Introduction

Graduate medical education (GME) training takes place in teaching hospitals and ambulatory offices under the supervision of practicing faculty physicians. These institutions make up approximately 6% of all of the hospitals within the United States.^1–3 Michigan alone has 52 GME training institutions or hospitals, ranking sixth in the nation, and educating just under 5000 residents per year.^4,5 In addition to providing a large amount of complex and acute patient care, one-fifth or more of all care that takes place in US hospitals occurs at teaching hospitals.^2,6

These training sites provide a safety net for the uninsured and underserved, as well as many advanced services with the latest technology that nonteaching hospitals cannot provide.^1,2,6 These include burn units, pediatric and neonatal intensive care units, transplant services, and cardiac surgery.¹ Teaching hospitals also play a pivotal role in ensuring the physician workforce for the future by providing the clinical environment for training and supporting scholarly activity in the form of research, patient safety, and quality improvement projects.⁶ These critical functions do not come without a price. Factoring in other costs to teaching hospitals outside of resident/fellow salaries and benefits (eg, more tests ordered, longer patient length of stay), the average cost of GME training is approximately $115,000 to $150,000 per resident or fellow per year.⁷

While GME is funded by multiple sources, the majority of dollars come from federal and state funding in the form of Medicare and Medicaid. There is bipartisan support for reducing Medicare costs, and cuts to GME funding are regularly proposed as a means to achieve this goal.^8,9 In addition, some states have reduced or proposed reductions in Medicaid support of GME, while others have eliminated this source of funding altogether.⁵

A reduction in GME funding could affect the number of newly trained physicians, which could exacerbate a projected physician shortage that is estimated to be 91 500 nationwide by 2020.¹ An additional consideration is that up to 33% of current practicing physicians are projected to retire in the next 5 years, further widening the gap between health care needs and the number of physicians available.^3,10

One method to provide supporting evidence for continuing Medicaid funding for GME is to demonstrate the value of retaining physicians within the community/state in which they completed their GME training. There are limited data available in the literature addressing in-state retention,^11–13 particularly as it relates to a single GME sponsoring institution. Thus, we analyzed the in-state retention rate of a single Michigan-based GME sponsoring institution by looking at the practice locations of graduates from its training programs over a 15-year period.

Methods

Results

The data set included 1296 graduates, and of these, 1168 (90.1%) met the criteria for inclusion. There were 128 residents and fellows who went on to another training program, who were still in training at the time of data collection, and subsequently were excluded from review. Demographic and educational data are shown in table 1. The majority of graduates were men, and the average age at the time of graduation was 33 years. Primary care program graduates (defined as family medicine, internal medicine, internal medicine-pediatrics, and pediatrics) made up slightly more than 40% of graduates. About one-third of the graduates attended medical school in Michigan, and one-fourth completed both their undergraduate and MD/DO degree in Michigan.

table 1 Demographic and Educational Information

View Fullscreen

An examination of practice location immediately after graduation was performed, excluding 391 graduates who went into a fellowship or another residency program following graduation (table 2). Michigan as a practice location immediately after graduation was the choice for more than half of the graduates (396 of 777), and more than half of primary care graduates (247 of 415) located their practice in Michigan immediately after training.

table 2 Practice Location After Graduation^a

View Fullscreen

Just under half (546 of 1168) of our sample practiced in Michigan at some point during their career (table 3). Almost 80% (279 of 358) of the graduates who completed medical school in Michigan practiced in Michigan, and nearly 90% (245 of 279) of that cohort also had completed a bachelor's degree in Michigan. Of the primary care graduates, almost 60% (284 of 476) practiced in Michigan during their career as a physician. Graduates who were practicing in Michigan in 2014 were further reviewed, excluding graduates who had died prior to 2014 or who had reentered training during 2014. As of fall 2014, 41.3% (479 of 1161) of GRMEP graduates were currently practicing in Michigan. Of this group, 36.7% (176 of 479) were employed by 1 of the teaching hospitals at which they completed their training. For graduates who practiced in Michigan immediately after graduation, 85.8% (315 of 367) were practicing in Michigan in 2014 (the 24 graduates from 2014 who were practicing in Michigan after graduation were excluded from this percentage).

table 3 Michigan as a Practice Location

View Fullscreen

Discussion

Our data show that over 45% of graduates from GRMEP GME programs have practiced medicine in Michigan at some point following graduation. Additionally, almost 60% of primary care graduates have practiced medicine within the state. This information on retention indicates that 1 of the advantages of providing GME training is the development and education of future physicians who are likely to practice in the state in which they trained.

These data are important to consider, as in recent years there have been several bills introduced in the US Congress to reduce GME payments.⁸ A reduction in, or elimination of, funding for GME programs would put more of a financial burden on teaching hospitals. This could adversely affect the supply of US physicians by forcing teaching institutions to reduce or cut GME programs and/or eliminate services not available elsewhere in the community.^3,14 These types of reductions, in an era when health care access is an issue, may adversely affect the quality of medical care in the nation.^3,10

Our findings suggest that more attention may need to focus on the importance of in-state retention of GME graduates. Teaching hospitals may experience significant savings through the recruitment of graduates from their programs, through reduced recruitment costs, faster integration into practice, and increased familiarity with both the institutional culture and electronic health record.¹⁵ In addition, a method to assess the value of in-state retention of GME graduates for state governments could be to evaluate the number of physicians who train at GME teaching institutions who stay or return to practice in the same state.

According to the Association of American Medical Colleges' Center for Workforce Studies, 44% of Michigan GME graduates were considered active physicians in Michigan in 2014, and 67% of active physicians who completed both GME and undergraduate medical education in Michigan practiced in Michigan in 2014.⁴ Our data showed that almost 80% of GRMEP GME graduates who practiced in Michigan after training graduated from a Michigan-based medical school, and that nearly 90% of these also had obtained a bachelor's degree in Michigan. New York showed similar findings in a study conducted in 2014, showing that for physicians who had confirmed practice plans, just under half (45%) were planning to practice in New York. A high percentage (80%) of the respondents who were staying in New York had strong ties to the state, having attended both high school and medical school in New York.¹¹

The similarity of findings is encouraging, suggesting our findings may be generalizable to other training programs in Michigan, as well as other states. Our results may be useful to other institutions who wish to demonstrate the inherent value at the community and state level in supporting GME training at teaching hospitals.

Limitations of the study include its retrospective nature and the fact that not all graduates had complete data for all study variables. In addition, there were 128 residents and fellows who were still in training at the time of data collection and could not be included in our data set. Although this study addressed in-state retention, the focus was not on the reason for retention or the geographic distribution of graduates. Further studies should address these important issues.

Conclusion

Our findings show that nearly 50% of graduates from a single Michigan GME institution chose Michigan as a practice location, and graduates were highly likely to practice in Michigan if they completed their pre-GME education in the state.