ORIGINAL CONTRIBUTION Current Status of Gender and Racial/Ethnic Disparities Among Academic Emergency Medicine Physicians Tracy E. Madsen, MD, ScM, Judith A. Linden, MD, Kirsten Rounds, RN, MS, Yu-Hsiang Hsieh, PhD, Bernard L. Lopez, MD, MS, Dowin Boatright, MD, MBA, Nidhi Garg, MD, Sheryl L. Heron, MD, MPH, Amy Jameson, MBA, Dara Kass, MD, Michelle D. Lall, MD, MHS, Ashley M. Melendez, MSPH, BSN, James J. Scheulen, PA, MBA, Kinjal N. Sethuraman, MD, MPH, Lauren M. Westafer, DO, MPH, and Basmah Safdar, MD, MSc ABSTRACT Objective: A 2010 survey identified disparities in salaries by gender and underrepresented minorities (URM). With an increase in the emergency medicine (EM) workforce since, we aimed to 1) describe the current status of academic EM workforce by gender, race, and rank and 2) evaluate if disparities still exist in salary or rank by gender. Methods: Information on demographics, rank, clinical commitment, and base and total annual salary for full-time faculty members in U.S. academic emergency departments were collected in 2015 via the Academy of Administrators in Academic Emergency Medicine (AAAEM) Salary Survey. Multiple linear regression was used to compare salary by gender while controlling for confounders. Results: Response rate was 47% (47/101), yielding data on 1,371 full-time faculty: 33% women, 78% white, 4% black, 5% Asian, 3% Asian Indian, 4% other, and 7% unknown race. Comparing white race to nonwhite, 62% versus 69% were instructor/assistant, 23% versus 20% were associate, and 15% versus 10% were full professors. Comparing women to men, 74% versus 59% were instructor/assistant, 19% versus 24% were associate, and 7% versus 17% were full professors. Of 113 chair/vice-chair positions, only 15% were women, and 18% were nonwhite. Women were more often fellowship trained (37% vs. 31%), less often core faculty (59% vs. 64%), with fewer administrative roles (47% vs. 57%; all p < 0.05) but worked similar clinical hours (mean  SD = 1,069  371 hours vs. 1,051  393 hours). Mean overall salary was $278,631 (SD  $68,003). The mean (SD) salary of women was $19,418 ($3,736) less than men (p < 0.001), even after adjusting for race, region, rank, years of experience, clinical hours, core faculty status, administrative roles, board certification, and fellowship training. From the Department of Emergency Medicine, Alpert Medical School of Brown University (TEM, KR), Providence RI; the Department of Emergency Medicine, Boston Medical Center/Boston University School of Medicine (JAL), Boston, MA; the Department of Emergency Medicine, Johns Hopkins University School of Medicine (YHH, JJS), Baltimore, MD; the Department of Emergency Medicine, Sidney Kimmel Medical College of Thomas Jefferson University (BLL), Philadelphia, PA; the Department of Emergency Medicine, Yale University (DB, BS), New Haven, CT; the Department of Emergency Medicine, Hofstra University School of Medicine (NG), Hempstead, NY; the Department of Emergency Medicine, Emory University School of Medicine (SLH, MDL), Atlanta, GA; the Department of Emergency Medicine, University of New Mexico (AJ), Albuquerque, NM; the Department of Emergency Medicine, New York University School of Medicine (DK), New York, NY; the Department of Emergency Medicine, University of Louisville (AMM), Louisville, KY; the Department of Emergency Medicine, University of Maryland School of Medicine (KNS), Baltimore, MD; and the Department of Emergency Medicine, Baystate Medical Center (LMW), Springfield, MA. Received March 18, 2017; revision received July 9, 2017; accepted July 16, 2017. Presented at the Society for Academic Emergency Medicine (SAEM) Annual Meeting, Orlando, FL, May 2017, and at the SAEM New England Regional Meeting, Worcester, MA, March 2017. The authors have no relevant financial information to disclose. DK is the founder and Editor-in-Chief of FemInEM, a national organization dedicated to supporting the development and advancement of all women in medicine. SH receives royalties from Springer as an editor of the textbook Diversity and Inclusion in Quality Patient Care. The other authors have no potential conflicts to disclose. Supervising Editor: Esther K. Choo, MD, MPH. Address for correspondence and reprints: Tracy E. Madsen, MD, ScM; e-mail: Tracy_Madsen@brown.edu. ACADEMIC EMERGENCY MEDICINE 2017;24:1182–1192. 1182 ISSN 1069-6563 PII ISSN 1069-6563583 © 2017 by the Society for Academic Emergency Medicine doi: 10.1111/acem.13269 ACADEMIC EMERGENCY MEDICINE • October 2017, Vol. 24, No. 10 • www.aemj.org 1183 Conclusions: In 2015, disparities in salary and rank persist among full-time U.S. academic EM faculty. There were gender and URM disparities in rank and leadership positions. Women earned less than men regardless of rank, clinical hours, or training. Future efforts should focus on evaluating salary data by race and developing systemwide practices to eliminate disparities. P romoting equity, inclusion, and diversity within a physician workforce is critical to ensure professional excellence and culturally competent patient care. It is also an important measure toward successful recruitment, engagement, and retention of high-quality medical students, residents, and faculty.1 Even though diversity in the physician workforce has been shown to positively influence patient outcomes, disparities continue to exist in salary and academic promotion for women and underrepresented minorities (URM) across the house of medicine.2–6 In emergency medicine (EM) specifically, survey data collected in 2010 from academic emergency departments (EDs) demonstrated that the salary for female faculty was, on average, 12% less than their male counterparts and that this difference widened as physicians were promoted.7 The salary disparity persisted across all academic ranks and most leadership positions.3,7 Similar disparities exist in the physician workforce by race and ethnicity. While blacks and Hispanics represent approximately 30% of the United States population, they represent less than 6% of practicing physicians and 9% of EM physicians in the United States.8 Whether these salary and rank disparities continue to exist among EM physicians is unknown.9,10 Data on URM in academic EM is also sparse. A contemporary data set that provides information on full-time academic EM faculty is needed to answer these questions. Prior reports have used the Association of American Medical Colleges (AAMC) annual salary survey as a conventional benchmark. The AAMC survey, however, includes only research and education faculty and not faculty on clinical tracks. In addition, many EM faculty are not represented in the AAMC data, as evidenced by significant differences in reported salaries for EM physicians between the AAMC and other sources.7 The Academy of Administrators in Academic Emergency Medicine (AAAEM), an academy within the Society for Academic Emergency Medicine (SAEM), annually conducts a salary survey with characteristics and data points for clinical faculty in addition to research and education faculty, providing a more comprehensive overview of the academic EM workforce. The AAAEM survey therefore allows a unique opportunity to track the current status of the EM academic workforce in relation to gender and race/ethnicity. Our objectives were to use data from the 2015 AAAEM salary survey to describe the current academic EM workforce by gender, race, and rank and to determine whether gender disparities in salary or rank currently exist. METHODS This was a cross-sectional observational study of academic EDs in the United States. This study was submitted to the institutional review board (IRB) of the institution that maintains the data set and was determined to be exempt from IRB approval. Participants Survey participants included academic EDs or divisions (adult and pediatric) throughout the United States. Representatives from each department or division were asked to provide deidentified information regarding their individual faculty members. Potential participants were identified by their membership in the AAAEM or the Association of Academic Chairs in Emergency Medicine (AACEM) and were invited to participate in completion of the AAAEM 2015 salary survey via the AAAEM online community e-mail system and the AACEM online community e-mail system. For participating centers, the department administrators were asked to log onto the Novi Survey portal to provide information regarding their faculty members for the previous academic year (July 1, 2014–June 30, 2015). A group of designated “super users” consisting of selected administrators and/or chairs followed up with all potential respondents during the collection period to encourage responses, answer questions, and enhance the response rate. Survey Description The AAAEM salary survey is an iteration of a salary survey that has been periodically conducted by SAEM since 1991.7,11 In 2009, AACEM and AAAEM became responsible for periodically conducting the survey. The survey instrument has been revised multiple times since its development. Each iteration has been 1184 approved by the SAEM Board of Directors; revisions since 2009 have been approved by AACEM and AAAEM leadership. Gender items were added to the administrator’s survey in 2013, and race and ethnicity were added to the survey in 2015. The 2015 salary survey included items regarding demographics, academic rank, years of experience, clinical hours worked, training and board certification, core faculty status, administrative roles, position, and salary for faculty in each department. Survey data were collected and maintained by a central site. Each data point was reviewed if it fell on the extreme end of the range submitted (either high or low); the accuracy of potential outliers was confirmed with the representative from the participant’s department. Variables/Definitions Salaries used in the present analysis were base and/or total salaries (base salary plus incentive compensation, additional compensation for overtime or other clinical premiums, and stipends for other work). Demographic data for each faculty member included gender, race, ethnicity, and years as EM faculty. Race was operationalized into six categories: white/non-Hispanic, black/ African American, Asian, Asian Indian, other (whiteHispanic, Pacific Islander, multiracial, some other race, and American Indian, each of which represented < 2% of the overall sample), and unknown. Ethnicity was categorized as Hispanic or non-Hispanic. The geographic region of the United States (northeast, midwest, south, west) was also reported for each department. Academic rank was categorized as instructor, assistant professor, associate professor, or professor. The total years of experience variable was defined as total years as EM faculty since completion of training. Yearly hours worked were defined as clinical hours and total hours. Fellowship trained was defined as having completed any fellowship (ACGME or nonACGME approved) following residency training. Board certified was defined as either certified or board eligible by the American Board of Emergency Medicine (ABEM). Core faculty, as defined by the ACGME, included faculty who provide clinical service and teaching, who devote the majority of their professional efforts to the program and who have enough protected time to meet the educational requirements of the program. Administrative roles were defined as having a role in the department with substantial time spent organizing, managing, or otherwise being responsible for the running of a program, division, Madsen et al. • DISPARITIES IN ACADEMIC EM WORKFORCE site, or educational activity. Job duties/positions reported in the current analysis include chair/chief, vice-chair, medical director, and assistant or associate medical director. Data Analysis Baseline data by gender and race/ethnicity were reported using descriptive statistics (frequencies/proportions, means/SD, and median/interquartile range [IQR], as appropriate). Analysis was restricted to fulltime faculty only. Mean and median base and total salaries were reported by gender, stratified by rank and geographic region. To protect responder anonymity, mean salaries were not reported for categories in which there were less than 10 faculty members. Salaries for those participants in gender/rank/region categories in which there were fewer than 10 participants, however, were reported in aggregate (reported by rank and gender but not stratified by region) and were included in regression models. Finally, average salaries were reported by gender, rank, and years as EM faculty and then by gender, rank, and clinical hours, which were divided into quartiles. Multivariate linear regression was conducted with total salary as the outcome and gender as the primary independent variable. Two models were created a priori in consideration of potential confounders and covariates on salary and number of variables in the model. We followed a rule of thumb, ensuring that there were at least 10 outcomes for every covariate in order to avoid overfitting. The full model included gender along with 11 other covariates, chosen a priori by authors based on previous literature7 and scientific plausibility: race/ethnicity, geographic region, academic rank, total years as EM faculty, type of primary site, core faculty status, administrative roles, total clinical hours, EM residency training status, fellowship training status, and ABEM board eligible/certified. The final model included gender and only other covariates if they were statistically significant (p < 0.05) using a stepwise selection approach and if they did not cause lack of fit for the model. A sensitivity analysis was conducted using base salary instead of total salary, again using multivariate linear regression. As in the analyses with total salary, potential confounders were chosen a priori and included in the final model using a stepwise selection approach. A second sensitivity analysis was conducted using total salary as the outcome but excluding chairs from the sample using multivariate linear regression. ACADEMIC EMERGENCY MEDICINE • October 2017, Vol. 24, No. 10 • www.aemj.org Multivariate logistic regression was also used to examine gender along with other potential factors associated with associate or full professor rank. The model included race/ethnicity, total years as EM faculty, primary site, status as core faculty, administrative roles, EM residency training, and fellowship training; the final model included gender and only statistically significant factors. For the purposes of the regression models, race was treated as a binary variable (white/ non-Hispanic vs. other) because of the small numbers of faculty in categories other than white/non-Hispanic. Further analysis of salary by race or ethnicity was not attempted because of the small number of faculty reported in each of the race categories other than white/non-Hispanic. A type I error rate of 0.05 was used in all analyses. All statistical analyses were performed using SAS version 9.4 (SAS Institute). Model diagnostics for the final regression models were performed using a lackof-fit test to determine goodness of fit of the models, variance inflation factor to determine multicollinearity, and an SAS diagnostic panel to identify presence of outliers and influential data points as well as to diagnose violations of the normality and homoscedasticity assumptions. Outliers, however, were not excluded from the regression models since these outliners may be more likely to be female and/or minority race. RESULTS Of the 101 departments invited to participate, 47 departments responded for a response rate of 47%. From a total of 1,599 faculty, 86% (n = 1,371) were full time and were included in the analyses. One male faculty member’s salary was not reported and was thus excluded from the study, and one faculty member was excluded because of gender marked as “unknown.” Of 1,371 faculty in the final analytical sample, 33% (n = 447) were women, and 22% (n = 305) were in categories other than white/non-Hispanic. Gender Comparison of EM Academic Workforce Table 1 shows descriptive characteristics of representative EM faculty by gender and race/ethnicity. Women and men differed significantly by academic rank, fellowship training, administration roles, position, and total years as EM faculty (all p < 0.05). Fewer women than men were professors (7% vs. 17%) and associate professors (19% vs. 24%), while more women were 1185 instructors or assistant professors (74% vs. 59%). More women than men were fellowship trained (36.9% vs. 31.1%), but fewer women had administrative roles (47% vs. 57%). Women less often held a chair/chief or vice-chair position (4% vs. 10%). Women worked similar clinical hours as men (1,068.6  371 vs. 1,051.3  393), as well as total hours (clinical + teaching + administrative; 1,928.5  523.7 vs. 1,899.3  493.9; all p > 0.05). Comparable numbers of men and women were also EM residency trained and board certified. Table 1A in Data Supplement S1 (available as supporting information in the online version of this paper, which is available at https://doi.org/ onlinelibrary.wiley.com/doi/10.1111/acem.13269/full) shows descriptive characteristics with faculty broken down into race/gender subgroups. Race and Ethnicity Comparison of EM Workforce Significant differences by faculty race included academic rank, core faculty status, administration roles, total hours worked, and clinical hours worked (all p < 0.05). Table 1 shows these characteristics divided into six categories by race. When treated as a binary variable, comparison of white/non-Hispanic versus all other (black/African American, Asian, Asian Indian, other) revealed that significantly more faculty in the “all other” category were classified in the clinical instructor and assistant professor category: 69% (n = 210), compared to 62% (n = 663) of white/nonHispanic faculty. In contrast, 15% (n = 155) of white/ non-Hispanic faculty were full professors compared to 10% (n = 30) of all other faculty. Of white/non-Hispanic faculty, 55% (n = 590) had administration roles. Among nonwhite faculty members, proportions of faculty with administration roles ranged from 44% of black faculty to 61% of Asian Indian faculty. Of the 113 chair/vice-chair positions, 18% were filled by nonwhite faculty. Both clinical and total hours also differed across racial groups with those in the “other race” category working the most number of hours and those with unknown race working the least mean number of hours (Table 1). With regard to ethnicity, 3% (n = 40) of participants were Hispanic. Of all Hispanic participants, three were instructor, 25 were assistant, seven were associate, and five were full professors. Of the 40 Hispanic participants, 15 were fellowship trained, 37 were EM residency trained, 26 were core faculty, and 40 were EM board certified. One of the Hispanic 1186 Table 1 Characteristics of Academic EM Physicians Women Overall breakdown 447 (32.6) Men 924 (67.4) White/ non-Hispanic 1066 (77.8) Black/African American 54 (3.9) Asian 62 (4.5) Asian Indian 36 (2.6) Other* 51 (3.7) Unknown 102 (7.4) Rank Instructor 39 (8.7) 65 (7.0) 65 (6.1) 5 (9.3) 12 (19.4) 5 (13.9) 3 (5.9) 14 (13.7) Assistant professor 291 (65.1) 478 (51.7) 598 (56.1) 31 (57.4) 34 (54.8) 19 (52.8) 31 (60.8) 56 (54.9) Associate professor 86 (19.2) 221 (23.9) 245 (23.0) 16 (29.6) Professor 29 (6.5) 156 (16.9) 155 (14.5) 2 (0.5) 4 (0.4) 3 (0.3) Missing Total years of experience 8.4  6.8 12.0  9.2 11.2  8.9 11 (17.7) 9 (25.0) 11 (21.6) 15 (14.7) 0 (0) 5 (8.1) 3 (8.3) 5 (9.8) 17 (16.7) 2 (3.7) 0 (0) 0 (0) 1 (2.0) 0 (0.0) 8.4  5 .6 8.9  6.8 9.7  5.8 9.6  7.6 11.3  8.6 Yearly hours Clinical 1068.6  371.0 1051.3  393.2 1064.8  375.8 1073.7  361.1 1079.1  334.6 1036.0  347.2 1133.4  354.4 924.0  517.0 Total 1899.3  493.9 1928.5  523.7 1947.0  494.4 1888.4  618.9 1849.1  431.6 1919.5  508.3 2076.4  446.3 1613.6  622.2 FTE (mean) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Fellowship trained 165 (36.9) 287 (31.1) 334 (31.3) 18 (33.3) 21 (33.9) 15 (41.7) 21 (41.2) 1.0 EM residency 405 (90.6) 838 (90.7) 961 (90.2) 50 (92.6) 57 (91.9) 33 (91.7) 47 (92.2) 95 (93.1) Board certified 432 (96.6) 894 (96.8) 1035 (97.1) 51 (94.4) 61 (98.4) 34 (94.4) 50 (98.0) 95 (93.1) Core faculty 264 (59.1) 593 (64.2) 682 (64.0) 35 (64.8) 37 (59.7) 24 (66.7) 36 (70.6) 43 (42.2) Admin duties (yes/no) 211 (47.2) 522 (56.5) 590 (55.4) 24 (44.4) 31 (50.0) 22 (61.1) 25 (49.0) 41 (40.2) 43 (42.2) Chair/chief 11 (2.5) 54 (5.8) 53 (5.0) 1 (1.9) 2 (3.2) 2 (5.6) 1 (2.0) 6 (5.9) Vice-chair 6 (1.3) 42 (4.6) 40 (3.8) 1 (1.9) 1 (1.6) 0 (0.0) 3 (5.9) 3 (2.9) 137 (30.7) 328 (35.5) 379 (35.6) 16 (29.6) 15 (24.2) 16 (44.4) 13 (25.5) 26 (25.5) 57 (12.8) 98 (10.6) 118 (11.1) 6 (11.1) 13 (21.0) 4 (11.1) 8 (15.7) 6 (5.9) Medial director Medical director– assistant/associate Data are reported as n (%) or mean  SD. *Other includes white/Hispanic, Pacific Islander, multiracial, and American Indian; all of these categories represent < 2%. Statistically significant (p < 0.05) differences between female and male EM faculty included academic rank, fellowship trained, administration roles, position, total years as EM faculty; statistically significant (p < 0.05) by race/ethnicity of EM faculty included academic rank, core faculty, administration roles, annual total hours worked (adjusted), and annual clinical hours worked (adjusted). Madsen et al. • DISPARITIES IN ACADEMIC EM WORKFORCE Position (duty as per survey) ACADEMIC EMERGENCY MEDICINE • October 2017, Vol. 24, No. 10 • www.aemj.org participants was a chair, three were vice-chairs, and 10 were medical directors. Salary Comparison of EM Workforce The mean (SD) and median (IQR) of base and total salary for an academic EM physician in our cohort were $233,419 ($62,321)/$233,810 ($204,803– $258,600) and $278,631 ($68,003)/$268,713 ($241,750–$303,596), respectively. In 2015, the mean (SD) and median (IQR) base and total salary for assistant level position were $219,410 ($46,966)/ $222,133 ($199,000–$249,373) and $260,223 ($48,089)/$254,448 ($233,360–$281,934), for associate professor were $250,849 ($56,151)/$249,373 ($228,000–$274,390) and $298,089 ($58,847)/ $281,250 ($263,878–$320,340), and for full professor were $274,656 ($95,408)/$266,000 ($226,744– $308,724) and $339,961 ($100,433)/$312,110 ($272,920–$372,008). Gender Comparison for Annual Salary Table 2 shows mean and median total salary stratified by gender, geographic region, and rank. The mean and median total salary across all geographic regions for instructors were $224,238/$225,912 for women compared to $266,333/$264,655 for men. For women at the assistant professor level, and mean and median salaries were $249,105/$246,530 compared to $266,991/$260,230 for men at the same level. For women at the associate professor level, mean and median salary was $273,441/$272,288 compared to $307,681/ $288,169 for men at the same level. For women at the full professor level, mean and median salary was $324,713/$294,050 for women compared to $342,795/$315,603 for men. With the exception of mean and median salary at the full professor level, all of these differences were statistically significant. Similar disparities by gender were found in mean and median base salaries (Data Supplement S1, Table 2A). Table 3 shows the multivariate linear regression of factors associated with total salary. After race, geographic region, academic rank, total years as EM faculty, primary site, core faculty, administrative roles, total clinical hours, residency training, fellowship training, and EM board certified/eligible were adjusted for, women made $19,418  3,736 less than men (p < 0.001). Other factors that were significant included geographic region, academic rank, total years as EM faculty, and administrative roles. In a similar 1187 regression model with base salary as the outcome, women made $13,109  3,298 less than men (p < 0.001; Data Supplement S1, Table 3A). Finally, when chairs were excluded from the model, results showed that women still made significantly less ($18,658  $3,075; Data Supplement S1, Table 3B, supplement). Mean salaries for chairs and vice-chairs by gender can be found in Table 2 in the supplement. Comparing starting salaries by gender, we found that women received a lower starting package at each academic rank compared to men (p < 0.001 for instructors, p = 0.015 for assistant professors). To account for the influence of factors such as fellowship training or the number of clinical commitments, we performed a stratified analysis for these confounders. We found that the gender disparity in starting salaries persisted when comparing fellowship-trained physicians (Data Supplement S1, Tables 2B and 2C; p < 0.05). Finally, salaries for women were less than men across almost all quartiles of clinical hours (Figure 1 and Data Supplement S1, Table 2D). Gender and Race Comparison of Rank Table 4 shows the multivariate regression of factors associated with a rank of associate or full professor. After geographic region was adjusting for, total years as EM faculty, primary site, core faculty, administrative roles, residency training, fellowship training, neither gender nor race influenced rank of associate or full professor (p = 0.375, p = 0.202). DISCUSSION Using a contemporary survey, we found that the proportion of women physicians in the current EM workforce (33%) is slightly higher than previous reports (31%).7 Despite this encouraging trend, we confirmed that significant disparities continue to exist in representation, rank, and salaries by gender and also by race among academic EM physicians. The majority of academic EM physicians were male (67%) and white/ non-Hispanic (78%). Outside of the white/non-Hispanic category, no other single racial or ethnic category comprised more than 5% of faculty. We also noted disparities in academic rank and position; fewer women and URM were professors, associate professors, chairs, and vice-chairs. Women on average earned $19,000 less than their male colleagues, even after controlling for academic rank, geographic region, 1188 Madsen et al. • DISPARITIES IN ACADEMIC EM WORKFORCE Table 2 Mean and Median Total Salary ($) by Gender/Region/Rank Women (n = 445) Men (n = 920) n Mean  SD/Median (IQR) n Mean  SD/Median (IQR) 13 257,565  46,763 15 255,688  51,251 139 282,063  47,277 56 304,957  55,269 33 384,937  104,758 Midwest Instructor 244,800 (222,107–280,063) Assistant professor 76 260,167  45,283 Associate professor 18 272,270  24,016 231,506 (219,425–279,200) 255,341 (231,843–282,264) 274,625 (251,220–300,686) 272,924 (255,602–279,504) Full professor 2 – 281,375 (267,366–336,791) 351,959 (299,789–449,519) Northeast Instructor 14 233,486  54,631 30 298,099  57,791 157 267,105  47,704 88 315,313  71,780 41 361,399  104,274 236,220 (215,645–266,583) Assistant professor 114 250,429  40,364 Associate professor 27 286,271  48,693 290,828 (264,655–311,246) 246,709 (224,430–268,163) 258,858 (240,091–296,461) 280,282 (252,392–320,340) Full professor 7 – 300,352 (278,043–342,305) 327,420 (293,289–401,039) South Instructor 2 – 10 268,532  61,391 101 266,693  49,090 258,100 (242,000–308,472) Assistant professor 60 246,978  33,445 243,599 (225,753–267,598) Associate professor 21 270,856  21,362 262,832 (233,915–293,330) 44 314,580  62,238 26 344,386  74,206 273,113 (263,600–282,253) Full professor 8 - 288,400 (275,678–331,370) 324,864 (276,600–331,370) West Instructor 10 164,423  55,032 10 184,804  48,100 81 241,279  48,773 195,000 (103,000–206,604) Assistant professor 41 228,033  51,577 203,325 (125,900–225,912) 239,786 (221,287–253,000) Associate professor 20 259,885  40,722 245,232 (224,426–258,779) 33 282,749  41,742 56 303,603  93,222 259,247 (235,354–275,134) Full professor 12 289,217  56,695 287,860 (253,336–311,884) 272,707 (257,880–285,739) 292,900 (266,177–315,603) Numbers represent raw, unadjusted salaries in dollars. Across all geographic regions: mean  SD/median (IQR) salary for instructors was $224,238  $61,742/$225,912 ($195,000–$266,583) for women compared to $266,333  $66,949/$264,655 ($227,595–$299,993) for men. For assistant professors, mean  SD/median (IQR) salary was $249,105  $43,083/$246,530 ($225,141–$271,042) for women compared to $266,991  $49,735/$260,230 ($239,330–$291,236) for men. For associate professors, mean  SD/median (IQR) salary was $273,441  $37,631/$272,288 ($251,979–$285,700) for women compared to $307,681  $62,767/$288,169 ($270,547–$334,743) for men. For full professors, mean  SD/median (IQR) salary was $324,713  $101,307/$294,050 ($267,220–$361,990) for women versus $342,795  100,342/$315,603 ($279,958–$377,804) for men. *Salaries not reported in cells with n < 10 to maintain anonymity. training, clinical hours worked, years of experience, and administrative roles. These factors help mitigate the assumption that disparity in salary is due to less experience, qualification, or hours worked. Our study findings affirm a gender gap in salary within EM that has been demonstrated across specialties, demonstrate that gender gaps in salary in academic EM have persisted over time and illustrate that women are compensated less even after controlling for key factors that have been speculated to contribute to salary differences.10,12 Prior studies have suggested that gender disparities in salary among academic 1189 ACADEMIC EMERGENCY MEDICINE • October 2017, Vol. 24, No. 10 • www.aemj.org Table 3 Multivariate Analysis of Factors Associated with Total Salary of EM Faculty First Model Variables Categories Intercept Gender Race Academic rank 19,418  3,736 Other Reference West 21,423  3,366 <0.001 0.197 Reference Reference <0.001 36,126  4,489 <0.001 Midwest 35,066  5,121 <0.001 44,099  4,742 <0.001 South 26,237  6,455 <0.001 32,064  5,005 <0.001 Instructor/assistant professor Reference Reference Associate professor 23,208  4,794 <0.001 31,141  3,913 <0.001 Professor 60,244  7,062 <0.001 76,266  4,896 <0.001 1,048  293 <0.001 Community Academic affiliated Academic Reference 1,786  10,300 –10,111  8,969 No Reference Yes 2,325  3,999 – – 0.260 – – 0.561 No Reference 23,144  3,862 <0.001 –1.2  5.6 0.832 Increase every hour EM residency trained No Reference Yes 9,033  7,000 No Reference Yes –1,646  3,681 No Reference Yes 1,664  9,981 – 0.862 Yes Total clinical hours (adjusted) EM board eligible/certified Reference <0.001 22,919  4,823 Primary site Fellowship trained p-value Northeast Increase every year Administrative roles 5,720  4,427 Salary  SD ($) 203,147  4,503 Reference Male Total years as EM faculty Core faculty Final Model* p-value 197,748  16,081 Female White-non-Hispanic Region Salary  SD ($) – Reference 25,119  3,259 <0.001 – – 0.197 – 0.655 – – – 0.868 First model has all variables included; final model has only significant (p < 0.05) variables. *p value of lack-of-fit test was 0.121 and R2 was 0.293 for final model. Figure 1. Average salary by gender and number of clinical hours. Number of clinical hours divided into quartiles (first, second, third, and fourth). Error bars represent SDs. First indicates the lowest quartile (lowest number of clinical hours); fourth indicates highest quartile physicians could be attenuated by adjusting for part time status, clinical hours worked, or leave time (i.e., maternity leave) during the course of physicians’ careers.12 In our study, we eliminated these issues by comparing salaries of only full-time faculty. We found that gaps in salaries persisted despite the fact that more women EM physicians were fellowship trained, worked similar number of clinical hours, and were EM residency trained. Men starting at instructor position earned 35% more than women and 10% higher starting in an assistant professor position. These trends persisted across increasing academic rank, each quartile of clinical hours worked, or fellowship training. Such gaps have consistently been reported before,7 but the primary focus on gender disparities in our study allowed us to conduct a methodologically rigorous study that controlled for potential confounding factors. Watts et al.7 reported that women EM faculty made 10% to 13% less than men in 2010. Concerningly, the gap in starting salary has not narrowed, as in 1995, earned 12% higher, on average, than women 1190 Madsen et al. • DISPARITIES IN ACADEMIC EM WORKFORCE Table 4 Multivariate Analysis of Factors Associated With Associate or Full Professor Rank of EM Faculty Full Model Variables Gender Race Region Categories OR (95% CI) Female 1.00 Male 1.18 (0.82–1.70) Other 1.00 White-non-Hispanic 1.32 (0.86–2.02) Final Model p-value OR (95% CI) p-value – 0.375 – – 0.202 Northeast 1.00 Midwest 1.37 (0.89–2.12) 0.157 1.26 (0.83–1.90) South 1.87 (1.06–3.32) 0.032 2.09 (1.24–3.52) 0.006 West 3.08 (1.94–4.88) <0.001 2.90 (1.90–4.44) <0.001 Total years as EM faculty Increase every year 1.25 (1.21–1.28) <0.001 1.23 (1.20–1.27) <0.001 Primary site Community 1.00 Academic affiliated 1.76 (0.45–6.92) 0.418 Academic 3.10 (0.87–11.01) 0.104 No 1.00 Core faculty 1.00 – Yes 1.73 (1.17–2.56) No 1.00 Yes 1.90 (1.33–2.71) EM residency trained No 1.00 Yes 1.86 (0.88–3.93) No 1.00 Yes 1.01 (0.71–1.45) No 1.00 Yes 1.16 (0.42–3.19) EM board eligible/certified – – 1.00 Administrative roles Fellowship trained 0.277 0.006 2.08 (1.45–2.97) <0.001 1.00 <0.001 1.92 (1.36–2.71) <0.001 – 0.102 – – 0.946 – – 0.771 – First model has all variables included; final model has only significant (p < 0.05) variables. EM physicians. Consistent with previous reports, we found fewer proportions of women in associate and full professor academic rank.13 Reasons for salary disparities by gender are unclear but may include gender differences in negotiation as well as the presence of conscious and unconscious biases. For example, previous data have shown that women negotiate less often, are less likely to initiate negotiations, and negotiate lower salaries compared to men.14–16 Complicating gender differences in negotiation are data showing that compared to men, women may be penalized for initiating negotiations.17 Women who initiate negotiations for higher compensation may be perceived as too “demanding” or less friendly, going against typical feminine stereotypes. Such perceptions may then lead to a reluctance on the part of women to negotiate higher salaries.17 Gender differences in negotiation could affect starting salaries as well as salary increases over time. Implicit, or unconscious, biases about the value of women’s contributions to the workforce could also contribute to disparities in salary.18 The lack of women and URM at higher ranks is consistent with recent cross-sectional data from the AAMC showing that although women constitute close to 50% of all graduating medical students, only 21% of full professors are female.3 Women have cited numerous reasons for leaving careers in academic medicine, including lack of adequate mentorship, high work–family conflict, gender discrimination, and interest in teaching over research;13,19 however, failure to advance or to receive promotion to leadership roles may also be a factor, as our study suggests. Future research is needed to delineate the issues of retention and advancement. Our study also adds to prior EM literature by including information on race and ethnicity, presenting a descriptive analysis of academic rank, salary, and leadership positions of URM faculty.7 Our findings are consistent with data showing that URM comprise a small proportion of the academic medicine workforce and are less likely to hold senior positions.2,8 A study of 128 academic medical centers found that black or Hispanic faculty constituted only 5% of new academic ACADEMIC EMERGENCY MEDICINE • October 2017, Vol. 24, No. 10 • www.aemj.org hires and they had significantly longer promotion timelines when compared to their white counterparts.2 Similar data were found when examining promotion rates for Asian Pacific Islander, URM (black, Mexican American, Puerto Rican, Native American, and Native Alaskan), and other Hispanic physicians. After factors such as gender, tenure status, degree, and NIH award status were adjusted for, URM faculty were still less likely to be promoted at all levels. In addition to the negative implications for academic EM physicians, a lack of diversity and equality among academic EM faculty has consequences for trainees and patients. A diverse workforce in EM is more representative of the larger population, and diverse providers can bring their unique experiences and understanding of patient backgrounds, leading to increased cultural sensitivity and improved care for the diverse ED patient population. Increasing diversity in our workforce has the potential to decrease treatment disparities.20 The importance of a diverse EM workforce has recently been highlighted by both the American College of Emergency Physicians (ACEP) and the SAEM.21–23 Part of the new ACEP strategic plan is “to promote and facilitate diversity, inclusion, and cultural sensitivity.”21 Similarly, the Academy for Diversity and Inclusion in Emergency Medicine (ADIEM), on behalf of SAEM, has established a goal of diversifying the EM physician workforce so that it better reflects the population of patients we serve.22,23 LIMITATIONS There are several limitations to our analysis. First, the response rate of 47% may have led to a bias in the sample of departments represented by this study. It is possible that departments that did participate were those more confident in their hiring, salary conferral, and promotion practices or those that had the data readily available due to prior audits of these processes; if so, it is possible that our study results underestimate existing gender disparities. Some misclassification by gender, race, or ethnicity may have occurred because in most circumstances, department administrators rather than individual faculty reported these characteristics. Future studies of disparities by race and ethnicity should report race and ethnicity as self-identified by faculty members. In addition, there may be factors that affect promotion or compensation that were not accounted for in our 1191 study, including family leave time prior to the survey which was not collected. CONCLUSIONS Despite the prioritization of increasing diversity by leading organizations of emergency medicine, gender disparities in academic rank and compensation and racial and ethnic disparities in academic rank (and composition) among academic emergency medicine physicians continue to exist. Further research is needed to identify other strategies to eliminate disparities and match the mission of ensuring inclusion, diversity, and equality in academic emergency medicine. The authors support the AWAEM, ADIEM, and AAAEM Membership for supporting this work. The authors acknowledge Susan Watts, PhD, and the AWAEM Board of Directors for their collaboration with AAAEM in the addition of gender items to the AAAEM administrator’s survey. References 1. Martin ML. The value of diversity in academic emergency medicine. Acad Emerg Med 2000;7:1027–31. 2. Nunez-Smith M, Ciarleglio MM, Sandoval-Schaefer T, et al. Institutional variation in the promotion of racial/ethnic minority faculty at US medical schools. Am J Public Health 2012;102:852–8. 3. The State of Women in Academic Medicine. The Pipeline and Pathway to Leadershi. Washington, DC: Association of American Medical Colleges, 2013:2013–2014. 4. Wehner MR, Nead KT, Linos K, Linos E. Plenty of moustaches but not enough women: cross sectional study of medical leaders. BMJ 2015;351:h6311. 5. Carr PL, Gunn CM, Kaplan SA, Anita R, Freund KM. Inadequate progress for women in academic medicine: findings from the national faculty study. J Women’s Health 2015;24:190–9. 6. Fang D, Moy E, Colburn L, Hurley J. Racial and ethnic disparities in faculty promotion in academic medicine. J Am Med Assoc 2000;284:1085–92. 7. Watts SH, Promes SB, Hockberger R. The Society for Academic Emergency Medicine and Association of Academic Chairs in Emergency Medicine 2009-2010 emergency medicine faculty salary and benefits survey. Acad Emerg Med 2012;19:852–60. 8. Castillo-Page L. Diversity in the Physician Workforce: Facts & Figures 2010. Washington, DC: Association of American Medical Colleges, 2010. 9. Kuhn GJ, Abbuhl SB, Clem KJ. Recommendations from the Society for Academic Emergency Medicine (SAEM) 1192 10. 11. 12. 13. 14. 15. 16. 17. 18. taskforce on women in academic emergency medicine. Acad Emerg Med 2008;15:762–7. Jagsi R, Griffith KA, Stewart A, Sambuco D, DeCastro R, Ubel PA. Gender differences in the salaries of physician researchers. JAMA 2012;307:2410–7. Kristal SL, Randall-Kristal KA, Thompson BM. 20012002 SAEM emergency medicine faculty salary and benefits survey. Acad Emerg Med 2002;9:1435–44. Freund KM, Raj A, Kaplan SE, et al. Inequities in academic compensation by gender: a follow-up to the national faculty survey cohort study. Acad Med 2016;91:1068–73. Westring A, McDonald JM, Carr P, Grisso JA. An integrated framework for gender equity in academic medicine. Acad Med 2016;91:1041–4. Small DA, Gelfand M, Babcock L, Gettman H. Who goes to the bargaining table? The influence of gender and framing on the initiation of negotiation. J Pers Soc Psychol 2007;93:600–13. Sarfaty S, Kolb D, Barnett R, et al. Negotiation in academic medicine: a necessary career skill. J Women’s Health 2007;16:235–44. Stevens CK, Bavetta AG, Gist ME. Gender differences in the acquisition of salary negotiation skills: the role of goals, self-efficacy, and perceived control. J Appl Psychol 1993;78:723–35. Bowles HR, Babcock L, Lai L. Social incentives for gender differences in the propensity to initiate negotiations. Organ Behav Hum Decis Process 2007;103:84–103. Carney D, Nosek BA, Greenwald AG, Banaji MR, Baumeister R, Vohs K. Implicit Association Test (IAT). Madsen et al. • DISPARITIES IN ACADEMIC EM WORKFORCE 19. 20. 21. 22. 23. In: Baumeister R, Vohs K, editors. Encyclopedia of Social Psychology. Thousand Oaks, CA: SAGE Publications Inc, 2007:463–6. Edmunds LD, Ovseiko PV, Shepperd S, et al. Why do women choose or reject careers in academic medicine? A narrative review of empirical evidence. Lancet 2016;388:2948–58. Smedley BD, Stith AY, Nelson AR. Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Washington, DC: National Academies Press, 2003. Parker RB, Stack SJ, Schneider SM, et al. Why diversity and inclusion are critical to the American College of Emergency Physicians’ future success. Ann Emerg Med 2017;69:714–7. James T. SAEM diversity position statement. The SAEM Diversity Interest Group. Acad Emerg Med 2000;7:1055. ADIEM. Academy for Diversity & Inclusion in Emergency Medicine. Available at: http://community. saem.org/communities/community-home?CommunityKey= a8136ffe-42c0-4939-956d-fa23d3b786f3. Accessed Mar 1, 2017. Supporting Information The following supporting information is available in the online version of this paper available at http:// onlinelibrary.wiley.com/doi/10.1111/acem.13269/full Data Supplement S1. Supplemental tables.