Perceived Discrimination Among Surgical Residents at Academic Medical Centers

Abstract

Background:

Residents of color experience microaggressions in the work environment, are less likely to feel that they fit into their training programs and feel less comfortable asking for help. Discrimination has been documented among surgical residents, but has not been extensively studied and largely remains unaddressed. We sought to determine the extent of perceived discrimination among general surgery residents.

Materials and Methods:

Residents who were enrolled in a randomized controlled trial investigating a cultural dexterity curriculum completed baseline assessments prior to randomization that included demographic information and the Everyday Discrimination Scale (EDS). Data from the baseline assessments were analyzed for associations of EDS scores with race, ethnicity, sex, socioeconomic level, language ability, and training level.

Results:

Of 266 residents across 7 residency programs, 145 (55%) were men. Racial breakdown was 157 (59%) White, 45 (17%) Asian, 30 (11%) Black, and 12 (5%) Multiracial. The median EDS score was 7 (range: 0–36); 58 (22%) fell into the High EDS score group. Resident race, fluency in a language other than English, and median household income were significantly associated with EDS scores. When controlling for other sociodemographic factors, Black residents were 4.2 (95% CI 1.62–11.01, p=0.003) times as likely to have High EDS scores than their White counterparts.

Conclusion:

Black surgical residents experience high levels of perceived discrimination on a daily basis. Institutional leaders should be aware of these findings as they seek to cultivate a diverse surgical training environment.

Introduction

In the United States, it is well-established that Black, Native American, and Latinx residents are underrepresented in medicine,⁵ and there is increasing evidence that residents of color are treated differently during their training. While older reports of differential experience included being called racial slurs,⁶ over time these differences have become more subtle and insidious.⁷ Black, Native American, and Latinx residents are less likely to feel that they fit into their training programs, feel less comfortable asking for help, and experience microaggressions in the work environment.^7,8 Given that residents of color are often in the minority, this unique experience may be overlooked.

Specifically, within surgery, discrimination has been documented among residents of color.^8–10 This has not been extensively studied, however, and it largely remains unaddressed. Most notably, discrimination has been associated with burnout and suicidal thoughts.¹⁰ Perceived discrimination, defined as stressful experiences of unfair treatment based on personal attributes,¹¹ have also been linked to adverse health outcomes.^1–4 Yet, perceived discrimination has been scarcely explored among healthcare providers.

Surgical residents are subject to demanding schedules and high acuity settings that foster intense work pressures. Within this culture, there are those who may experience further internal and external pressures due to discrimination, bias, and structural racism. In particular, race is a factor associated with attrition from surgery,¹² and those who leave the field cite a culture lacking safe spaces.¹³ It is crucial to understand the experiences of those who may be most vulnerable in order to cultivate a diverse and productive training environment. We sought to determine the existence and extent of everyday perceived discrimination (a measure of day-to-day discrimination both in and out of the hospital) among general surgery residents at academic medical centers and to identify risk factors for experiencing such discrimination.

Materials and Methods

PACTS Trial.

This study is part of a prospective, cluster-randomized crossover trial designed to evaluate the Provider Awareness and Cultural dexterity Toolkit for Surgeons (PACTS) curriculum. The trial is implemented across eight academic general surgery training programs each at different institutions from across the U.S., but largely concentrated in the Northeast. The PACTS curriculum is a novel curriculum that utilizes flipped classroom model, spaced education, and performance tracking and focuses on topics relevant to surgeons, including 1) caring for patients with limited English proficiency, 2) informed consent, 3) pain management, and 4) establishing trust to improve surgical residents’ knowledge, skills and attitudes regarding caring for a diverse patient population in an effort to improve surgical outcomes.^14,15 The study was approved by the Mass General Brigham Institutional Review Board (Protocol #2018P001237). Informed consent for enrollment in the trial and affiliated assessments was obtained from residents at the time that the baseline assessment was completed. Due to data-sharing restrictions, data from only seven of the eight sites were available for this analysis.

Resident Demographics.

Baseline demographic information was collected from residents enrolled in the study via an online self-assessment prior to undergoing any cultural dexterity training and prior to any randomization. Residents self-reported their age, institution, sex (male, female, nonbinary, or self-described), post-graduate year (PGY), fluency in any language other than English, race (American Indian/Alaska Native, Asian, Black and/or African American, African, Native Hawaiian/Pacific Islander, White/Caucasian, self-describe, or decline to answer), and ethnicity (Hispanic/Latinx, Non-Hispanic/Non-Latinx, or decline to answer). PGY-1 and -2 residents were grouped together as junior residents, and PGY-3, -4 and -5 residents were grouped together as senior residents. Due to the small sample size and to allow for further de-identification, race was categorized as follows: Black and/or African American and African were grouped together as Black; Asian and Native Hawaiian/Pacific Islander were grouped together as Asian; those who checked more than one box or who self-described as “Mixed” or “Biracial” were grouped together as Multiracial; and those who declined to report their race or who self-described as “Latino” or “Hispanic” (given a separate ethnicity category) were grouped together as Other.

Given that all surgical residents’ salaries are similar, we opted to capture childhood socioeconomic status (SES) using two different measures: 1) median income of childhood ZIP code and 2) parents’ highest academic degree. We felt that childhood SES would better represent lived experience since earning a medical degree and a resident salary results in similar current SES, despite different hurdles or approaches to reach this point. While SES is noted to be difficult to fully capture and inconsistently measured, education is a common proxy that generally predicts income, housing, and occupation.¹⁶ More specifically, childhood SES has been previously determined using either parent education or occupation.¹⁷ Thus, residents were asked to report the highest level of education for one or two parent(s)/legal guardian(s). Options included no schooling completed, nursery school, grades 1 through 11, 12th grade-no diploma, high school diploma, GED, less than one year of college, one or more years of college credit-no degree, associate’s degree, bachelor’s degree, master’s degree, professional degree, or doctorate degree. If residents provided education level of two parents, the highest level was utilized. These were grouped into two categories: college and higher or less than college. Some college without a bachelor’s degree was grouped with less than college. A second proxy of childhood SES was utilized by collecting the zip code that correlates to the area where residents grew up or spent the most time during childhood. Zip code level median household income (MHI) has been used as a proxy of SES.¹⁸ Census data were utilized to determine median household income of the zip code reported. The MHIs of the residents were categorized into the national MHI quartiles, with first quartile as the lowest and fourth as the highest.

Perceived Discrimination.

To quantify perceived discrimination, we utilized the nine-item Everyday Discrimination Scale (EDS).¹⁹ This scale captures daily experiences of discrimination, regardless of source (Figure 1). Response options included never/less than once a year, a few times a year, a few times in a month, at least once a week, and almost every day. The EDS was coded using frequency-based coding, in which each response in the scale is given an increasing value for increasing frequency (range: never/less than once a year = 0 to almost every day = 4).²⁰ Responses to questions were totaled, giving a range of 0 to 36; higher scores represented greater perceived discrimination. The EDS scores were divided into tertiles as done in prior work,²⁰ reflecting low (0–12), middle (13–24) or high scores (25–36). Given the distribution of the data and to provide sufficient numbers in each group, groups were further collapsed to a Low group (0–12) and a High group (13–36).

Figure 1.

Everyday Discrimination Scale

Statistical Analysis.

Data were analyzed for associations of EDS scores with race, ethnicity, sex, socioeconomic data, language ability, training level and program using Fisher exact tests (bivariate analysis) and logistic regression models (multivariate analysis). Statistical analysis was performed using STATA 16 (StataCorp LP, College Station, TX), and the level of significance was set as p < 0.05.

Results

Across seven academic general surgery residency programs, 266 residents completed the assessment (Table 1). Each training program contributed 10–21% of the total number of residents. There were 115 (43%) senior residents (32 PGY-5, 36 PGY-4, and 57 PGY-3), and 151 (57%) junior residents (58 PGY-2, 93 PGY-1). Racial breakdown was 157 (59%) White, 45 (17%) Asian, 30 (11%) Black, 12 (5%) Multiracial, and 22 (8%) Other. A total of 150 (56%) residents reported spending most of their childhood in zip codes with the fourth or highest MHI quartile. Regarding parent educational attainment, 238 (90%) residents had at least one parent with a college degree or higher and of these, 195 (73% of all residents) had at least one parent who had a master’s degree or higher.

Table 1.

Surgical Resident Characteristics and Everyday Discrimination Scale Scores

Variable	Overall n (%)	Low n (%)	High n (%)	P value
Age (years, median)	29	29	30	0.406
Resident Sex
Male	145 (55)	112 (77)	33 (23)	0.766
Female	121 (45)	96 (79)	25 (21)
Resident Language
English Only	174 (65)	144 (83)	30 (17)	0.019*
Fluency in Language not English	92 (35)	64 (70)	28 (30)
Resident Training Level
Senior	115 (43)	88 (77)	27 (23)	0.653
Junior	151 (57)	120 (79)	31 (21)
Resident Race
White	157 (59)	132 (84)	25 (16)	0.007*
Asian	45 (17)	33 (73)	12 (27)
Black	30 (11)	17 (57)	13 (43)
Multiracial	12 (5)	11 (92)	1 (8)
Other	22 (8)	15 (68)	7 (32)
Resident Ethnicity
Non-Hispanic/Latinx	233 (88)	185 (79)	48 (21)	0.135
Hispanic/Latinx	22 (8)	17 (77)	5 (23)
Unknown	11 (4)	6 (55)	5 (45)
Childhood Socioeconomic Status
Childhood Median Household Income
First Quartile	18 (7)	14 (78)	4 (22)	0.046*
Second Quartile	13 (5)	8 (62)	5 (38)
Third Quartile	35 (13)	31 (89)	4 (11)
Fourth Quartile	150 (56)	122 (81)	28 (19)
Unknown	50 (19)	33 (66)	17 (34)
Parent Education Level
College Degree and Beyond	238 (89)	187 (79)	51 (21)	0.634
No College Degree	28 (11)	21 (75)	7 (25)
Total	266 (100)	208 (78)	58 (22)

^*Denotes statistical significance (p < 0.05)

The median EDS score was 7, with scores ranging from 0–36. A total of 208 (78%) residents reported Low EDS scores and 58 (22%) reported High scores. Of note, all residents in the High group had EDS scores in the 75^th percentile or higher among our study subjects, including 33 (23%) men and 25 (21%) women. Of the residents who spoke two or more languages fluently, 28 (30%) were in the High group, compared to 30 (17%) who spoke English only. Regarding race, 25 (16%) White, 12 (27%) Asian, and 13 (43%) Black residents reported High levels of perceived discrimination. In terms of markers of SES, 4 (19%), 7 (9%), 4 (11%), 28 (19%) individuals from the first, second, third, and fourth MHI quartiles reported High EDS scores, respectively. Of the residents whose parents had completed at least a college degree, 51 (21%) had High EDS scores, compared to 7 (25%) of those residents whose parents did not. In our bivariate analysis, EDS score was associated with language (p=0.019), race (p=0.007), and childhood MHI (p=0.046). EDS score was not significantly associated with residency program (p=0.561) and this remained true when stratified by race (White: p=0.827, Asian: p=0.871, Black: p=0.074, Multiracial: p=0.658, and Other: p=0.904).

When controlling for other demographic and socioeconomic factors (age, sex, ethnicity, parent education level, childhood MHI, language ability), resident race remained significantly associated with EDS scores (Table 2). In this multivariate analysis, Black residents were 4.2 (95% CI 1.62–11.01, p=0.003) times as likely to have High EDS scores than their White counterparts. Asian residents were 1.4 times as likely to have High EDS scores than White residents, but this difference was not statistically significant (p=0.420).

Table 2.

Multivariate Model: Likelihood of High versus Low Everyday Discrimination Scale Scores

Variable	Odds Ratio	95% Confidence Interval	P value
Resident Age	1.03	0.95 – 1.12	0.443
Resident Sex
Female	1.05	0.55 – 2.03	0.879
Male	ref
Resident Language
Fluency in Language not English	1.89	0.89 – 4.00	0.879
English Only	ref
Resident Training Level
Junior	0.86	0.43 – 1.73	0.671
Senior	ref
Resident Race
Asian	1.43	0.60 – 3.44	0.420
Black	4.23	1.62 – 11.01	0.003*
Multiracial	0.49	0.05 – 4.17	0.511
Other	1.44	0.36 – 5.79	0.602
White	ref
Resident Ethnicity
Hispanic/Latinx	0.70	0.19 – 2.51	0.580
Unknown	3.13	0.57 – 17.20	0.188
Non-Hispanic/Latinx	ref
Childhood Socioeconomic Status
Resident Childhood Median Household Income
First Quartile	0.59	0.14 – 2.50	0.477
Second Quartile	2.77	0.76 – 10.09	0.121
Third Quartile	0.46	0.13 – 1.48	0.192
Unknown	1.18	0.47 – 2.96	0.720
Fourth Quartile	ref
Resident Parent Education Level
No College Degree	1.15	0.41 – 3.27	0.780
College and Beyond	ref

^*Denotes statistical significance (p < 0.05)

Discussion

In this study of surgical residents at seven academic medical centers, we found that race was significantly associated with High EDS scores. Specifically, Black residents were four times as likely to report High EDS scores compared to their White counterparts. Ultimately, Black surgeons are not immune to the bias experienced by people of color in the general population. The EDS captures chronic, daily discrimination that often manifests as microaggressions and also describes the level of discrimination experienced. Therefore, not only are residents of color more likely to experience discrimination based on prior studies, but our data also show that Black residents experience higher levels of discrimination on a daily basis. Our study also suggests that Asian residents experience higher levels of discrimination, although this trend did not reach statistical significance. Additionally, it is important to note that EDS does not simply measure discrimination during training, but rather reflects discrimination experienced by residents both in and out of the hospital.

Our findings are consistent with prior studies that have documented resident experiences of race-based discrimination within the bounds of the hospital.^6,7,9,21 Over the course of 30 years, these studies have captured how reports of racial discrimination evolved from blatant to subtle experiences. Early studies of race-based discrimination, including a national survey of residents from varied specialties published in 1994, documented that nearly 40% of Black, Latinx, and Native American residents had been called racial/ethnic slurs during their first year of residency.⁶ In a 1995 study, nearly 60% of Black and Asian emergency medicine residents reported experiencing racial discrimination.²² However, a decade later, a qualitative study documented that Black surgery and medicine residents had few experiences with outward racism. Instead, residents of color reported more subtle forms of discrimination such as social isolation from White colleagues and harsher punishments compared to their White colleagues.²³ Despite limited studies, this narrative has been echoed in more recent research and in 2018, a qualitative study of Black, Hispanic, and Native American residents across specialties and their experiences with race/ethnicity in the workplace and found that the occurrence of microaggressions was a major theme.⁷ Microaggressions may seem less harmful than outward racism (which also does still occur), but such experiences infiltrate daily interactions in medicine and surgery.²⁴ Though they occur at private and personal levels, microaggressions reinforce “in” groups and existing exclusionary power structures.²⁴

Among general surgery residents specifically, there are no qualitative studies and only three quantitative studies (cross-sectional national surveys) regarding the occurrence of discrimination.^8–10 In 2004, one study found that the experience of residents varied by race – compared to their White counterparts, Black and Asian residents were less likely to feel they fit in at their training programs and were more likely to report that attendings would think worse of them if they asked for help.⁸ While this national survey did not explicitly ask residents about discrimination, it was the first study of to examine the impact of race on the experiences of general surgery residents and the subtle difference in treatment by race suggested microaggressions at play. In 2019, a survey of surgical residents focused on mistreatment and burnout reported that 16.6% of residents had experienced racial discrimination.¹⁰ This was the first report of race-based discrimination among surgical residents. However, this percentage was reported for all survey respondents and resident race was not included as a variable. Most recently, in 2020, a study found that nearly 24% of all surgical residents experience discrimination, but when stratified by race, only 12% of White residents reported discrimination as compared to 71% of Black, 46% of Asian, and 25% of Latinx residents.⁹ Our study provides further depth regarding levels of discrimination experienced by residents by race. However, unlike these prior studies, our survey is of residents in academic surgery training programs and not nationally representative. The surgical community continues to lack a deep and nuanced understanding of these experiences.

An important consideration when contextualizing our findings is the underrepresentation of individuals from diverse backgrounds in medicine. Specifically, within surgery, only 2% of full professors and 6.2% of general surgery residents are Black, as compared to 12.4% of the U.S. population. Similarly, the U.S. population is 17.4% Latinx, but only 4% of full professors and 8.5% of general surgery residents identify as Latinx.²⁵ Nationally, Black academic surgeons are less likely to be promoted compared to their White counterparts.²⁶ More generally in medicine, it has been found that within a given institution, White faculty have significantly higher rates of promotion compared to Latinx and Black faculty,²⁷ and in national samples, non-White faculty are significantly less likely to be promoted despite controlling for sex, degree, tenure status, receipt of National Institutes of Health (NIH) research awards, department, and medical school type.²⁸ Even after controlling for an applicant’s educational background, country of origin, training, previous research awards, publication record, and employer characteristics, Black investigators across the U.S. are less likely to be awarded NIH funding.²⁹ There seems to be a lack of upward movement of Black individuals at multiple levels of career advancement. Augmenting diversity in the surgical workforce has received increasing attention,³⁰ especially since non-White self-identification remains an independent predictor of lower likelihood of being selected for residency interviews.³¹ Moreover, Black medical students report the predominantly White environment as a barrier to pursuing surgery, further limiting the applicant pool. Changing the culture of surgery, creating a more nurturing environment for trainees of color, and supporting promotion of Black trainees will not only improve the wellbeing of those in the system, but may also improve the pathway to becoming a surgeon.

Burnout is a problem in surgery across backgrounds given the demands of the profession. However, one cross-sectional national survey found that surgical residents who experienced discrimination, abuse, or harassment at least a few times per month were approximately three times more likely to develop burnout or suicidal thoughts that those who did not report any mistreatment.¹⁰ As residents have navigated the last several months amid a COVID-19 crisis and national Black Lives Matters protests, residents of color are left not only defending their communities, but also their patients and themselves. Especially in this setting, mistreatment may impact the wellbeing of residents and their desire to continue in this profession. Not only is racial and ethnic diversity limited, but the surgical community is at risk of losing residents of color to attrition. Particularly among female surgery residents, race is the number one factor that determines risk for attrition, with White women experiencing significantly lower rates of attrition compared to non-White women.¹²

In order to create substantial cultural change and provide a supportive environment for trainees, it will be important to understand the sources of discrimination. From prior work in the surgical and medical community, it is known that residents often attribute the major source of discrimination to patients and their families.^7,10,32,33 In one study, 98% of second and third year internal medicine residents reported witnessing biased patient behavior at least once a year and experiences with such behavior were more common among women, Black or Latinx, and Asian residents.³² Biased behavior in this study included belittling or demeaning stereotypes, role questioning, explicit epithets or refuse of care, and sexual harassment.³² Specific recurrent experiences of residents who are underrepresented in medicine with patients range from being mistaken for ancillary staff such as the environmental services workers to being told to go back to where they came from.⁷ In the above mentioned survey of surgical residents regarding burnout, the most frequently attributed source of racial discrimination was patients and patients’ families.¹⁰ Not only is it crucial to critically interrogate such experiences of discrimination by patients, but it is key to utilize this emerging data to develop protocols to assess and debrief such incidents in order to create organization cultural change.³⁴ Residency program leadership may not be able to prevent patient bias, however, protective policies can be established to support residents.³⁵ While the EDS measures discrimination both in and out of the hospital, patients blur this divide and developing a safe environment for residents will necessitate understanding and addressing discrimination from all sources.

Our work found that the majority of surgery residents grew up in communities in the highest MHI quartile. This reflects a known lack of socioeconomic diversity in medicine, with 73–79% of all medical school matriculants in the last three decades coming from the top two household income quintiles nationally.³⁶ In fact, in 2017, only 5% of all individuals entering medical school were from the lowest household income quintile in contrast to a quarter of matriculants coming from the top 5% of household incomes in the U.S.³⁶ While MHI was significantly correlated to EDS scores, this did not remain true in our multivariate analysis. Unlike the work done on discrimination and race in medicine, there remains a paucity of literature on childhood SES and discrimination among medical and surgical trainees. Entering the field of medicine remains a costly venture (exam prep, application fees, etc.), and our study confirms the lack of socioeconomic diversity in medicine.

Our study is not without limitations. Foremost, there are limitations inherent to the EDS instrument. Though this scale is among the most widely used to study discrimination,³⁷ the coding of this instrument has not been standardized²⁰ and so, there is a potential for misclassification of data in the process of converting to a categorical variable. Additionally, more recent studies have questioned the validity of the scale to make meaningful comparisons across social groups, especially other than race.³⁸ Despite multiple prior publications documenting the prevalence of gender-based discrimination in surgical training,^10,39–41 our study did not find a difference in EDS scores by sex. Although the EDS instrument is thought to capture discrimination based on any aspect of one’s identity, it was initially designed to capture race-based discrimination³⁸ and we caution readers to not simply conclude that our sample of residents do not experience different levels of discrimination based on sex.

Given IRB restrictions, data were only available for seven of eight sites. Regarding generalizability, the residencies surveyed are those involved in the PACTS Trial, do not have universal geographic representation, and are exclusively at training programs at academic medical centers. Since the residencies have opted to enroll in a randomized trial regarding cultural competency, it is possible that there is overall a greater commitment to diversity and inclusion at these particular programs, which would risk underrepresenting the true levels of perceived discrimination among surgical residents nationally. Additionally, not all residents at each site participated in the trial, further limiting generalizability. While Asian residents had increased odds of experiencing higher levels of discrimination, this was not significant and given prior studies documenting experiences of discrimination in the Asian resident community,^8,9,22 this likely reflects the study’s small sample size. For a more granular understanding of this heterogenous group of residents, future research should include a larger sample size and greater geographic diversity.

Within the race, ethnicity and MHI variables, we had missing data that we accounted for as an “other” category. We opted not to pursue imputation as the race and ethnicity missingness was less than 10% and with a small population, imputing an independent variable could bias our findings. Median MHI of communities is not an ideal proxy of childhood SES given variability of SES of families within a zip code. Moreover, residents were asked to recall the zip code that they spent the most time and this does not fully capture the experiences of those who moved while growing up. For example, it is possible that residents spent fewer, but more formative years in a given zip code that was not reported. Lastly, childhood zip code does not capture SES of those who immigrated to the U.S. Therefore, despite missing data, we felt that given that level of heterogeneity it would not be appropriate to impute MHI. Additionally, using only the highest education level of residents’ parents may not be fully representative of actual family structures and assumes that parent provided financially for the resident.

The EDS is a measure of daily experiences of discrimination, but our study does not capture the exact sources of discrimination and how much of the discrimination is in the work environment. We also did not collect data on faculty diversity and its impact of a more diverse faculty on resident perceived discrimination. Lastly, it is important to note that while race remains a social construct, racism is a mediator through which race has a very real impact.⁴² Ultimately, race as a variable is flawed. However, despite the complexities of the race variable, we attempt to capture the diverse experiences of residents in relation to their race.

Conclusions

In this study of seven surgical residencies, Black surgical residents are more likely to report higher levels of discrimination compared to White surgical residents. This study captures the level of discrimination among surgical residents both in and out of the hospital. It is essential to understand life experiences of surgical residents, to consider providers holistically, and study the impact of provider discrimination on surgical care delivery, as well as resident wellness (burnout/attrition). Institutional leaders should be aware of these findings as they seek to cultivate a diverse surgical training environment.

It will also be important to understand the impact of perceived discrimination on personal wellbeing, resident attitudes and knowledge about treating a diverse range of patients, and patient clinical outcomes. Specifically, it will be important to understand how experiences with perceived discrimination influence adoption and impact of the PACTS curriculum. Future work through the PACTS Trial will correlate resident EDS scores with cultural dexterity.

Highlights:

• 22% of surgical residents at academic centers report high levels of perceived discrimination

• Resident race, fluency in a language other than English, and median household income were significantly associated with EDS scores

• Black surgical residents were 4 times more likely to have high levels of perceived discrimination compared to their White counterparts