Unfilled in emergency medicine: An analysis of the 2022 and 2023 Match by program accreditation, ownership, and geography

Abstract

Background

The two most recent National Resident Matching Program (NRMP) Match cycles saw a high number of initially unfilled emergency medicine (EM) residency positions. We sought to identify the risk of EM residency program characteristics including accreditation duration, primary clinical site ownership status, and geography pertaining to not initially filling all positions.

Methods

We performed a repeated cross‐sectional observational study of EM residency programs participating in the 2022 and 2023 NRMP Match cycles and used publicly available data from the NRMP, the Accreditation Council for Graduate Medical Education, the Centers for Medicare & Medicaid Services, and the U.S. Department of Housing and Urban Development. Our primary outcome was the proportion of EM residency programs that did not initially fill positions, with analyses stratified by accreditation duration (>5 or ≤5 years), primary clinical site ownership status, and geographic core‐based statistical areas (CBSAs).

Results

A total of 219 of 2921 (7.5%) positions in the 2022 Match and 554 of 3010 (18.4%) positions in the 2023 Match were initially unfilled. Over the 2‐year period, EM residency programs accredited within the past 5 years had more than double the risk (relative risk [RR] 2.08, 95% confidence interval [CI] 1.69–2.57, chi‐square p < 0.001) of not filling all positions compared to those accredited more than 5 years previously. EM residency programs with a primary clinical site under for‐profit ownership had a 50% greater risk of not filling all positions when compared to those under nonprofit or governmental ownership (RR 1.50, 95% CI 1.14–1.98, chi‐square p = 0.009). In 2023, several CBSAs had a high number of both offered and unfilled positions.

Conclusions

EM residency programs accredited within the past 5 years or those with a primary clinical site under for‐profit ownership had a greater risk of not filling all positions within the past two Match cycles.

INTRODUCTION

The 2022 and 2023 National Resident Matching Program (NRMP) Match cycles saw an unprecedented amount of initially unfilled emergency medicine (EM) residency positions prior to the Supplemental Offer and Acceptance Program (SOAP). ¹ Interest in EM has declined considerably over the past two Match cycles; the proportion of medical students applying for EM residency positions dropped 16.8% from 2021 to 2022 and further declined an additional 18.1% from 2022 to 2023. ² Suggested contributors to the Match results include the projected workforce surplus by 2030, continued growth in residency program numbers, and escalating levels of burnout and attrition among emergency physicians. ³ , ⁴

Understanding the relationship between residency program–specific characteristics and their relation to Match results would provide important foundational data to the NRMP, the Accreditation Council for Graduate Medical Education (ACGME), EM residency program leadership, and potential EM residency applicants. However, available literature has primarily assessed Match results from the perspective of the EM residency applicant, including geographic mobility and the impact of the COVID‐19 pandemic. ⁵ , ⁶ , ⁷ To date, no existing work describes the characteristics of EM residency programs that successfully fill their positions before the SOAP versus those that do not. Therefore, we sought to identify the risk of EM residency program characteristics including accreditation duration, primary clinical site ownership status, and geography pertaining to not filling positions before the SOAP.

METHODS

Study design and data sets

We performed a repeated cross‐sectional observational study of EM residency programs participating in the 2022 and 2023 NRMP Match cycles. This study was deemed exempt by the institutional review board and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. ⁸

We relied on four publicly available sources of data. First, we used NRMP data to identify EM residency programs participating in the 2022 and 2023 Match cycles along with information on program location and total number of pre‐SOAP positions offered, filled, and unfilled. ⁹ Second, we similarly used the ACMGE's Accreditation Data System (ADS) to identify ACGME‐accredited programs along with information on program location, original accreditation date, sponsoring institution, participating clinical sites, and the number of clinical months per year allocated to each participating site (thereby allowing identification of primary clinical site); ADS is the primary site for program directors and institutions to submit information related to compliance with ACGME accreditation standards. This information was obtained by direct request to the ACGME. Third, we used data from cost report files from the Centers for Medicare & Medicaid Services to identify the ownership status of the primary clinical site for each EM residency program. Consistent with prior work, ¹⁰ we categorized ownership status as nonprofit, for‐profit (proprietary), or governmental. Fourth, we used crosswalks from the U.S. Department of Housing and Urban Development to link zip codes to the 929 core‐based statistical areas (CBSAs) across the country. Inclusive of metropolitan statistical areas and micropolitan statistical areas, CBSAs are an important level of analysis used by federal statistical agencies to understand demographic, economic, and health care characteristics of different regions.

Outcomes

Our primary outcome was the proportion of EM residency programs that did not completely fill positions before the SOAP. We stratified analyses by accreditation duration (>5 or ≤5 years) and primary clinical site ownership status. We compared for‐profit ownership status to either nonprofit or governmental ownership status. Geographically, we then identified the proportion of EM residency positions that did not fill within individual CBSAs, thereby detecting potential oversaturation within certain CBSAs.

Data analyses

We used descriptive statistics to describe EM residency programs within the 2022 and 2023 Match. We used the two‐sided chi‐square test for statistical analyses and present relative risk ratios between relevant groups. We performed data management and analyses using Stata (version 16, StataCorp).

RESULTS

The 2022 Match included 277 EM residency programs, offered 2921 positions, and resulted in 69 programs not filling a total of 219 (7.5%) positions before the SOAP. The 2023 Match included 287 EM residency programs, offered 3010 positions, and resulted in 132 programs not filling a total of 554 (18.4%) positions before the SOAP.

In 2022, a total of 41 (46.1%) of the 89 programs accredited within the 5 years prior to the Match did not initially fill all positions. This reflected a 3.21 (2.11–4.89, chi‐square p < 0.001) times greater risk than the 26 (14.4%) of the 181 programs accredited more than 5 years prior. In 2023, a total of 49 (72.1%) of the 68 programs accredited within the past 5 years did not initially fill positions during the Match. This reflected a 1.85 (1.48–2.32, chi‐square p < 0.001) times greater risk than the 81 (38.9%) of the 208 programs accredited more than 5 years prior (Figure 1).

FIGURE 1.

Proportion of unfilled positions in the 2022 and 2023 Match for individual EM residency programs, stratified by years since accreditation. Violin plots are shown, with individual dots representing programs. In the 2022 Match, 270 EM residency programs had reliable accreditation date data. In the 2023 Match, 276 EM residency programs had reliable accreditation date.

In 2022, a total of 30 (10.8%) programs were identified as having a for‐profit primary clinical site with 13 (43.3%) not initially filling during the Match. This reflected a 1.91 (1.20–3.06, chi‐square p = 0.014) times greater risk than the 56 (22.7%) programs with a nonprofit or governmental primary clinical site. In 2023, a total of 31 (10.8%) programs were identified as having a for‐profit primary clinical site, with 18 (58.1%) not initially filling during the Match. This reflected a 1.30 (0.94–1.81, chi‐square p = 0.153) times greater risk than the 114 (44.5%) programs with a nonprofit or governmental primary clinical site (Table S1).

In 2023, a total of 287 EM residency programs were present within 144 CBSAs, with the number of positions offered ranging from five to 383 (mean 20.9). CBSAs with at least 50 positions offered and the highest unfilled proportions included: Detroit‐Warren‐Dearborn (40.7% unfilled), Miami‐Fort Lauderdale‐West Palm Beach (30.3% unfilled), and Philadelphia‐Camden‐Wilmington (18.7% unfilled; Figure S1).

DISCUSSION

In this cross‐sectional study on pre‐SOAP Match cycle outcomes, we identified three main findings. First, EM residency programs accredited within the past 5 years had more than double the risk of not filling all positions when compared to those accredited more than 5 years previously. Second, EM residency programs with a primary clinical site under for‐profit ownership had a 50% greater risk of not filling all positions when compared to those under nonprofit or governmental ownership. Third, in 2023, several CBSAs had a high number of both offered and unfilled positions.

This work is the first of its kind, identifying key characteristics—recent accreditation and for‐profit ownership status of the primary clinical site—that confer an increased risk of EM residency programs not filling all Match positions prior to the SOAP. Despite clear warning signs from the 2022 Match cycle, eight new EM residency programs were identified on ACGME's ADS to be newly accredited during the 2022–2023 academic year, with four having received accreditation even after the results of the 2023 Match. In light of our findings, these newly accredited programs are at a considerably increased risk of not filling all positions in subsequent Match cycles. Furthermore, for‐profit entity involvement in EM graduate medical education has exhibited considerable growth within the past 7 years, ⁸ with our work potentially reflecting applicant's recognition of fiscal instability within for‐profit programs given the recent bankruptcy filing of a large for‐profit EM physician‐staffing group that sought to create several new residencies. ¹¹

Within the context of the two recent EM Match cycles, these findings provide evidence of a pressing need to right‐size the supply and demand of practicing emergency physicians. As evidenced by anesthesiology in the 1990s when the proportion of residency positions filled nadired at 35%, yet rebounded to >95% for all years after 2005, interest in different medical specialties ebbs and flows. ¹² With the ACGME's work primarily devoted to approving programs that meet its accreditation standards, an independent regulatory body or specialty‐specific residency accreditation process may be beneficial in the current landscape to curb the current oversupply of EM residency positions and avoid new program residency accreditation in oversaturated CBSAs.

LIMITATIONS

There are limitations of our study. With the SOAP subsequently having filled many positions, the long‐term ramifications of a high number of unfilled positions in the initial Match remains unclear. Second, our ownership analysis is hospital‐centric, with this characteristic being distinct from residency ownership or affiliation. However, our approach and identified proportions were consistent with prior EM residency program characterizing the distribution of for‐profit, nonprofit, and governmental ownership. ¹⁰ Finally, program‐level analyses also may not detect patterns at the applicant‐level, with our work absent of the degree to which programs ranked and received applicants.

CONCLUSIONS

In summary, we identified that emergency medicine residency programs accredited within the past 5 years or those with a primary clinical site under for‐profit ownership had a marked greater risk of not filling all positions within the last two Match cycles prior to the Supplemental Offer and Acceptance Program. These findings warrant a specialty‐wide reflection and reaction to ensure that the future workforce is adequate and that emergency medicine continues to attract talented medical school graduates to sustain the future of emergency care across the nation.

AUTHOR CONTRIBUTIONS

Study concept and design: Cameron J. Gettel, Christopher L. Bennett, and Arjun K. Venkatesh. Acquisition of the data: Cameron J. Gettel and Christopher L. Bennett. Analysis and interpretation of the data: Cameron J. Gettel, Christopher L. Bennett, Craig Rothenberg, Jessica L. Smith, Katja Goldflam, Wendy W. Sun, and Arjun K. Venkatesh. Drafting of the manuscript: Cameron J. Gettel and Arjun K. Venkatesh. Critical revision of the manuscript for intellectual content: Cameron J. Gettel, Christopher L. Bennett, Craig Rothenberg, Jessica L. Smith, Katja Goldflam, Wendy W. Sun, and Arjun K. Venkatesh. Statistical expertise: Cameron J. Gettel, Craig Rothenberg, and Arjun K. Venkatesh. Acquisition of funding: Cameron J. Gettel and Arjun K. Venkatesh.

CONFLICT OF INTEREST STATEMENT

JLS serves on the Board of Directors of the 2023–2024 Council of Residency Directors in Emergency Medicine (CORD); authorship is as an individual and not representative of the organization. The other authors declare no conflicts of interest.

Supporting information

Figure S1.

Table S1.

Gettel CJ, Bennett CL, Rothenberg C, et al. Unfilled in emergency medicine: An analysis of the 2022 and 2023 Match by program accreditation, ownership, and geography. AEM Educ Train. 2023;7:e10902. doi: 10.1002/aet2.10902