Assessment of the Prevalence of Medical Student Mistreatment by Sex, Race/Ethnicity, and Sexual Orientation

Key Points

Question  Does the self-reported prevalence of medical student mistreatment vary based on student sex, race/ethnicity, and sexual orientation?

Findings  In this cohort study of 27 504 graduating medical students, the following students reported a higher prevalence of mistreatment than male, white, and heterosexual students: female students; Asian, underrepresented minority, and multiracial students; and lesbian, gay, or bisexual students.

Meaning  These findings suggest that there is a differential burden of mistreatment that must be addressed to improve the medical school learning environment.

Importance  Previous studies have shown that medical student mistreatment is common. However, few data exist to date describing how the prevalence of medical student mistreatment varies by student sex, race/ethnicity, and sexual orientation.

Objective  To examine the association between mistreatment and medical student sex, race/ethnicity, and sexual orientation.

Design, Setting, and Participants  This cohort study analyzed data from the 2016 and 2017 Association of American Medical Colleges Graduation Questionnaire. The questionnaire annually surveys graduating students at all 140 accredited allopathic US medical schools. Participants were graduates from allopathic US medical schools in 2016 and 2017. Data were analyzed between April 1 and December 31, 2019.

Main Outcomes and Measures  Prevalence of self-reported medical student mistreatment by sex, race/ethnicity, and sexual orientation.

Results  A total of 27 504 unique student surveys were analyzed, representing 72.1% of graduating US medical students in 2016 and 2017. The sample included the following: 13 351 female respondents (48.5%), 16 521 white (60.1%), 5641 Asian (20.5%), 2433 underrepresented minority (URM) (8.8%), and 2376 multiracial respondents (8.6%); and 25 763 heterosexual (93.7%) and 1463 lesbian, gay, or bisexual (LGB) respondents (5.3%). At least 1 episode of mistreatment was reported by a greater proportion of female students compared with male students (40.9% vs 25.2%, P < .001); Asian, URM, and multiracial students compared with white students (31.9%, 38.0%, 32.9%, and 24.0%, respectively; P < .001); and LGB students compared with heterosexual students (43.5% vs 23.6%, P < .001). A higher percentage of female students compared with male students reported discrimination based on gender (28.2% vs 9.4%, P < .001); a greater proportion of Asian, URM, and multiracial students compared with white students reported discrimination based on race/ethnicity (15.7%, 23.3%, 11.8%, and 3.8%, respectively; P < .001), and LGB students reported a higher prevalence of discrimination based on sexual orientation than heterosexual students (23.1% vs 1.0%, P < .001). Moreover, higher proportions of female (17.8% vs 7.0%), URM, Asian, and multiracial (4.9% white, 10.7% Asian, 16.3% URM, and 11.3% multiracial), and LGB (16.4% vs 3.6%) students reported 2 or more types of mistreatment compared with their male, white, and heterosexual counterparts (P < .001).

Conclusions and Relevance  Female, URM, Asian, multiracial, and LGB students seem to bear a disproportionate burden of the mistreatment reported in medical schools. It appears that addressing the disparate mistreatment reported will be an important step to promote diversity, equity, and inclusion in medical education.

Medical student mistreatment remains a prevalent¹ and damaging² experience. Mistreatment encompasses a spectrum of abusive behaviors,³ including discrimination, assault, verbal abuse, and sexual harassment, and has been associated with burnout, depression, alcohol abuse, increased cynicism, and medical school attrition.^3-5 These detrimental outcomes are especially concerning for women, racial/ethnic minorities, and sexual minorities. Physicians who are female, racial/ethnic minorities, and sexual minorities make important contributions to medical practice that improve health care access and quality,^6-16 and mistreatment in these groups could have substantial consequences on physician workforce diversity.^8,17-21

Since 1991, the Association of American Medical Colleges (AAMC) has surveyed graduating medical students about mistreatment in its Graduation Questionnaire (GQ).²² Although the AAMC publishes GQ data yearly, these data have been presented only in aggregate, and little is known about how the prevalence of mistreatment varies by medical student demographic characteristics. It is possible that students with identities historically marginalized in medicine in terms of representation,^17,23,24 compensation,^25-27 career advancement,^28-40 and exposure to discrimination^{8,18,20,21,41-47} (women, racial/ethnic minorities, and sexual minorities) experience a higher burden of mistreatment than other groups.

Although some prior studies^48,49 have described mistreatment secondary to medical student sex, race/ethnicity, and sexual orientation, this work has tended to be limited by small sample size or low response rate or was conducted at only 1 or a few sites, limiting generalizability. To address this gap, we examined the prevalence and types of mistreatment reported by a large national cohort of medical students by student sex, race/ethnicity, and sexual orientation. Our study is unique in its large nationally representative sample, inclusion of a wide breadth of mistreatment types, and focus on the connection between membership in a marginalized group and the experience of mistreatment.

We conducted a retrospective cohort study of medical student responses to the AAMC-GQ. The AAMC-GQ is a survey that is administered annually to students graduating from all 140 accredited allopathic medical schools in the United States. The AAMC-GQ includes questions about student demographics, educational experience, finances, and career plans.^1,22 Our study was deemed exempt from review by the Yale Institutional Review Board because the data were deidentified.

The AAMC-GQ contains items that assess negative behaviors by medical school faculty, nurses, residents and interns, additional institutional employees or staff, and other students. The questions address students’ experience of general negative behaviors and 3 types of discrimination related to gender, race/ethnicity, and sexual orientation (eFigure in the Supplement).¹ These negative and discriminatory behaviors are seen as indicators of mistreatment and include experiencing public humiliation and being subjected to unwanted physical advances and bigoted remarks.²²

Data from student responses to the 2016 and 2017 AAMC-GQ were collected. Students from newer medical schools that did not participate in both the 2016 and 2017 AAMC-GQ were excluded from the data set provided to us by the AAMC. Descriptive statistics were computed for demographic variables, mistreatment type, and prevalence. For all descriptive statistics and analyses, respondents who did not answer all AAMC-GQ questions concerning general negative behaviors and discrimination were excluded.

The AAMC linked student AAMC-GQ responses to student self-reported sex and race/ethnicity via AAMC data applications and services. Students were excluded from descriptive statistics or analyses by race/ethnicity if the AAMC did not have data for their race/ethnicity or if race/ethnicity was reported only as “other.” Student sexual orientation was identified by self-report on the AAMC-GQ. Students who did not report sexual orientation were excluded from descriptive statistics and analyses by sexual orientation. Although the AAMC-GQ includes questions about gender identity, these data were not made available to the study team secondary to the small number of students self-reporting as transgender and concerns for privacy.

Demographic variables included sex (male vs female), race/ethnicity (white vs Asian vs underrepresented minority [URM] vs multiracial), and sexual orientation (heterosexual vs lesbian, gay, or bisexual [LGB]). Students were classified as URM if their self-reported race/ethnicity was American Indian, Alaska native, black, African American, Hispanic, Latino, Spanish, Native Hawaiian, or Pacific Islander. Students who reported multiple races/ethnicities were classified as multiracial.

The AAMC-GQ response choices for all questions regarding negative behaviors and discrimination use a 4-point scale (never, once, occasionally, and frequently). Based on the frequency distribution of reported mistreatment, “occasionally” and “frequently” were combined to create a 3-point scale consisting of never, once, and more than once. By combining “once” and “more than once,” we also created a dichotomous “ever/never” variable to describe if students reported ever having been mistreated.

Although the AAMC-GQ includes questions that examine a range of negative behaviors, the analysis is focused on the negative behaviors that we believed to be most detrimental to the student learning environment (Box). The prevalence and frequency of 8 types of mistreatment were assessed by student sex, race/ethnicity, and sexual orientation. These 8 types of mistreatment included 5 general negative behaviors and 3 types of discrimination associated with the specific student demographic category. For each student, we also evaluated the total number of the 8 types of mistreatment reported by student sex, race/ethnicity, and sexual orientation. Based on the frequency distribution of the number of self-reported mistreatment types, the following 4 categories were created: zero or no mistreatment, 1 type of mistreatment, 2 types of mistreatment, and 3 or more types of mistreatment.

Box Section Ref ID

We used Pearson χ² test or Fisher exact test, as appropriate, to compare differences in the prevalence and frequency of self-reported medical student mistreatment and the total number of mistreatment types by student sex, race/ethnicity, and sexual orientation. We applied a Bonferroni correction to account for multiple comparisons (corrected α = 0.00250). Because prior literature has suggested that much of the greater discrimination reported by URMs is directed at underrepresented women,⁵⁰ the interaction of URM status and sex on reports of any racial/ethnic discrimination were also tested using logistic regression. A 2-sided Wald test at P = .05 level of significance was used for logistic regression. For all other tests, P = .003 was used to indicate significance. Stata/SE, version 15.1 (StataCorp LP), was used to obtain predicted probabilities after logistic regression. All other analyses were performed using SPSS, version 26 (IBM).

The initial study cohort included 30 651 respondents. Of these respondents, 3147 students (10.3%) were excluded because they did not respond to all mistreatment questions. Of those excluded for failing to answer mistreatment questions, 2517 (80.0%) did not answer any of the mistreatment questions. Among students who provided demographic information, nonrespondents to the mistreatment questions were more likely than respondents to be male, Asian, URM, and multiracial. Therefore, our analysis included 27 504 unique student surveys, representing 72.1% of the 38 160 graduates from allopathic US medical schools in 2016 and 2017.

The final sample included the following: 13 351 female students (48.5%); 16 521 white students (60.1%), 5641 Asian students (20.5%), 2433 URM students (8.8%), and 2376 multiracial students (8.6%); and 1463 LGB students (5.3%). The demographic characteristics of the study cohort are summarized in Table 1.

Among respondents, 35.4% reported experiencing at least 1 type of mistreatment. The most commonly reported type of mistreatment was public humiliation (21.1%). Of all students, 18.5% reported experiencing discrimination secondary to gender, 8.8% secondary to race/ethnicity, and 2.3% secondary to sexual orientation. The eTable in the Supplement lists complete frequency data.

Female students reported a higher prevalence of mistreatment than male students across several domains (Table 2). Overall, a larger proportion of female students compared with male students reported at least 1 episode of mistreatment (40.9% vs 25.2%, P < .001). Female students reported a higher prevalence of public humiliation (22.9% vs 19.5%, P < .001) and unwanted sexual advances (6.8% vs 1.3%, P < .001). Moreover, female students reported higher rates of gender-based discrimination than male students (28.2% vs 9.4%, P < .001), including being denied opportunities for training or rewards based on gender (6.7% vs 4.7%, P < .001), being subjected to sexist remarks or names (24.3% vs 3.4%, P < .001), and receiving lower evaluations or grades solely because of gender (6.8% vs 4.6%, P < .001). In addition, a higher percentage of female students compared with male students reported 2 or more types of mistreatment (17.8% vs 7.0%, P < .001).

Compared with white students, Asian, URM, and multiracial students reported higher rates of mistreatment (24.0%, 31.9%, 38.0%, and 32.9%, respectively; P < .001) and discrimination based on race/ethnicity (3.8%, 15.7%, 23.3%, and 11.8%, respectively; P < .001) (Table 3). These reports of racial/ethnic discrimination included being denied opportunities for training or rewards based on race/ethnicity (1.5% [white students], 4.4% [Asian students], 7.3% [URM students], and 3.6% [multiracial students]; P < .001), being subjected to racially/ethnically offensive remarks or names (2.5%, 12.9%, 18.9%, and 9.6%, respectively; P < .001), and receiving lower evaluations or grades solely because of race/ethnicity (0.7%, 5.0%, 9.6%, and 3.4%, respectively; P < .001). Furthermore, Asian, URM, and multiracial students (4.9% white, 10.7% Asian, 16.3% URM, and 11.3% multiracial) reported a higher prevalence of experiencing 2 or more types of mistreatment compared with white students (P < .001). In addition, we found that the association of URM status with reports of racial/ethnic discrimination differed statistically significantly by sex (P = .01), with URM female students reporting the highest levels of racial/ethnic discrimination (26.5%) compared with URM male students (19.2%), non-URM women (7.8%) and non-URM men (6.8%).

For 3 of 5 general negative behaviors, LGB students reported a higher prevalence of mistreatment than heterosexual students (Table 4). Overall, 43.5% of LGB students reported an episode of mistreatment compared with 23.6% of heterosexual students (P < .001). A greater proportion of LGB students compared with heterosexual students said that they had been publicly humiliated (27.1% vs 20.7%, P < .001) and subjected to unwanted sexual advances (7.7% vs 3.7%, P < .001). In addition, LGB students reported higher rates of discrimination based on sexual orientation than heterosexual students (23.1% vs 1.0%, P < .001), including being denied opportunities for training or rewards based on sexual orientation (3.2% vs 0.3%, P < .001), being subjected to offensive remarks or names related to sexual orientation (21.8% vs 0.8%, P < .001), and receiving lower evaluations or grades solely because of sexual orientation rather than performance (4.0% vs 0.3%, P < .001). Furthermore, 16.4% of LGB students reported 2 or more types of mistreatment compared with 3.6% of heterosexual students (P < .001).

The major findings of our national study include not only a high prevalence of medical student mistreatment but also differences in the prevalence of mistreatment by student sex, race/ethnicity, and sexual orientation. Furthermore, results of the present study suggest an injurious interaction between URM status and sex, with URM female medical students reporting the highest prevalence of racial/ethnic discrimination. The differential treatment reported by medical students in this study suggests a noninclusive learning environment, which could have profound implications for the well-being and academic success of students. In fact, prior literature has shown an association between exposure to mistreatment and discrimination and student reports of decreased physical and mental health, worsening grades, and lower academic motivation and persistence.^51-59

Although the disproportionate burden of mistreatment reported by female, URM, multiracial, and LGB students is disquieting, these findings also demonstrate that several particularly harmful behaviors remain common in medical school. These reported behaviors include, but are not limited to, unwanted sexual advances (6.8% of female students and 7.7% of LGB students), lower evaluations secondary to bias and discrimination (6.8% of female students and 9.6% of URM students), and being subjected to sexist or bigoted comments (24.3% of female students, 18.9% of URM students, and 21.8% of LGB students). Moreover, 28.2% of female students, 23.3% of URM students, and 23.1% of LGB students reported an experience of discrimination.

The inequitable environment described by our study builds on prior research examining both academic medicine and community practice settings. These prior studies have demonstrated differences by race/ethnicity and sex in receipt of academic awards,^29,31,60 ratings on performance evaluations,^28,61,62 rates of promotion,^30,32 compensation,^25-27,30 National Institutes of Health funding,^35,37-39 and reports of discrimination.^18,41-45,48 The findings of the present study add to prior literature^48,49 on the topic by providing additional detail and nuance to the experience of mistreatment and discrimination encountered by a large national sample of medical students accounting for sex, race/ethnicity, and sexual orientation.

Some of the most commonly reported forms of mistreatment in this study were offensive remarks based on gender, race/ethnicity, and sexual orientation. Studies have shown that bigoted remarks can have negative consequences both on people targeted by the remarks^63-66 and on bystanders,^64,67-69 including reports of depression, fear, anger, and lower self-esteem^63-65,67 and lower job satisfaction, decreased productivity, and work-related depression.^64,67-69 Bigoted remarks have the potential to harm organizational morale and compromise institutional diversity efforts.^64,67-69

Equally concerning, bigoted comments that may be pervasive in medical school can have pernicious consequences in the patient setting. Bigoted statements, especially when spoken by medical faculty and supervising residents, represent a form of negative role modeling, which has been shown to be associated with racial/ethnic and anti–LGB and transgender bias in medical students.^70-77 This association is important because racial/ethnic minorities and sexual minorities experience substantial disparities in health and health care outcomes,^78-82 and physician bias has been identified as a key contributor to these disparities.^82,83

Another concerning negative experience reported by students was missed opportunities or lower grades because of discrimination. These experiences may have incremental consequences as trainees advance through their medical careers. Research has shown that small disadvantages can accumulate over time and prevent career advancement.^84,85 This accumulation could partially explain findings that women and racial/ethnic minorities are less likely than their counterparts to be promoted and receive academic awards and honors.^{23,32-34,40,51-59,86-89}

In addition, a key finding from the present study is that higher proportions of female, Asian, URM, multiracial, and LGB students experienced 2 or more types of mistreatment during their medical school career compared with their male, white, and heterosexual counterparts. Although mistreatment was prevalent across all demographic groups, this result may demonstrate that students with marginalized identities are more likely to be subjected to a deleterious environment.

This study has implications for medical schools, leaders in academic medicine, national medical organizations, and medical school accrediting bodies. Medical student mistreatment has been well documented over the last 30 years,^3-5,48,90-95 and its persistence remains a source of serious concern for students and educators. Despite efforts to curtail its occurrence, mistreatment remains prevalent.⁹⁶ Given the differential burden of mistreatment reported by women, racial/ethnic minorities, and sexual minorities, future interventions to reduce medical student mistreatment would benefit by incorporating strategies to address bias and discrimination in medical education.

Although there remains much work to be done to identify evidence-based practices to reduce bias and discrimination, potential interventions include implicit bias^97-100 and bystander intervention^101-103 training, better protections for individuals who have been subjected to and report instances of bias and discrimination,^5,96,104-109 and greater transparency in policies for reporting and remediating instances of bias and discrimination.^{5,92,96,106-110} As concerns about medical student wellness grow, the unequal prevalence of mistreatment described herein is noteworthy. Because prior literature has demonstrated an association between mistreatment and depression,⁴ burnout,³ and a desire to leave medical school,⁵ the disproportionate burden of mistreatment reported by female, URM, multiracial, and LGB students could hinder efforts to recruit and retain individuals from diverse backgrounds.

Several medical schools have recently developed an institutional role for a chief wellness officer.¹¹¹ Although medical schools continue to develop strategies to improve student wellness, greater collaboration between an institution’s chief diversity officer and chief wellness officer may be warranted. The chief diversity officer and chief wellness officer could work synergistically to develop novel interventions to address the association of mistreatment with the well-being of students from diverse backgrounds.

The reports of mistreatment and discrimination described in this study offer an opportunity for leaders in academic medicine and medical school governing bodies to reflect on the current climate of diversity, equity, and inclusion in medical education. Although recent data demonstrate that the diversity of medical school matriculants increased after the introduction in 2009 of the Liaison Committee on Medical Education diversity accreditation standards,¹¹² results from the present study indicate that much work remains to make medical school an inclusive and equitable environment. Attention to a medical school’s climate of equity and inclusion may represent an opportune focus for future medical school accreditation standards.

We believe our study findings also provide a basis for the AAMC, which has long advocated for diversity and inclusion,^112,113 to reconsider how it reports student mistreatment. Although the AAMC releases GQ mistreatment data annually, these data are presented in aggregate.^1,114 Given the disparate burden of mistreatment found in our study, a more granular breakdown of student mistreatment by sex, race/ethnicity, and sexual orientation may provide a more transparent reflection of the climate of diversity, equity, and inclusion present in medical education. Medical schools could use these data to more accurately monitor progress in reducing student mistreatment. In addition, these data could help identify and disseminate best practices among medical schools that have excelled in creating inclusive and equitable learning environments.

The results of the present study may also have implications for the Accreditation Council for Graduate Medical Education (ACGME), which released its first diversity accreditation standard in the 2019-2020 academic year.¹¹⁵ There are currently more than 124 000 resident physicians,¹¹⁶ and study findings have suggested that residents experience mistreatment and discrimination.^{21,43,45,48,117-120} As the ACGME establishes guidelines to assess adherence to its diversity accreditation standards, attention to how physician trainees from all backgrounds are treated during residency training will be critical. Because medical students and residents work in the same clinical environment, it may also be important to align questions on the AAMC-GQ and the ACGME wellness survey that trainees complete each year¹²¹ to better understand the similarities and differences in how mistreatment influences individuals at different levels of training.

In addition, the ACGME could consider adding questions to their annual program updates that ask participating residency programs how they monitor trainee mistreatment and what mechanisms are available to trainees who experience mistreatment to seek assistance. These narratives, in conjunction with mistreatment data, could possibly identify evidence-based practices to reduce trainee mistreatment in graduate medical education.

Findings from this study suggest that much work remains to address medical student mistreatment. To better tailor interventions to students’ experiences of mistreatment, future studies should further explore the sources of mistreatment and how the mistreatment source affects student well-being. As medical schools strive to create more inclusive learning environments, investigation into the experience of mistreatment by additional student demographic characteristics, such as age, religious beliefs, socioeconomic status, or disability, may be indicated. In addition, people do not experience aspects of their identities in isolation.^50,122 Our analysis examining the interaction of URM status and sex suggests that a student’s unique combination of multiple identities may contribute most substantially to mistreatment. Future studies should explore the effect of the intersectionality of student sex, race/ethnicity, and sexual orientation on the prevalence of mistreatment.

This study has several limitations. Graduating medical students may be less likely to remember mistreatment occurring early in their medical school career.⁹⁶ Because of the association between mistreatment and burnout, it is possible that some students who experienced mistreatment left medical school before graduation or were overrepresented among the students who graduated during the study period but failed to complete the GQ. Consequently, this study may underreport medical student mistreatment.

In addition, male, Asian, URM, and multiracial students were overrepresented among nonrespondents to mistreatment questions, and it is possible that this study does not capture the full experience of mistreatment in these groups. Nevertheless, to our knowledge, this study represents the largest investigation of medical student mistreatment to date accounting for medical student sex, race/ethnicity, and sexual orientation. Another limitation of the study is that individuals from diverse backgrounds were combined to create the URM, multiracial, and LGB categories, which may obscure differences in the mistreatment experience of any single group.

All data on mistreatment were obtained by medical student self-report, and it is possible that students may classify their experiences differently than medical school faculty, staff, residents, or their peers. Nevertheless, prior literature has shown that medical students perceive mistreatment events in a similar manner to other members of the academic medicine community, including residents and attending physicians.¹²³

Our findings indicate that medical student mistreatment remains common. Women, racial/ethnic minorities, and sexual minorities appear to experience a disproportionate burden of the mistreatment reported in medical schools. This differential burden of mistreatment may have substantial implications for the medical school learning environment and the diversity of the physician workforce.

Accepted for Publication: December 24, 2019.

Corresponding Author: Katherine A. Hill, BA, BS, 2022 MD Candidate, Yale School of Medicine, 333 Cedar St, Sterling Hall of Medicine, Room IE-61, New Haven, CT 06520 (katie.hill@yale.edu).

Published Online: February 24, 2020. doi:10.1001/jamainternmed.2020.0030

Author Contributions: Ms Hill and Dr Boatright had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hill, Samuels, Gross, Sitkin Zelin, Wong, Boatright.

Acquisition, analysis, or interpretation of data: Hill, Samuels, Desai, Sitkin Zelin, Latimore, Huot, Cramer, Wong, Boatright.

Drafting of the manuscript: Hill, Boatright.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hill, Desai, Cramer, Boatright.

Obtained funding: Hill, Latimore, Wong, Boatright.

Administrative, technical, or material support: Hill, Samuels, Sitkin Zelin, Huot, Boatright.

Supervision: Gross, Boatright.

Conflict of Interest Disclosures: Dr Samuels reported receiving grants from the Society of Academic Emergency Medicine (SAEM) Foundation/Academy for Diversity & Inclusion in Emergency Medicine (ADIEM) and from the Association of American Medical Colleges (AAMC) Northeast Group on Educational Affairs (NGEA). Dr Gross reported receiving grants from the National Comprehensive Cancer Network/Pfizer and from Johnson & Johnson and receiving support for travel to and speaking at a scientific conference from Flatiron Inc. Dr Wong reported receiving grants from the SAEM Foundation/ADIEM and from the National Institutes of Health (NIH) National Center for Advancing Translational Sciences (grant KL2TR001862). Dr Boatright reported receiving grants from the AAMC NEGA. No other disclosures were reported.

Funding/Support: This study was supported by the SAEM Foundation/ADIEM, the AAMC NEGA, and the NIH National Institute of Diabetes and Digestive and Kidney Diseases (award T35DK10468 [Ms Hill]).

Role of the Funder/Sponsor: The funding sources had no substantive involvement in any aspect of the study, including the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation of the manuscript.

Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Additional Contributions: Additional critical review of the manuscript was performed by David Matthew, PhD, and Marie Caulfied, PhD, at the AAMC, Washington, DC. They were not compensated for their contributions.