Clinical duties and compensation for U.S. emergency medicine fellows at academic centers: A descriptive, cross‐sectional survey

Abstract

Objective

Fellowship training is increasingly popular among residency graduates and critical to the advancement of academic emergency medicine (EM). Little is known about the clinical hours worked and financial compensation received by fellows during training. We sought to describe the clinical duties and financial compensation of EM fellows at U.S. academic centers.

Methods

This cross‐sectional study surveyed U.S. academic EM department administrators who were members of the Society for Academic Emergency Medicine's Academy of Administrators in Academic Emergency Medicine (AAAEM) regarding their fellowship programs and fellows. We electronically distributed the validated survey instrument to 73 member sites between October 2022 and January 2023. Survey domains included fellow and fellowship demographics, base and total annual clinical hours, and base and total annual compensation. We calculated descriptive statistics and compared fellows by accreditation (Accreditation Council for Graduate Medical Education [ACGME] or non‐ACGME) using chi‐square and Wilcoxon rank‐sum testing. We conducted a secondary analysis of base and total salary by gender and accreditation using Wilcoxon rank‐sum testing.

Results

We received 38 institutional responses (response rate 52%), which represented 217 individual fellows. Nearly three‐fourths (n = 158, 72.8%) of fellows enrolled in non‐ACGME fellowships, worked 33% more base hours annually than ACGME fellows (median 571 h vs. 768 h, p < 0.001), and received base compensation 20% higher than ACGME fellows ($88,540 vs. $70,777, p < 0.001). Accounting for additional compensation, the median total annual compensation for non‐ACGME fellows remained 11% higher than ACGME fellows ($105,000 vs. $93,853, p = 0.004). We observed no significant differences salary when stratified by gender.

Conclusions

Most EM fellows at U.S. academic institutions enrolled in non‐ACGME fellowships with significantly higher base hours and financial compensation than ACGME fellowships. These results represent the first description of the clinical hours and financial compensation of academic EM fellows and should be considered in ongoing benchmarking efforts by AAAEM.

INTRODUCTION

Background

Fellowship training plays a critical role in the advancement of academic emergency medicine (EM) through the development of EM subspecialties and generation of knowledge specific to these niches. ¹ , ² Fellowship training options are increasing for EM‐trained residency graduates: the most recent American Board of Emergency Medicine report on Accreditation Council for Graduate Medical Education (ACGME)‐accredited EM fellowships estimated that there are 157 programs in 2023 (up from 151 in 2022) and continued growth is expected in years to come. ³ , ⁴

Importance

Despite the growth of EM fellowship training, little is known about the EM physicians choosing to pursue fellowship training, the clinical duties performed, and financial compensation received. Data available through the American Board of Emergency Medicine do not include fellows in non–ACGME‐accredited EM fellowships. ³ , ⁴ A limited number of studies have described demographics of specific subsets of EM fellows, including emergency medical services (EMS), ultrasound, global EM, and simulation. ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ An even smaller number of studies have examined resident‐identified pertinent considerations, including salary, when considering a fellowship training program. ¹² There is a paucity of literature characterizing the clinical and nonclinical work hours as well as compensation for the hours worked while in fellowship.

Goals of this investigation

We sought to describe the clinical duties and financial compensation of EM fellows at U.S. academic centers. This work will provide necessary information for departmental benchmarking and program growth.

METHODS

Study design, setting, and participants

This cross‐sectional study surveyed U.S. academic EM departments regarding their available fellowship programs and enrolled fellows. The survey population was administrators of the department or division of EM who were members of the Society for Academic Emergency Medicine's (SAEM) Academy of Administrators in Academic Emergency Medicine (AAAEM). ¹³ We evaluated surveys with two primary objectives: demographics of enrolled fellows and details of fellowship programs. In the analysis of enrolled fellows, we included filled fellow positions who received the majority of their compensation through the division or department of EM. In the analysis of fellowship programs, we included all programs offering a fellowship opportunity of a full‐time equivalent (FTE) greater than 0.5.

Survey design, sampling, and implementation

The survey was designed by members of AAAEM and the Association of Academic Chairs of Emergency Medicine (AACEM) benchmarking survey committee based on previously validated faculty salary survey questions. ¹⁴ The fellowship survey instrument is included in Appendix A.

We electronically distributed the survey to the EM administrator at each of 73 member sites using a single‐stage sampling technique between October 2022 and January 2023. If we did not receive a response, we sent up to five repeated reminders at standard intervals before considering the site a nonresponder. The survey was hosted on Roundtable Analytics and completed by each member site's EM administrator between October 2022 and January 2023. We prevented multiple participation using a unique code for each site and attempted to minimize human error during data entry by minimizing free‐text fields wherever possible. To further minimize errors in free‐text fields, the survey host contacted the survey respondent about any outliers to confirm a specific data point's accuracy.

Survey respondents were aware of the intended use of survey results. Following receipt of the data by the AAAEM/AACEM benchmarking committee, all data were deidentified so that responses could not be traced back to a single institution or fellow. The analysis of the deidentified survey data set was determined not to be human subjects research by The Ohio State University Institutional Review Board (#2023E0544).

Variables, data sources, and measurement

The survey included five sections: demographics, training, FTE productivity, hours worked, and salary/other compensation. Demographics included gender, race (U.S. Census Bureau definitions), ethnicity (Hispanic, Latino, Spanish, or none of the above), and region (American Association of Medical Colleges [AAMC] standard definitions for Northeast, Central, South, or West). ¹⁵ Training questions included length of prior residency training, length of fellowship, type of fellowship, and ACGME accreditation of fellowship. FTE productivity included FTE and total work relative value units (RVUs). Hours were defined as annual base clinical hours, annual nonclinical hours, and annual extra clinical hours; these were compiled as annual total hours. Compensation was defined as salary based on postgraduate year (PGY) scale or not and then further specified by base compensation, additional compensation, bonus/incentive, total salary, and standard professional enrichment allotment. Additional compensation was defined as the amount paid for extra clinical hours and/or shift differential. Bonus/incentive was defined as a variable amount of compensation, which is unknown at the beginning of the year. Total salary was defined as the sum of base compensation, additional compensation, and bonus/incentive. Professional enrichment allotment was defined as continuing medical education funds and advanced degree tuition assistance (Appendix A).

Statistical methods

We calculated descriptive statistics presented as medians and interquartile ranges (IQRs) for continuous measures and frequency (%) for categorical measures. Fellowship comparisons were stratified by accreditation status (ACGME and non‐ACGME) and gender as part of preplanned analyses. Comparisons between groups were made using chi‐square or Wilcoxon rank‐sum testing. All analyses were completed using STATA IC Version 17.

RESULTS

Descriptive statistics

A total of 38 of 73 institutions (response rate 52%) participated and completed the survey (Appendix B). These institution‐based responses represented 217 individual EM fellows. Demographics of the enrolled fellow population are noted in Table 1. Fellows were mostly White (n = 138, 78%), non‐Hispanic (n = 167, 95%), and graduates of 3‐year EM residency programs (n = 125, 63%). Gender identity was evenly divided between male (n = 113, 52.3%) and female (n = 103, 47.7%). Fellows were most often in the Northeast (n = 66, 30%) or West region (n = 71, 33%) rather than South (n = 44, 20%) or Midwest (n = 36, 17%).

TABLE 1.

EM fellow demographics (n = 217).

Variable	Frequency (%)
Gender
Male	113 (52)
Female	103 (48)
Missing	1
Race
White	138 (78)
Other, including multiracial	32 (18)
Black or African American	7 (4)
Missing	40
Ethnicity
Not Hispanic, Latino, or Spanish	167 (95)
Hispanic, Latino, or Spanish	8 (5)
Missing	42
Region
West	71 (33)
Northeast	66 (30)
South	44 (20)
Midwest	36 (17)
Length of residency
3 years	125 (63)
4 years	75 (37)
Missing	17

Program results

Nearly three‐fourths (n = 158, 72.8%) of fellows were enrolled in non‐ACGME EM fellowships, as defined in Appendix C. Most common fellowship training was in ultrasound (n = 50, 23.0%), EMS (n = 29, 13.4%), education (n = 24, 11.1%), global EM (n = 21, 9.7%), or administration (n = 19, 8.8%). Fellows were most often enrolled in 1‐year programs (n = 120, 55.3%) that did not provide an advanced degree with fellowship training (n = 122, 65.9%; Table 2).

TABLE 2.

Fellowship descriptions by survey respondents (n = 217).

Fellowship accreditation	Frequency (%)
Non‐ACGME	158 (73)
ACGME	59 (27)
Type of fellowship
Ultrasound	50 (23)
EMS	29 (13)
Education	24 (11)
Global EM	21 (10)
Administration	19 (9)
Medical toxicology	10 (5)
Research	10 (5)
Simulation	8 (4)
Sports medicine	8 (4)
Critical care	7 (3)
Wilderness medicine	6 (3)
Pediatrics	3 (1)
Addiction medicine	2 (1)
Clinical informatics	2 (1)
Disaster medicine	2 (1)
Faculty development	2 (1)
Neurovascular/neurologic emergencies	2 (1)
Other	12 (5)
Duration of fellowship
1 year	120 (55)
2 years	94 (43)
>2 years	3 (2)
Advanced degree with fellowship training
No	122 (66)
Yes	63 (34)
Required	16 (25)
Optional	47 (75)
Missing	32
Advanced degree tuition assistance
Dollar amount, median (IQR)	$40,655 ($15,000–$50,000)

Abbreviations: ACGME, Accreditation Council for Graduate Medical Education; IQR, interquartile range.

Clinical hours and compensation based on ACGME or non‐ACGME accreditation

Clinical hours and compensation were evaluated by the program's ACGME accreditation status. Fellows enrolled in non‐ACGME EM fellowships worked nearly 33% more annual base clinical hours than ACGME fellows (571 h vs. 768 h, p < 0.001) and generated 64% more RVUs annually (4446 units vs. 2707 units, p < 0.001). Annual nonclinical hours also differed between non‐ACGME and ACGME fellows (1136 h vs. 1392 h, p = 0.032), though the total annual hours (base clinical hours + nonclinical hours + any extra clinical hours) were not (1689 h for non‐ACGME vs. 1935 h for ACGME, p = 0.6518; Figure 1A).

FIGURE 1.

(A) Box plot of base clinical hours, nonclinical hours, and total work hours between ACGME and non‐ACGME fellows. (B) Box plot of base annual compensation and total annual compensation between ACGME and non‐ACGME fellows. Statistical comparisons used Wilcoxon rank‐sum testing, where *p < 0.05 and **p < 0.001. ACGME, Accreditation Council for Graduate Medical Education.

Fellows enrolled in ACGME‐accredited EM fellowships were often paid based on PGY scale, while this was less common among non‐ACGME fellows (67.9% vs. 24.3%, p < 0.001). The median annual base compensation for non‐ACGME fellows was 20% higher than ACGME fellows ($88,540 vs. $70,777, p < 0.001). After additional compensation (including incentives and bonuses) was accounted for, the median total annual compensation for fellows was 11% higher for non‐ACGME fellows than ACGME fellows ($105,000 vs. $93,853, p = 0.004; Figure 1B). The proportion of fellows who received additional compensation or bonus/incentive was similar between ACGME and non‐ACGME fellows as was the value of the professional enrichment allotment (Table 3).

TABLE 3.

Financial compensation characteristics, stratified by fellowship accreditation.

Variable	ACGME (n = 59)	Non‐ACGME (n = 158)	p‐value
Compensation based on PGY scale
Yes	38 (67.9)	36 (24.3)	<0.001
No	18 (32.1)	112 (75.7)
Did not respond	3	10
Base compensation	$70,777 ($65,440–$87,300)	$88,540 ($75,000–$100,000)	<0.001
Additional compensation	34 (58)	84 (53)	0.557
Bonus/incentive	18 (31)	35 (22)	0.202
Total compensation	$93,853 ($77,255–$118,295)	$105,000 ($91,000–$132,155)	0.004
Professional enrichment allotment	$3238 ($3000–$5000)	$3334 ($3000–$5000)	0.776
Missing	3	12

Note: Data are reported as n (%) or median (IQR).

Abbreviations: ACGME, Accreditation Council for Graduate Medical Education; IQR, interquartile range.

We considered the potential impact of ACGME accreditation on pay disparities between genders and among regions. We observed no significant differences in median total annual salary when stratifying total salary of non‐ACGME fellows by gender ($105,000 male vs. $107,977 female, p = 0.218) and ACGME fellows by gender ($94,658 male vs. $93,730 female, p = 0.370). However, pay disparities existed among regions for both ACGME and non‐ACGME fellows. For non‐ACGME fellows, those in the Midwest earned a median of $63,000 more total annual salary than those in the South (p = 0.004); for ACGME fellows, those in the West earned $50,000 more total annual salary than those in the South (p = 0.002; Tables 4 and 5).

TABLE 4.

Median annual total compensation of non‐ACGME fellows, stratified by gender and region (n = 158).

Variable	n	Annual total compensation, median (IQR)	p‐value
Gender
Male	79	$105,000 ($88,600–$132,155)	0.218
Female	78	$107,977 ($93,730–$133,638)
Region			0.004
West	56	$100,000 ($91,000–$125,270)
Northeast	52	$105,000 ($100,000–$132,155)
South	29	$91,344 ($77,000–$110,648)
Midwest	21	$154,500 ($112,447–$165,643)

Abbreviations: ACGME, Accreditation Council for Graduate Medical Education; IQR, interquartile ranges.

TABLE 5.

Median annual total compensation of ACGME fellows, stratified by gender and region (n = 59).

Variable	n	Annual total compensation, median (IQR)	p‐value
Gender			0.370
Male	34	$94,658 ($77,255–$132,155)
Female	25	$93,730 ($78,130–$112,610)
Region			0.002
West	15	$114,790 ($104,550–$151,933)
Northeast	14	$91,700 ($87,300–$132,155)
South	15	$67,537 ($62,570–$93,852)
Midwest	15	$93,730 ($77,805–$104,305)

Abbreviations: ACGME, Accreditation Council for Graduate Medical Education; IQR, interquartile range.

DISCUSSION

In this cross‐sectional survey of U.S. EM fellows, most enrolled in non‐ACGME fellowships worked 33% more annual base hours and earned $11,000 (11%) more each year than their non‐ACGME counterparts. There were no notable differences in total salary based on gender when stratified by ACGME or non‐ACGME accreditation, though we observed regional variations in pay. These results represent the first description of the work hours and financial compensation of academic EM fellows and should be considered in ongoing benchmarking efforts by AAAEM.

The predominant demographic descriptions in our study of EM fellows (52% male, 78% White) differ slightly from the demographic descriptions of EM residents in 2022 ABEM survey of ACGME fellows (63% male, 67% White). ³ Compared to a 2015 AAAEM evaluation of gender and racial/ethnic diversity among academic EM physicians, we observed more female fellows (48% in our study vs. 33% in the faculty survey) and unchanged proportions of racial and ethnic diversity among our sample. ¹⁴ These point estimates may represent year‐to‐year variation, though they do suggest a trend toward increased gender and ethnic diversity among fellowship‐trained EM physicians. This trend is also supported by literature on specific emergency fellowships, including EMS and ultrasound, which have seen annual increases in gender and racial/ethnic diversity. ⁶ , ⁷ , ⁸ , ⁹

This study also provides some of the first insights into non‐ACGME EM fellows, who were typically excluded from prior national descriptions of EM fellows. ³ , ⁴ Compared to their ACGME counterparts, fellows in non‐ACGME programs tended to have similar demographic descriptions but more often enroll in fellowship programs in ultrasound, education, or administration.

We noted significant increase in both clinical work hours and total compensation among non‐ACGME EM fellows. This was not an unexpected result, given the increased work hours and compensation flexibility in the absence of ACGME regulation. It was, however, interesting to note that some ACGME programs chose not to pay on PGY scale and some non‐ACGME programs chose to pay on PGY scale. In contextualizing these findings, it is important to note that, in the absence of ACGME regulations, some fellowships have established a separate accreditation process (e.g., Emergency Ultrasound Fellowship Accreditation Council) that provide similar work hour, education, and compensation regulations and may explain these findings. ¹⁶ The median salary for ACGME EM fellows ($70,000) in our study was similar to those reported in the AAMC's national salary survey of all residents and fellows ($67,700 for PGY‐4 and $70,769 for PGY‐5), including those in non‐EM programs. ¹⁷ These are expected findings given the standardization of ACGME program salaries based on regional cost of living and PGY pay scale. Regional variation in salary was also noted in the AAMC survey, with lowest salaries noted in the southern region; however, the differences (PGY‐4 median $65,827 and PGY‐5 median $69,288) were less pronounced than in our study (South region median $67,537 vs. other regions median $91,700). Although pronounced, this regional variation in academic EM salaries appears to be consistent with the 2023 AAAEM faculty salary survey, which demonstrated that the South region also had the lowest average total salary among academic EM physicians. ¹⁸ Given the heterogeneity noted in our sample, future research should further examine the association between these trends, the specific fellowship subspeciality, and the accreditation body (whether ACGME, non‐ACGME, or none).

A key secondary finding in our analysis was the absence of a gender‐based pay gap and median salaries for female fellows were slightly higher than male fellows in the non‐ACGME fellowships. This finding suggests that the less‐regulated payment structure of non‐ACGME fellowships does not result in a gender‐associated pay disparity among EM fellows. This also suggests that the previously noted gender‐associated pay disparities in academic EM originate after fellowship. ¹⁹

LIMITATIONS

The internal validity of this study is limited by the survey respondent being the department administrator, which may have resulted in misclassification bias (e.g., demographic data, external moonlighting that department administrator is not aware of). However, given the primary study outcomes of work hours and compensation, the study design required the department administrator to complete this survey. We are limited by selection bias and are unable to determine if there are any structural or organizational aspects of institutions that may have led to survey participation. The external validity of this study is also limited by the survey response rate, potential clustering and overrepresentation among larger institutions with multiple fellowship programs, and the fact that the survey represents a single year. Although we note interesting associations between accreditation body, work hours, and compensation, this study was not designed to infer causality between these variables. Furthermore, this study was not designed to understand if any of the observed associations impact resident choice of fellowship or the quality of fellowship training received.

CONCLUSIONS

In summary, emergency medicine fellows at U.S. academic institutions sampled in this single‐year survey were heterogenous but tended to enroll in non–Accreditation Council for Graduate Medical Education fellowships that often required more clinical hours for significantly higher financial compensation than Accreditation Council for Graduate Medical Education fellowships. Although pay gaps exist between Accreditation Council for Graduate Medical Education and non–Accreditation Council for Graduate Medical Education fellows, gender‐based pay differences were not observed. These findings have important implications for future benchmarking efforts by Academy of Administrators in Academic Emergency Medicine and departmental direction.

AUTHOR CONTRIBUTIONS

Gregory M. Archual and Ashish R. Panchal conceived, designed, and obtained regulatory permission for the study. Alyssa Tyransky and Gregory M. Archual designed the data collection and acquired data. Alyssa Tyransky, Alexander J. Ulintz, and Ashish R. Panchal performed the statistical analysis. Alexander J. Ulintz, Chelsea B. Kadish, and Ashish R. Panchal interpreted and contributed significantly to contextualizing the data. Alexander J. Ulintz drafted the manuscript, and all authors contributed substantially to its revision; all authors provided final approval of the version to be published. Alexander J. Ulintz and Ashish R. Panchal take responsibility for the paper as a whole.

CONFLICT OF INTEREST STATEMENT

AT received funding from the Academy of Academic Administrators of Emergency Medicine to assist with development and implementation of the survey described in this article. The authors declare no other conflicts of interest.

Supporting information

Appendix A.

Appendix B.

Appendix C.

Ulintz AJ, Tyransky A, Archual GM, Kadish CB, Panchal AR. Clinical duties and compensation for U.S. emergency medicine fellows at academic centers: A descriptive, cross‐sectional survey. AEM Educ Train. 2024;8:e10959. doi: 10.1002/aet2.10959