the residents in nonpatient care direct GME activities over the course of the time study may then be extrapolated to apply to the rest of the academic year.
   
   Accordingly, CMS is not requiring that time studies completed by teaching physicians in non-hospital sites for the purpose of determining the 90% cost threshold meet the requirements in CMS Pub. 15-1, Section 2313.2.   Since the teaching physician may not know the percentage of time spent on nonpatient care direct GME activities at the time the written agreement between the hospital and the non-hospital site is being entered into (since the written agreement must be in place before the rotation begins), the written agreement can be made based upon either the 3-hour per week proxy or an estimated percentage (based on the prior year’s rotations, if applicable), and the percentage may be modified during the academic year if necessary.
   
   Further, the teaching physician, or the non-hospital site employer, and the hospital should modify the calculation of the 90% cost threshold and the written agreement in order to reflect the actual percentage by June 30 of that academic year. The source documentation used to determine the amount of teaching physician compensation should be made available to the Medicare contractor upon request during audit.

Additional Example

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

Other Issues To Be Considered

Under the existing regulations at 42 CFR § 413.78(e), a hospital is permitted to count residents training in non-hospital sites only if the residents spend their time in patient care activities, and the hospital must comply with either of the following: (a) It must pay all or substantially all of the costs of the training program in the non-hospital site by the end of the third month following the month in which the training in the non-hospital site occurred; or (b) it must have a written agreement with the non-hospital site that states that the hospital will incur the cost of the resident’s salary and fringe benefits while the resident is training in the non-hospital site and the hospital is providing reasonable compensation to the non-hospital site for supervisory teaching activities. The written agreement must indicate the compensation the hospital is providing to the non-hospital site for supervisory teaching activities.

If a hospital chooses to make concurrent payments; that is, pay the training costs by the end of the third month following the month in which the training occurred, then the hospital must be able to document for audit purposes that the concurrent payments it makes reflect “all or substantially all” of the costs, in accordance with the new definition at 42 CFR § 413.75(b).

Alternatively, if the hospital chooses to maintain a written agreement with the non-hospital site the agreement must be in place before the residents begin training at a non-hospital site. 42 CFR § 413.78(f) requires that the written agreement must indicate that the hospital will incur at least 90% of the total of the costs of the resident’s salary and fringe benefits (including travel and lodging where applicable) while the resident is training in the non-hospital site and the portion of the cost of the teaching physician’s salary attributable to direct GME.  The agreement should also specify the total compensation amount the hospital will incur to meet the 90% “all or substantially all” threshold, and whether this amount reflects only residents’ salaries and fringe benefits (including travel and lodging where applicable), or reflects an amount for teaching physician compensation as well. The written agreement should specify the total amount of non-hospital site training costs the hospital will incur and specify what costs are included in that amount because the hospital would need to determine up front the amount it must pay to meet the 90% threshold and incur “all or substantially all” of the cost in accordance with the definition.

Regardless of whether a hospital chooses to make concurrent payments to the non-hospital site, or to have a written agreement, the hospital must demonstrate that it is paying for at least 90% of the costs of each program at each non-hospital site according to the following formula (although actual data may be used in place of the proxies):

0.90 x [(sum of each FTE resident’s salary + fringe benefits (including travel and lodging where applicable)) plus the portion of the teaching physician’s compensation attributable to nonpatient care direct GME activities].  The portion of the teaching physician’s compensation attributable to nonpatient care direct GME activities may be calculated as follows: (3/number of hours non-hospital site is open per week) x (national average salary for each teaching physician).

If there are no teaching costs (because, for example, the residents are rotating to a non-hospital site where the teaching physician is a solo practitioner), then the written agreement should indicate that the specified compensation amount reflects only residents’ salaries and fringe benefits (including travel and lodging where applicable) because there are no teaching physician costs (since the teaching physician is a solo practitioner).

If the hospital does choose to have a written agreement with the non-hospital site, the hospital must, at a minimum, liquidate the costs identified in the written agreement in accordance with the regulations 42 CFR § 413.100(c)(2)(i).

In addition, under current policy, a hospital may choose to provide non-monetary, in-kind compensation rather than provide direct financial compensation to the non-hospital site for supervisory teaching activities. Under the new definition of “all or substantially all,” a hospital would still be permitted to provide in-kind compensation to the non-hospital site, but, as under current policy, the hospital must be able to document that the value of the in-kind compensation is at least equivalent monetarily to the portion of the actual or proxy-based costs for that teaching physician attributable to nonpatient care direct GME activities. That is, the hospital must show that the value of in-kind compensation is sufficient to meet the 90% threshold using the formula stated above in this section.

CMS has modified their policy with respect to written agreements (for cost reporting periods beginning on or after July 1, 2007).  Current policy requires that the written agreement be in place prior to the time that the residents begin training in the non-hospital site (that is, signed by both the hospital and the non-hospital site). Since residents rotate to various non-hospital sites at different points in the residency year, a written agreement may or may not have to be in place with a particular non-hospital site by July 1. Rather, the agreement should be in place by the day before the rotation is scheduled to begin. For example, if a resident is scheduled to rotate to Clinic A on July 1, then the written agreement between the hospital and Clinic A must be in place by June 30 (that is, the day before July 1, not the end of the following residency year). However, if residents first rotate to Clinic B on December 1, then the written agreement between the hospital and Clinic B would have to be in place by November 30.

Additionally, hospitals are permitted to modify the 90% threshold calculations in their written agreements by the end of the academic year (that is, June 30) to reflect that the hospital is meeting the requirement to incur at least 90% of the costs associated with the actual training program rotations. If the hospital opts to use actual data and not the proxies, the hospital may use the prior year’s cost amounts as a placeholder upon entering into the written agreement, and must modify the agreements by June 30 of that residency year to properly reflect the actual costs that the hospital must incur in accordance with the 90% threshold for “all or substantially all” of the costs of the training program in the non-hospital setting. In addition, in the event that hospitals send residents to unanticipated or originally unscheduled rotations in non-hospital sites, the hospitals may make their “best estimate” by the day before the rotations occur (the hospital may use the prior year’s rotation experiences as a model), and must make modifications by the end of the academic year to ensure that they have properly met the 90% threshold.

Assistance:

We strongly recommend that you review in detail the May 11, 2007 federal register pages 26949-26976.  Should you have questions on this bulletin, please contact: