Uncategorized

Diversity of leadership in academic emergency medicine: Are we making progress?

a b s t r a c t

Background: Faculty who identify as women or racial/ethnic groups underrepresented in medicine (URiM) are less likely to occupy senior leadership positions or be promoted. Recent attention has focused on interventions to decrease this gap; thus, we aim to evaluate changes in leadership and academic promotion for these popula- tions over time.

Methods: Successive cross-sectional observational study of six years (2015 to 2020) of data from the Academy of Administrators/Association of Academic Chairs of Emergency Medicine- BenchMark Survey. Primary analyses fo- cused on gender/URiM differences in leadership roles and academic rank. Secondary analysis focused on dispar- ities during the first 10 Years of practice. statistical modeling was conducted to address the primary aim of assessing differences in gender/URiM representation in EM leadership roles/rank over time.

Results: 12,967 responses were included (4589 women, 8378 men). Women had less median years as faculty (7 vs 11). Women and URiM were less likely to hold a leadership role and had lower academic rank with no change over the study period. More women were consistently in the early career cohort (within 10 years or less as fac- ulty) : 2015 =-75.0% [95% CI:+- 3.8%] v 61.4% [95% CI:+- 3.0%]; 2020 =-75.1% [95% CI: +- 2.9%] v 63.3%, [95% CI::

+- 2.5%]. Men were significantly more likely to have any leadership role compared to women in 2015 and 2020 (2015 = 54.3% [95% CI: +- 3.1%] v 44.8%, [95% CI: +- 4.3%]; 2020 = 43.1% [95% CI:+- 2.5%] v 34.8 [95% CI:+- 3.1%]).

Higher academic rank (associate/professor) was significantly more frequent among early career men than women in 2015 (21.1% [95% CI:+- 2.58%] v 12.9%; [95% CI:+- 3.0%]) and 2020 (23.1% [95% CI:+- 2.2%] v 17.4%;

[95% CI:+- 2.5%]).

Conclusions: Disparities in women and URiM faculty leadership and academic rank persist, with no change over a six-year time span. Men early career faculty are more likely to hold leadership positions and be promoted to higher academic rank, suggesting early career inequities must be a target for future interventions.

(C) 2022

* Corresponding author.

E-mail addresses: [email protected] (J.A. Linden), [email protected] (J. Baird), [email protected] (T.E. Madsen), [email protected] (K. Rounds), [email protected] (M.D. Lall), [email protected] (N.P. Raukar), [email protected] (A. Fang), [email protected] (M. Lin), [email protected] (K. Sethuraman), [email protected]

(V.A. Dobiesz).

  1. Introduction

Despite consistent increases in women matriculating in medical schools (comprising more than half of medical school enrollment), mul- tiple studies reveal that women and faculty from racial and ethnic groups that are underrepresented in medicine (URiM) continue to be underrepresented in higher academic ranks and in senior leadership

https://doi.org/10.1016/j.ajem.2022.04.009

0735-6757/(C) 2022

positions, such as departmental chairs, deans, chief medical officers, and chief executive officers of Hospital systems [1-5]. Academic emergency medicine mirrors these disparities, with women continuing to be un- derrepresented in emergency medicine (EM) residencies (36%) and ac- ademic EM faculty (37%) [6-8]. Racial and ethnic populations that are underrepresented in medicine (URiM) continue to be poorly repre- sented in EM residencies (14%) and academic faculty (10%) [2,6,7].

Diversity in academic leadership is critical for increasing innovation, incorporating varied experiences and viewpoints, providing excellent patient care, and for mentoring underrepresented faculty, students and residents [9-11]. Successful women and URiM leaders and academi- cians serve as role models, encouraging students and residents from un- derrepresented groups to consider academics. Increasing women and underrepresented minorities in leadership may also increase represen- tation of these groups in EM faculty – one study reported that depart- ments with women chairpersons had a higher proportion of women faculty [12]. It is critical to track changes over time, in order to ensure that our specialty is making progress toward achieving the goal of in- creasing gender and underrepresented minority physician representa- tion in academic EM, especially in leadership and higher academic rank. This data can also help determine whether recent efforts to im- prove diversity and inclusion have been successful.

This study examines the temporal change in women and URiM composition with respect to leadership role and academic rank, over a six-year time span between 2015 and 2020. In addition, we examine the factors associated with leadership such as clinical hours worked and faculty experience (years since graduating residency). As a second- ary analysis, we focus on the first 10 years of professional practice following residency as the pipeline to leadership roles, specifically dif- ferences in rank and leadership experience by gender and URiM status, given recent evidence suggesting disparities in leadership attainment are associated with differences in early career opportunities [13,14].

  1. Methods
    1. Study design and setting/participants

This was a secondary analysis of six years of data from a successive cross-sectional survey (2015-2020) collected from academic emer- gency departments (EDs) participating in the annual Academy of Ad- ministrators in Academic Emergency Medicine /Association of Academic Chairs of Emergency Medicine (AAAEM/AACEM) Benchmark Survey. This study was deemed not human subjects research by the Bos- ton University Medical Center Institutional Review Board.

    1. Survey

The AAAEM/AACEM Benchmark Survey is completed on an annual basis by emergency department administrators and provides informa- tion about each physician in the department. Responses are de- identified as to individuals’ identities, but include the individuals’ gen- der, race, and ethnicity as reported by the departmental administrator. The methodology of this survey has been previously published [15]. The survey includes questions on departmental volume, operations, fac- ulty salaries, administrative and clinical roles, and annual hours of clin- ical practice. This survey was initiated in 2013, with additional questions regarding gender and underrepresented in medicine status (URiM) added in 2015. The survey includes validation checks including a review of data with significant change from prior years or data deemed out of range (with ranges established based on average re- sponses from prior years), educating responders with clear category definitions, and super-users who contact each respondent to ensure un- derstanding of these definitions. The most consistent category for all definitions is the role of “primary clinician with no leadership role”, as physicians with any leadership role would have been identified in one of the myriad roles available as a leadership role. If the role was not

identical to the listed options, responders were instructed to use the role that most closely represented the role, even if that role was more generic than the actual role, which captures any leadership role as op- posed to those with no leadership role

    1. Measurements

The following data for each survey year was extracted for the study: demographic characteristics of physicians at each survey site (gender, race, ethnicity), faculty rank, years as faculty, Fellowship training, annual clinical hours worked, and geographical region. Physicians’ race and ethnicity was recorded as URiM, to be consistent with the Association of American Medical Colleges (AAMC) definition. AAMC describes URiM status as ‘racial and ethnic populations that are underrepresented in the medical profession relative to their numbers in the general popu- lation’ [16]. White or Asian was coded as not URiM, with black/African American, or other race and/or Hispanic ethnicity as URiM. The primary outcome variable of interest – leadership role, was attained by re-coding survey responses of primary job duty into four main categories; no lead- ership role (primary position clinical only), operations leadership, educa- tion leadership (residency program/medical student education), research leadership, and executive leadership (chair/vice chair). This cod- ing was assigned based on consensus agreement from the study authors after discussion. If there was disagreement, this was discussed until con- sensus was reached

    1. Data Analysis

Data were de-identified for individual ED sites and downloaded into Microsoft Excel for primary data cleaning. Analysis was conducted using SAS(R) analysis software (Version 9.4, Carey, NC). Primary descriptive analysis was conducted prior to statistical modeling and reported as mean (with standard deviation [SD]), median (interquartile range [IQR]), and frequency (as percentages). Statistical analyses of demo- graphic characteristics were performed. UnAdjusted covariates em- ployed t-tests, Wilcoxon tests, and Pearson’s chi square test where appropriate. Data were excluded if the physician worked at primarily a pediatric or free-standing ED. We excluded pediatric and freestanding EDs as the composition of these subsets often varies from the larger cadre of academic EDs and due to small numbers of each in the data set. Data were also excluded if the primary job of the participant was not available

The survey data were reported on physicians across the survey years. Faculty were counted more than once in different years, if they re- mained at any of the participating AACEM/AAAEM institutions across the survey years. The survey methodology and de-identification of indi- vidual faculty did not allow individual physicians to be followed over time. To account for the correlation in these data we used a clustered regression model, with the ED site as the clustering effect. The effect of site variance was assessed by calculating the intraclass correlation (ICC). Adjusted logistic regression modeling was conducted to address the primary aim of the study: assessing differences in gender/URiM representation in EM leadership roles. With this predictive modeling, physicians’ gender (woman/man) was used as the primary predictor of leadership role. Academic rank (professor/associate professor versus other), years as faculty (median centered), fellowship training, base clinical hours (median centered), URiM status, and survey year were covariates in the series of Multivariable logistic regression models

    1. Missing data

An examination of patterns of missing data was conducted and de- termined to be missing at random. Pairwise deletion of categorical data was implemented to allow for maximum available data for all anal- ysis. Case wise deletion was used for data missing on gender. For contin- uous data, i.e., clinical hours worked and years as faculty, multiple

imputation techniques were utilized from available complete data to estimate missing values

  1. Results

Table 2

Demographic Characteristics of Study Sample by Gender Across All Survey Years (2015-2020)

Characteristic Women Men

In total, there were 13,034 physician data points from 76 unique ED sites that participated across all survey years. Gender was missing from

Race/Ethnicity

n = 4589 n (%)

n = 8378 n (%)

67 survey responses resulting in 12,967 physician data points being uti- lized in these analyses. Table 1 shows the number of sites participating

and the response rate for each year. Of the sites, 24 (32%) participated in all years, and 56 (74%) participated in 4 or more years of the survey. The

White

Multi Other*

3289

21

76

71.7%

0.5%

1.7%

6521

52

148

77.8%

0.6%

1.8%

average response rate across all survey years was 55.4%, with a range of

Missing

420

9.2%

710

8.5%

47.5 to 64.5%

Hispanic**

132

2.9%

201

2.4%

Table 2 shows the demographic characteristics of the sample. Across survey years 35.4% were identified as women, and this was consistent

URiM***

Academic Rank Instructor

382

427

8.3%

9.3%

442

605

5.3%

7.2%

across survey years with the proportion of women ranging from 33.1

Assistant Professor

2883

62.8%

4266

50.9%

to 37.1%. Most ED physicians were identified as white (71.7% women,

Associate Professor

838

18.3%

2055

24.5%

77.8% men), with significantly more women identified as Black or

Professor

291

6.3%

1248

14.9%

African American (5.4% [95%CI: +- 0.9%] vs. 2.9%, [95%CI: +- 0.5%]) and

Geographic Region

Midwest

886

19.3%

1907

22.8%

as URiM (8.3% [95%CI: +- 0.8%] vs. 5.3% [95%CI: +- 0.5]). Men had a

Northeast

1810

39.4%

2932

35.0%

higher median number of years as faculty compared with women (11

South

1035

22.6%

2014

24.0%

[IQR 5,19] vs. 7 [IQR 4,13] p < 0.001), and were more likely to be asso-

West

858

18.7%

1525

18.2%

ciate/full professors (39.4% [95%CI: +- 1.9%] vs. 24.6 [95%CI: +- 1.2%]). Median annual clinical hours worked were higher for women (998,

Leadership role None

Education

2863

832

62.4%

18.1%

4291

1103

51.2%

13.2%

IQR 735, 1250) than men (960, IQR 670, 1248, p = 0.01), but clinical

Operations

330

7.2%

1075

12.8%

hours worked did not differ by URiM status (URiM median = 960

Executive

474

10.3%

1676

20.0%

[IQR 693,1237], non-URiM median = 960 [IQR 691, 1237], p = 0.74).

Research

68

1.5%

184

2.2%

Asian 533 11.6% 703 8.4%

Black 250 5.4% 244 2.9%

Physicians identified as women or URiM had less median years as fac- ulty compared to men and non-URiM physicians (women and URiM

= 6 [IQR 2,11] vs. men and non-URiM = 7 [IQR 3,14], p < 0.004). Al-

though all included sites were academic, 5.9% of physician data points (n = 764) were reported as having worked at a community ED

    1. Leadership roles

Fig. 1 shows the gender representation of four leadership roles be- tween 2015 and 2020. In both years, women were significantly more likely than men to have no leadership role (mean = 60.2% [95%CI: +- 3.3%] vs. 51.3% [95%CI: +- 2.8 %]). This gender difference in leadership role was significant across all survey years, with women in no leader- ship role ranging from 53.5% to 63.5%, and correspondingly 47.6% to 58.9% for men. Of those in leadership roles, women were significantly more likely than men to have an education leadership role across all survey years (17.7% [95%CI: +- 1.1%] vs. 13.1% [95%CI: +- 0.7%]), while men were more likely to have an operations leadership role (12.5% [95%CI: +- 1.0%] vs. 7.2% [95%CI: +- 1.1%]) or an executive leadership role (19.9% [95%CI: +- 0.9%] vs. 10.2% [95%CI: +- 0.9%]). A series of

Cochrane-Armitage trend tests were conducted to determine if there were any significant trends over time by gender for any leadership role compared to no leadership role, as well as trends over time by gen- der for each type of leadership role. The proportion of gender represen- tation across all leadership roles in this sample of academic EDs did not significantly change over time

Table 1

Academic Emergency Departments Sites Participating in the AAAEM/AACEM Benchmark Survey by Year.

Survey Year

Total # of AAAEM/AACEM sites

AAAEM/AACEM sites participating in the Benchmark Survey

2015

99

47 (47.4%)

2016

112

67 (59.8%)

2017

107

69 (64.5%)

2018

118

68 (57.6%)

2019

118

62 (52.5%)

2020

117

59 (50.4%)

Fellowship Training

1693

36.9%

2919

34.8%

Median

IQR

Median

IQR

Base Clinical Hours/year

998

735,1250

960

670,1248

Years @ Faculty

7

4,13

11

5,19

IQR = interquartile range; *other: Native Hawaiian/Pacific Islander/American Indian/Alas- kan native; ** separate question from Race (numbers are a subset of the race totals; URiM*** = underrepresented in medicine = combination of Black + Hispanic (not dupli- cated with Racial Identity).

In examining the effect of URiM status in leadership representation (Fig. 2), in 2015 and 2020, those who were identified as URiM were sig- nificantly more likely than non-URiM to have no leadership role (57.2% [95%CI: +- 3.4%] vs. 53.5% [95%CI: +- 0.9%]). Across survey years URiM identified faculty were less likely to be in executive roles (URiM = 14.3% [95% CI: +- 2.9%] vs. 17.6% [95% CI: +- 0.7%]) and education leader- ship roles (URiM = 12.8% [95% CI: +- 2.2%] vs. 15.3% [95% CI: +- 0.6%])

Image of Fig. 1

Fig. 1. Types of Leadership Role by Gender, 2015 and 2020.*

* Leadership roles by gender for all years from 2015 to 2020 are shown in Appendix A. Cochrane-Armitage trend tests demonstrated no significant change over time by gender for each leadership role, including no leadership role

Image of Fig. 2

Image of Fig. 4Fig. 2. Types of Leadership Role by Underrepresented in Medicine Status, 2015 and 2020.* Leadership roles by URiM for all years from 2015 to 2020 are shown in Appendix B

    1. Academic rank

Differences in academic rank by gender and URiM status in 2015 and 2020 are shown in Figs. 3 and 4, respectively. Across survey years phy- sicians who were identified as men were significantly more likely to have academic ranks at professor or associate professor level (men = 40.3% [95% CI: +- 1.1%] vs. women = 25.3% [95% CI: +- 1.3%]), as were

non-URIM (non-URiM 36.7% [95% CI: +- 0.9%] vs. URiM 25.2% [95% CI:

+- 3.0%]. A series of Cochrane-Armitage trend tests showed that the trends in difference between men and women and URiM and non- URiM did not significantly change over time

    1. Predictive modeling

The results of the clustered multivariable logistic regression models of leadership roles are displayed in Table 3. For these analyses only those physicians identified as having a leadership role at any survey year were included. Three models were analyzed examining the effects

Image of Fig. 3

Fig. 3. Academic Rank by Gender, 2015 and 2020.

      • Academic rank by gender for all years from 2015 to 2020 are shown in Appendix C. Cochrane-Armitage trend tests demonstrated no significant change over time by gender for each academic rank

Fig. 4. Academic Rank by Underrepresented in Medicine Status, 2015 and 2020.

      • Academic rank by URiM status for all years from 2015 to 2020 are shown in Appendix C. Cochrane-Armitage trend tests demonstrated no significant change over time by URiM for each academic rank

of gender on the log odds of holding an executive vs other leadership role, operations versus other leadership roles, and education versus other leadership roles, while controlling for years as faculty, faculty rank, base clinical hours, and URiM status. The effect of change over time was also included. The ICC for each model was calculated to assess the effect of physicians clustered within each site on the variance ex- plained in these models

For Model 1, after adjusting for all other variables in the model, women had significantly lower odds of being in an executive role (ap- proximately 45% lower odds), compared with other leadership roles, than men (AOR = 0.55 [95% CI 0.48-0.64]). The odds of a physician hav- ing an executive leadership role, compared to other roles, was signifi- cantly lower if the physician was at the academic rank of assistant professor or instructor (AOR = 0.54 [95% CI 0.44-0.66]), but URiM sta- tus did not have a significant effect on having an executive leadership role (AOR = 0.94 [95% CI 0.79-1.20]). These results were also reflected in the model assessing the odds of women being in operations leader- ship roles compared to men (AOR = 0.63 [95% CI 0.53-0.73]). The third model showed that after adjusting for all other variables in the model, women were 2.5 times more likely than men to have an educa- tional leadership role (AOR = 2.58 [95% CI 2.25-2.95]). The effect of time was also included in these three models and demonstrated that, in comparison to the first year of study data (2015), the log odds of time on the model was not significant. The ICC is shown for each model, and while significant, these values show that the effect of physi- cians clustered within site accounts for a small proportion of the vari- ance explained in the model (executive model = 5%, operations model = 6%, education model = 2%). The number of URiM physicians across the leadership domains by survey year was insufficient to include this as a primary predictor in the statistical modeling

    1. Leadership and academic rank in early career stages

As can be seen in Table 4 disproportionately more women are in an early career cohort (with 10 years or less as faculty) compared to men, in 2015 and 2020 (2015: women = 75.0% [95% CI:: +- 3.8%] vs. men = 61.4% [95% CI:: +- 3.0%]; 2020 women = 75.1% [95% CI:: +- 2.9%] vs. men

= 63.3%, [95% CI:: +- 2.5%]). Among these early career cohorts, men were significantly more likely to have any leadership role compared to women in both 2015 and 2020 (2015 = 50.1% [95% CI: +- 3.1%] vs.

Table 3

Multivariable logistic regression, leadership roles among academic emergency medicine faculty.

Predictor (Referent)

Model 1

Executive Leadership vs other

Model 2

Operations Leadership vs other

Model 3

Education Leadership vs other

Odds ratio (95% CI),

Odds ratio (95% CI),

Odds ratio (95% CI),

Gender (Male)

0.55 (0.48, 0.64)

0.63 (0.53, 0.73)

2.58 (2.25, 2.95)

Years at faculty (Median centered)

1.02 (1.01, 1.03)

1.03 (1.02, 1.04)

0.96 (0.95, 0.97)

Academic rank (Professor/Associate)

0.54 (0.44, 0.66)

0.52 (0.44, 0.62)

0.71 (0.60, 0.83)

Base clinical hours- median centered

1.00 (0.98, 0.99)

1.00 (1.00, 1.01)

1.00 (1.00, 1.01)

URiM Status (non URiM)

0.94 (0.79, 1.20)

1.02 (0.78, 1.34)

0.79 (0.61, 1.03)

Time (2015)

2016

1.05 (0.87, 1.26)

0.98 (0.76, 1.25)

1.12 (0.88, 1.41)

2017

0.91 (0.76, 1.09)

1.07 (0.83, 1.36)

1.10 (0.87, 1.38)

2018

0.98 (0.82, 1.17)

0.98 (0.77, 1.25)

1.07 (0.85, 1.35)

2019

0.84 (0.70, 1.01)

0.94 (0.73, 1.21)

1.18 (0.93, 1.51)

2020

0.81 (0.68, 0.97)

0.91 (0.71, 1.18)

1.37 (1.09, 1.74)

ICC site effect

0.05, p < 0.001

0.06, p < 0.001

0.02, p = 0.01

41.1%, [95% CI: +- 4.3%]; 2020 = 38.2 [95% CI: +- 2.5%] vs. 31.1 [95% CI: +-

3.1%]). Higher academic rank (associate or professor) was also signifi- cantly more frequent among early career men than women in both 2015 (21% [95% CI: +- 2.58%] vs. 13%; [95% CI: +- 3.0%]) and 2020

(23.1% [95% CI: +- 2.2%] vs. 17.4%; [95% CI: +- 2.5%])

Representation of early career physicians identified as URiM are shown in Table 4. Analysis was conducted to compare early career phy- sicians identified as URiM and those not. URiM physicians were not sig- nificantly more likely to be in the early stages of their career compared to non-URiM in 2015 (73.6% [95% CI: +- 9.1%] vs. 66.5% [95% CI: +- 2.61%] or in 2020 (74.3% [95% CI: +- 6.61%] vs. 67.3% [95% CI: +- 2.1%]). In 2015

and 2020 URiM and non-URiM physicians were not significantly differ- ent in proportions who had no leadership role (2015 = 61.2% [95% CI: +- 10.0%] vs. 52.0% [95% CI: +- 2.7%; 2020 = 66.9% [95% CI: +- 7.0%] vs.

66.5% [95% CI: +- 1.8%]). There were no significant differences in types of leadership roles by physician URiM status in 2015 or 2020. Similarly, in 2015 and 2020, the proportional representation of URiM physicians across academic ranks was not significantly different

Table 4

Early Career (time at faculty <10 years) Physician Characteristics in 2015 and 2020 by Gender and URiM Status

Gender

2015

2015

2020

2020

Women n (%)

Men n (%)

Women n (%)

Men n (%)

Early Career

371 (75.0)

613 (61.4)

634 (75.1)

908 (63.3)

* Leadership role

None

216 (58.2)

306 (49.9)

437 (68.9)

552 (61.8)

Education

76 (20.5)

103 (16.8)

101 (15.9)

123 (13.6)

Operations

25 (6.7)

82 (13.4)

35 (5.5)

95 (10.5)

Executive

43 (11.6)

113 (18.4)

45 (7.1)

115 (12.7)

*Rank

Instructor/Assistant

322 (87.1)

484 (78.9)

509 (82.6)

683 (77.0)

Associate

39 (10.5)

92 (15.0)

96 (15.6)

158 (17.7)

Professor

9 (2.4)

37 (6.1)

11 (1.8)

47 (5.3)

URiM Status

2015

2015

2020

2020

n (%)

URiM

Non-URiM

URiM

Non-URiM

n = 91

n = 1270

n = 171

n = 1864

??Early Career

67 (73.6)

845 (66.5)

127 (74.3)

1254 (67.3)

*Leadership role

None

41 (61.2)

439 (52.0)

85 (66.9)

791 (63.1)

Education

7 (10.5)

161 (19.1)

17 (13.4)

188 (15.0)

Operations

6 (9.0)

95 (11.2)

10 (7.9)

107 (8.5)

Executive

10 (15.0)

134 (15.9)

12 (9.5)

134 (10.7)

*Rank

Instructor/Assistant

58 (86.6)

635 (81.6)

99 (79.2)

962 (79.0)

Associate

9 (13.4)

114 (13.5)

25 (20.0)

207 (17.0)

Professor

0

41 (4.9)

1 (0.80)

49 (4.0)

?? Missing data 2015 = 17, 2020 missing data = 14.

  1. Discussion

Our findings demonstrate a lack of progress and persistent un- derrepresentation of women in academic EM leadership positions, and women and URiM in higher academic ranks among a large dataset of faculty in academic Emergency Medicine from 2015 through 2020. While disparities in leadership are well documented, our findings are novel because we identify two key drivers of persis- tent inequities for women: 1) gender disparities in leadership attain- ment among early career faculty (within their first ten years of practice) and 2) differences in leadership roles held by men and women emergency medicine faculty.

While the “first promotion” gap is a well identified phenomenon across industries, our study is among the first to examine leadership dif- ferences within the first ten years of practice. One recent study, also identified disparities in promotion of women and URiM which begins early in their careers, at the first tier of promotion [17]. Differences in ca- reer accomplishments are less likely to explain gender differences in leadership positions earlier in the career trajectory. In fact, women are more likely than men to have completed ACGME-accredited emergency medicine fellowships [18]. Gender differences in early career leadership and promotion may be partially explained by women receiving less mentorship, sponsorship, and career advancing opportunities, disad- vantageous institutional parental leave and promotion policies, implicit biases, gender discrimination, sexual harassment, and workplace cul- ture and environment. [19-27] Implicit bias that values leadership po- tential over performance during hiring may also be a contributing factor that preferentially benefits men and impedes women, a phenom- enon well documented in organizational science. Women are often val- ued higher for prior/current leadership performance, which may hinder advancement into new career paths or leadership domains beyond their current roles [28]. In prioritizing leadership potential over performance during hiring, men are considered for leadership roles (often in opera- tions or research domains), where they may have little to no prior expe- rience. These positions often come with greater protected time and/or a higher salary at the start of their academic careers putting them at a career-long advantage compared to women of the same rank and abil- ity. This and another study have documented fewer average clinical hours worked by men academic emergency physicians [despite higher pay] [29]. This additional protected nonclinical time facilitates increased academic productivity contributing to further leadership attainment and academic promotion and widening of this gender gap. The lack of early career leadership is the “broken first rung” in a ladder toward ac- ademic leadership; when women and URiM are excluded or start as- cending later, their trajectory will be slower, and they may be more likely to leave academic emergency medicine altogether. Prior studies have shown that URiM are less likely to be promoted and retained in ac- ademic medicine [30].

Early career disparities in leadership roles are more likely to accrue and have compounded effects over time, exacerbating disparities in promotions, salary increases, and leadership opportunities that con- tinue throughout the academic career. Potential solutions to address bias in early career leadership attainment and promotion are well doc- umented and include but are not limited to: standardized recruitment practices for leadership positions at all ranks, including public posting of leadership opportunities, implicit bias training for search committees and academic chairs, salary and hours transparency, paid family leave, and equitable sponsorship [20,21].

This study shows that the number of both women and men in leader- ship positions decreased from 2015 to 2020, presumably due to increased faculty, with a fixed number of leadership positions. Some have argued for leadership position term limits, to allow more opportunity for women and URiM, who are more likely to be early career faculty [31]. Cur- rently, the average term length for department chairs and deans is 8.7 years for men, and 6.1 years for women. For those in leadership positions for greater than 12 years, only 7.3% are women, and 9.5% are non-white/ non-Asian [31]. Leadership term limits, as recently instituted by the Na- tional Institutes of Health [32] can lead to more rapid turnover, typically every 6-8 years, enabling a diversity of leaders, ideas, and leadership styles. The long term impact of term limits for increasing diversity in lead- ership has not yet been studied in academic medicine, but has been stud- ied in business and politics with mixed results [33-35].

Our findings also show that among academic emergency physicians in leadership, women are far more likely to be in educational positions (residency director or clerkship director) than in other leadership posi- tions, while men are more likely to be in Operational leadership roles (clinical chief, ED director, chair) than any other role. We found no dif- ference by URiM status. These findings are consistent with recent work demonstrating men chairs in academic EDs were more likely to have held operational roles, while women were more likely to have held ed- ucational leadership positions prior to becoming chair [13]. While these differences may be explained by individual choice, they may also be attributed to stereotyping women into more nurturing educational roles by senior faculty responsible for appointing faculty to leadership positions. These differences are notable because academic chairs interviewed in prior studies have suggested that teaching and educa- tional contributions have lower value for academic promotion relative to other forms of scholarly productivity [20]. In addition, education leadership roles are associated with lower salaries after adjusting for clinical hours and lower salary growth over time relative to opera- tions and executive leadership roles [29] suggesting that differences in roles exacerbates gender-based pay disparities. Taken together, these findings suggest institutions prioritizing diverse leadership should equitably recruit and sponsor women and URiMs for operations and executive leadership roles

Our study expands on existing literature demonstrating persistent gaps in leadership attainment, promotion, and salary among women academic emergency physicians. Results from an analysis of a 2015 AAAEM/AACEM Benchmark Survey revealed that while women made up 33% of the workforce in academic EM, they were under- represented at the full professor (7% vs. 17%) and associate professor rank (19% vs. 24%) as compared to men, and over-represented in the in- structor/assistant rank (74% vs. 59%) [36]. A recent article utilizing the same dataset documented decreased leadership roles, increased clinical hours worked, and pay gaps ranging from $17,703 to $54,609 for women depending on the leadership role [15]. Differences in starting salary between men and women physicians are difficult to justify and were recently found to accrue up to $2 million in lost earnings among women physicians over their careers [37]. Our findings suggest that fo- cusing on early career equity in leadership opportunities and diversity of roles for women and URiM faculty, in addition to other evidence- based strategies, may help reduce disparities

There are some limitations to our study. While the number of aca- demic programs participating in the AAAEM/AACEM Benchmark Survey

increased each year, this may not be representative of the general pop- ulation of academic EDs. The proportion of women in this cohort (35.4%) is consistent with that reported in EM by the AAMC (36.9%) and other studies [6,7]. This study encompasses a relatively short win- dow in time for academic careers, and perhaps over a longer period of time, there will be greater ability to detect statistically significant changes. Participants likely were counted more than once, if they remained at one of the participating programs, however, we analyzed each survey year separately in all descriptive statistics and adjusted for clustering by year in all regression models. Given the anonymous nature of the de-identified data, we are unable to assess an individual’s change in academic rank or leadership position over time. We were unable to delve into the research leadership role in a meaningful way due to the low numbers of research leaders for both genders reported in this cohort. We did not include some education executive leadership roles (for example, Dean and Designated Institutional Officer), since these were not collected in the AAAEM/AACEM Benchmark Survey data. Physicians practicing at exclusively pediatric ED sites were excluded from our analysis due to the small numbers; since Pediatric EM providers are more likely to be women (615 women; 445 men from the AAAEM/AACEM Benchmark Survey), the leadership character- istics may be different than general EM and is a ripe area for further research

  1. Conclusions

This study found continued disparities in leadership and promotion for women and URiMs in academic emergency medicine over a six-year time span. Women are significantly more likely to have no leadership role or an education role. Furthermore, these disparities start within the first 10 years of the career and persist over time. Additional research, effort, and resources are needed to identify and implement effective strategies early in the careers of underrepresented faculty to mitigate these persistent barriers to the retention and advancement of women and URiMs in academic emergency medicine

Author contribution

JAL, JB, TM, KR, MDL, MPL, NPR, AF, ML, KS, VAD conceived and de-

signed the study, JB and TM performed statistical analysis, JAL, KR, MDL, JB,

Credit authorship contribution statement

Judith A. Linden: Supervision, Writing – original draft, Writing – review & editing, Conceptualization, Data curation, Investigation, Meth- odology. Janette Baird: Formal analysis, Writing – review & editing, Methodology, Investigation, Data curation, Conceptualization. Tracy E. Madsen: Methodology, Writing – review & editing, Investigation, Conceptualization, Data curation, Formal analysis. Kirsten Rounds: Data curation, Conceptualization, Investigation, Methodology, Writing – review & editing. Michelle D. Lall: Writing – review & editing, Meth- odology, Investigation, Data curation, Conceptualization. Neha P. Raukar: Writing – review & editing, Data curation, Investigation. Andrea Fang: Writing – review & editing, Investigation. Michelle Lin: Writing – review & editing, Formal analysis, Data curation, Conceptual- ization. Kinjal Sethuraman: Conceptualization, Data curation, Writing – review & editing. Valerie A. Dobiesz: Writing – review & editing, Writing – original draft, Methodology, Investigation, Data curation, Conceptualization.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influ- ence the work reported in this paper

Appendix A: Leadership Role by Gender Across Survey Years

Leadership Role

2015

2016

2017

2018

2019

2020

N

%

N

%

N

%

N

%

N

%

N

%

1494

100

2004

100

2417

100

2341

100.

2207

100.

2256

100.

Executive

293

20%

370

18%

397

16%

419

18%

343

16%

308

14%

Women

67

13.5

76

11.1%

86

10.2%

86

10.3%

80

9.8

73

8.8%

Men

226

22.6%

294

22.2%

311

19.7%

333

22.1%

263

18.9%

235

16.5%

Operations

186

12%

220

11%

277

11%

266

11%

220

10%

208

9%

Women

38

7.7%

45

6.6%

72

8.6%

70

8.4%

54

6.6%

53

6.4%

Men

148

14.8%

175

13.2%

205

13.0%

196

13.0%

166

11.9%

155

10.9%

Education

246

16%

318

16%

359

15%

347

15%

299

14%

337

15%

Women

103

20.8%

131

19.2%

147

17.5%

149

17.8%

131

16.1%

150

18.0%

Men

143

14.3%

187

14.1%

212

13.4%

198

13.2%

168

12.1%

187

13.1%

Research

39

3%

32

2%

53

2%

39

2%

35

2%

50

2%

Women

14

2.8%

10

1.5%

13

1.5%

9

1.1%

9

1.1%

14

1.7%

Men

25

2.5%

22

1.7%

40

2.5%

30

2.0%

26

1.9%

36

2.5%

None

730

49%

1064

53%

1331

55%

1270

54%

1310

59%

1353

60%

Women

273

55.2%

420

61.6%

522

62.1%

523

62.5%

540

66.3%

543

65.2%

Men

457

45.7%

644

48.7%

809

51.3%

747

49.7%

770

55.3%

810

56.9%

Appendix B: Leadership Role by URiM status Across Survey Years

LeadershipRole

2015

2016

2017

2018

2019

2020

N

%

N

%

N

%

N

%

N

%

N

%

1399

100%

1741

100%

2074

100%

2055

100%

1765

100%

2042

100.0%

Executive

279

20%

340

20%

359

17%

380

18%

291

16%

289

14%

URiM

19

19.8%

22

19.3%

21

13.%

22

14.8%

17

11.5

21

12.2%

Non-URiM

260

20.0%

318

19.5%

338

17%

358

18.8%

274

16.%

268

14.3%

Operations

177

13%

195

11%

235

11%

227

11%

182

10%

187

9%

URiM

8

8.3%

10

8.8%

22

13.%

14

9.4%

17

11.5%

13

7.6%

Non-URiM

169

13.0%

185

11.4%

213

11.%

213

11.2%

165

10.2%

174

9.3%

Education

234

17%

277

16%

307

15%

307

15%

253

14%

313

15%

URiM

14

14.6%

16

14.0%

20

12.%

20

13.4%

14

9.5%

25

14.5%

Non-URiM

220

16.9%

261

16.0%

287

15.%

287

15.1%

239

14.8%

288

15.4%

Research

37

3%

27

2%

47

2%

36

2%

33

2%

46

2%

URiM

3

3.1%

5

4.4%

9

5.6%

7

4.7%

6

4.1%

7

4.1%

Non-URiM

34

2.6%

22

1.4%

38

2.0%

29

1.5%

27

1.7%

39

2.1%

None

672

48%

902

52%

1126

54%

1105

54%

1006

57%

1207

59%

URiM

52

54.2%

61

53.5%

88

55%

86

57.7%

94

63.5%

106

61.6%

Non-URiM

620

47.6%

841

51.7%

1038

54.2%

1019

53.5%

912

56.4%

1101

58.9%

Appendix C: Faculty Rank Separated by Gender and URiM across Survey Years

Faculty Rank

2015

N

%

2016

N

%

2017

N

%

2018

N

%

2019

N

%

2020

N

%

1491

2201

2409

2356

2152

2231

Professor

202

13.6

278

13.8

262

10.9

275

11.6

263

12.2

266

12

Women

30

6.1

52

7.6

50

5.9

50

5.9

52

6.6

57

6.9

Men

172

17.2

226

17.0

212

13.6

222

14.8

211

15.4

209

14.8

URiM

5

5.2

8

6.8

14

9

10

6.7

8

5.6

13

7.7

Non-URiM

181

13.9

249

15.2

223

11.7

232

12.2

203

12.9

229

12.4

Associate Professor

330

22.1

454

22.5

514

21.3

539

22.8

504

23.4

549

24.6

Women

91

18.5

112

16.3

143

16.9

161

19.1

155

19.8

179

21.8

Men

239

23.9

342

25.7

371

23.7

378

24.9

349

25.5

370

26.2

URiM

21

22

19

16.2

20

12.8

29

19.5

22

15.5

40

23.5

Non-URiM

294

22.6

387

23.6

443

23.2

440

23.2

389

24.8

458

24.8

Assistant

838

56

1137

57

1397

58

1356

57.3

1180

54.3

1242

55.7

Women

320

65

454

66

552

65.3

551

65.3

493

63

521

63.5

Men

515

51.6

683

51.3

845

54

805

52.9

687

50.2

721

51.1

URiM

59

61.5

78

66.7

105

67.3

98

65.8

91

64.1

98

57.7

Non-URiM

725

55.8

894

54.5

1057

55.4

1075

56.7

891

56.8

1053

57

Instructor

118

8.3

151

7.5

236

9.8

195

8

205

9.5

174

7.8

Women

51

10.4

70

10.2

100

11.8

82

9.7

83

10.6

64

7.8

Men

73

7.3

81

6.1

136

8.7

113

7.4

122

8.9

110

7.8

URiM

11

11.5

12

10.3

17

10.9

12

8.1

21

14.8

19

11.2

Non-URiM

100

7.7

112

6.8

185

9.7

148

7.8

86

5.5

107

5.8

References

  1. AAMC. Table B-3: Total U.S. Medical School Enrollment by Race/Ethnicity (Alone) and Sex, 20162017 through 2020-2021. AAMC; 2020. Accessed November 22, 2021. https://www.aamc.org/media/6116/download.
  2. AAMC. U.S. Medical school Faculty: Table 19: Distribution of Full Time U.S. Medical School Faculty by Sex, Race,/Ethnicity, Rank and Department. Accessed February 1, 2021. https://www.aamc.org/media/8901/download; 2020.
  3. Bennett CL, Raja AS, Kapoor N, et al. Gender differences in faculty rank among aca- demic emergency physicians in the United States. Acad Emerg Med. 2019;26(3): 281-5. https://doi.org/10.1111/acem.13685.
  4. Lautenberger D, Raezer C, Bunton S. The underrepresentation of women in leader- ship positions at U.S. medical schools. AAMC Analys Brief. 2015.;15(2).
  5. Lautenberger D, Dandar V. The state of women in academic medicine 2018-2019. AAMC; 2020. Accessed December 17, 2020. https://store.aamc.org/the-state-of- women-in-academic-medicine-2018-2019-exploring-pathways-to-equity.html.
  6. Nelson LS, Keim SM, Baren JM, et al. American Board of Emergency Medicine Report on residency and fellowship training information (2017-2018). Ann Emerg Med. 2018;71(5):636-48. https://doi.org/10.1016/j.annemergmed.2018.03.037.
  7. ACGME. ACGME Data Resource Book, Academic year 2019-2020. Accessed Decem- ber 23, 2020. https://www.acgme.org/About-Us/Publications-and-Resources/ Graduate-Medical-Education-Data-Resource-Book.
  8. AAMC. Table 13: Medical School Faculty Be Sex, Rank, and Department 2018. Accessed January 3, 2021. https://www.aamc.org/system/files/2020-01/201 8Table13.pdf.
  9. 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(6):714-7. https://doi.org/10.1016/j.annemergmed.2016.11.030.
  10. Alsan M, Garrick O, Graziani GC. Does Diversity Matter for Health? Experimental Ev- idence from Oakland. National Bureau of Economic Research; 2018..
  11. Greenwood BN, Carnahan S, Huang L. Patient-physician gender concordance and in- creased mortality among female heart attack patients. Proc Natl Acad Sci U S A. 2018;115(34):8569-74. https://doi.org/10.1073/pnas.1800097115.
  12. Cheng D, Promes S, Clem K, Shah A, Pietrobon R. Chairperson and faculty gender in academic emergency medicine departments. Acad Emerg Med. 2006;13(8):904-6. https://doi.org/10.1197/j.aem.2006.01.025.
  13. Sethuraman KN, Lin M, Rounds K, et al. Here to chair: gender differences in the path to leadership. Acad Emerg Med. 2021;28(9):993-1000. https://doi.org/10.1111/ acem.14221.
  14. McKinsey & Company. Women in the Workplace Report. Accessed November 9, 2021. https://wiw-report.s3.amazonaws.com/Women_in_the_Workplace_2021.pdf.
  15. Wiler JL, Rounds K, McGowan B, Baird J. Continuation of gender disparities in pay among academic emergency medicine physicians. Choo EK, ed. Acad Emerg Med. 2019;26(3):286-92. https://doi.org/10.1111/acem.13694.
  16. AAMC. Association of American Colleges. Underrepresented in Medicine Definition. Accessed December 30, 2020. https://www.aamc.org/initiatives/urm/.
  17. Hobgood C, Fassiotto M. Using the rank equity index to measure emergency medi- cine faculty rank progression. Acad Emerg Med. 2021;28(9):966-73. https://doi. org/10.1111/acem.14268.
  18. Partiali B, Oska S, Turner-Lawrence D. An analysis of sex diversity in ACGME emer- gency medicine fellowships. Am J Emerg Med. 2021;46:720-1. https://doi.org/10. 1016/j.ajem.2020.10.015.
  19. Agrawal P, Madsen TE, Lall M, Zeidan A. Gender disparities in academic emergency medicine: strategies for the recruitment, retention, and promotion of women. AEM Educ Train. 2020;4(Suppl. 1):S67-74. https://doi.org/10.1002/aet2.10414.
  20. Madsen TE, Heron SL, Rounds K, et al. Making promotion count: the gender perspec- tive. In: Kline JA, editor. Academic Emergency Medicine; 2019 https://doi.org/10. 1111/acem.13680. Published online February 19.
  21. Choo EK, Kass D, Westergaard M, et al. The development of best practice recommen- dations to support the hiring, recruitment, and advancement of women physicians in emergency medicine. Acad Emerg Med. 2016;23(11):1203-9. https://doi.org/10. 1111/acem.13028.
  22. Lu DW, Lall MD, Mitzman J, et al. #MeToo in EM: a multicenter survey of academic emergency medicine faculty on their experiences with gender discrimination and sexual harassment. West J Emerg Med. 2020;21(2):252-60. https://doi.org/10. 5811/westjem.2019.11.44592.
  23. Roth VR, Theriault A, Clement C, Worthington J. Women physicians as healthcare leaders: a qualitative study. J Health Organ Manag. 2016;30(4):648-65. https:// doi.org/10.1108/JHOM-09-2014-0164.
  24. Graham EM, Ferrel MN, Wells KM, et al. Gender-based barriers to the advancement of women in academic emergency medicine: a multi-institutional survey study. West J Emerg Med. 2021;22(6):1355-9. https://doi.org/10.5811/westjem.2021.7. 52826.
  25. Yedidia MJ, Bickel J. Why aren’t there more women leaders in academic medicine? The views of clinical department chairs. Acad Med. 2001;76(5):453-65. https:// doi.org/10.1097/00001888-200105000-00017.
  26. Hobgood C, Draucker C. Gender differences in experiences of leadership emergence among emergency medicine department chairs. JAMA Netw Open. 2022;5(3): e221860. https://doi.org/10.1001/jamanetworkopen.2022.1860.
  27. Riano NS, Linos E, Accurso EC, et al. Paid family and childbearing leave policies at top US medical schools. JAMA. 2018;319(6):611-4. https://doi.org/10.1001/jama.2017. 19519.
  28. Player A, Randsley de Moura G, Leite AC, Abrams D, Tresh F. Overlooked leadership potential: the preference for leadership potential in job candidates who are men vs women. Front Psychol. 2019;10:755. https://doi.org/10.3389/fpsyg.2019.00755.
  29. Wiler JL, Wendel SK, Rounds K, McGowan B, Baird J. Salary disparities based on gen- der in academic emergency medicine leadership. Acad Emerg Med. 2022;29(3): 286-93. https://doi.org/10.1111/acem.14404. Epub 2021 Dec 9. PMID: 34689369.
  30. Kaplan SE, Raj A, Carr PL, Terrin N, Breeze JL, Freund KM. Race/ethnicity and success in academic medicine: findings from a longitudinal multi-institutional study. Acad Med. 2018;93(4):616-22. https://doi.org/10.1097/ACM.0000000000001968.
  31. Beeler WH, Mangurian C, Jagsi R. Unplugging the pipeline – a call for term limits in academic medicine. N Engl J Med. 2019;381(16):1508-11. https://doi.org/10.1056/ NEJMp1906832.
  32. Kaiser J. NIH limits reign of chiefs. Science. 2019;364(6439):423. https://doi.org/10. 1126/science.364.6439.423.
  33. Carroll SJ, Jenkins K. Do term limits help women get elected? Soc Sci Q. 2001;82(1): 197-201. https://doi.org/10.1111/0038-4941.00017.
  34. Rosenblum D, Nili Y. Board Diversity by Term Limits. Ala. L. Rev, 71; 2021; 211 [boards]. Published online 2019.
  35. Schraufnagel S, Halperin K. Term limits, electoral competition, and representational diversity: the case of Florida. State Politics Policy Q. 2006;6(4):448-62. https://doi. org/10.1177/153244000600600405.
  36. Madsen TE, Linden JA. Rounds K, et al. current status of gender and racial/ethnic dis- parities among academic emergency medicine physicians. Choo EK, ed. Acad Emerg Med. 2017;24(10):1182-92. https://doi.org/10.1111/acem.13269.
  37. Whaley CM, Koo T, Arora VM, Ganguli I, Gross N, Jena AB. Female physicians earn an estimated $2 million less than male physicians over a simulated 40-year career: study examines estimated career gap in pay between female physicians and male physicians. Health Aff. 2021;40(12):1856-64. https://doi.org/10.1377/hlthaff.2021. 00461.