Race, Ethnicity, and Medical Student Well-being in the United States FREE

A racially/ethnically diverse physician population is important to optimally address health care disparities among patients.^1- 3 Minority physicians can improve access to care for patients of similar racial/ethnic backgrounds,^4- 5 may be more likely to conduct research pertinent to concerns of minority patients,¹ and are pivotal to teaching students about cultural competence.¹ To achieve such diversity, medical schools have implemented prematriculation programs^6- 7 as well as ongoing interventions for minority students to help them succeed.^8- 9 Despite these efforts, minority students are more likely to struggle academically,^10- 12 which threatens the success of “pipeline” efforts aimed at promoting diversity in the physician workforce.^13- 14

While academic difficulty among minority students during medical school has been both recognized and investigated, the ways in which the training experiences of minority students compare with those of their colleagues remain relatively unexamined. An increasing body of literature suggests that, independent of race/ethnicity, medical students in the United States¹⁵ and abroad^16- 17 experience significant distress (eg, depression, anxiety, and burnout). This distress has both personal and professional consequences and parallels the growth of cynicism that is observed during medical training.^18- 19 Although we and others have reported that stressful personal life events²⁰ and curricular factors¹⁵ contribute to medical student distress, little is known about what role minority status plays in this process.

Five studies, enrolling a total of 228 minority students, have evaluated the relationship between race/ethnicity and distress among US medical students.^21- 25 Three reported no differences in students' mental health by race/ethnicity,^22- 24 while 1 found that Hispanic students were 3.4 times more likely to be depressed than non-Hispanic students.²⁵ Each of these studies was small and conducted at a single medical center. In our 2004 survey of medical students attending 1 of 3 schools in Minnesota, no differences were found in overall burnout or symptoms of depression by minority status, although minority students did report lower mental quality of life (QOL) scores (P = .05).²¹ The study, however, was limited by a lower response rate among minority students, and the generalizability of the findings beyond a single Midwestern state is unclear.

Among the general population, differences have been noted in the prevalence of burnout²⁶ and QOL by race/ethnicity, with minorities being more resilient,²⁷ whereas no differences have been found in the prevalence of depression.²⁸ Based on these observations and our previous findings, we hypothesized that burnout and QOL among medical students would vary by race/ethnicity, while symptoms of depression would not. To better understand the prevalence of distress among minority and nonminority students and to gain insight into the effect of race/ethnicity on the training experience, we conducted a multicenter study of more than 3000 medical students attending 5 medical schools in the United States.

The institutional review board at each institution approved this study. All medical students (N = 3080) at Mayo Medical School (Rochester, Minnesota), University of Washington School of Medicine (Seattle), University of Chicago Pritzker School of Medicine (Chicago, Illinois), University of Minnesota Medical School (Minneapolis), and University of Alabama at Birmingham School of Medicine were invited to participate. Students were sent an e-mail message with a cover letter that linked to the Web-based survey. Participation was elective and responses were anonymous.

In the survey, students were asked to classify themselves as Caucasian, African American, Hispanic, Asian, Native American, Pacific Islander, or other non-Caucasian and to complete the Maslach Burnout Inventory (MBI),²⁹ the Primary Care Evaluation of Mental Disorders (PRIME MD),³⁰ and the Medical Outcomes Study Short Form (SF-8)³¹ to identify burnout, symptoms of depression, and QOL, respectively. According to convention, we considered a score of 27 or higher on the emotional exhaustion (EE) subscale score and/or 10 or higher on the depersonalization (DP) subscale an indicator of professional burnout.²⁹

To further explore the impact of race/ethnicity on students' experiences, students were asked, “Has your race adversely affected your medical school experience?” Students who responded affirmatively were asked to describe how their race had adversely affected their medical school experience in a free entry text box and whether they had discussed the matter with their Office of Minority Affairs (OMA). Those who indicated they had contacted their OMA were asked to comment about the responsiveness and helpfulness of the OMA, while those who did not access their OMA were asked to comment about why they had not done so.

The primary analyses involved comparing the prevalence of burnout, a positive depression screen result, and mental and physical QOL by minority status. Differences by race were evaluated using χ² tests or Kruskal-Wallis tests. All tests were 2-sided, with type I error rates of 0.05. A forward stepwise logistic regression was used to evaluate independent associations among minority status, age (≥25 years), sex, marital status, and parental status, with indicators of distress where differences were noted by race/ethnicity.

Content analysis of the students' comments was conducted using a constant comparative approach.³² Three of us (L.N.D., M.R.T., and T.D.S.) independently identified themes and categories from half of the students' comments, using sentences as the unit of analysis. Each author's codes were compared and differences were negotiated. The revised codes were independently applied to the remaining student comments. Differences in code application were discussed and the codes were revised, where needed. Finally, 1 of us (L.N.D.) applied the updated code to all students' comments, and 2 of us (M.R.T. and T.D.S.) reviewed and confirmed the coding. Paradigmatic quotes were selected to illustrate major themes and to provide rich description.

Correct e-mail addresses were confirmed for 99.9% (3076 of 3080) of the students. In total, 1701 students (response rate, 55%) responded to the survey. Nonresponders were more likely to be male and aged 24 years or older (both P ≤ .001). Demographic characteristics of responders are shown in Table 1. Twelve students did not respond to the race/ethnicity question, yielding a total sample of 1689 for analysis related to race/ethnicity. Of these individuals, 410 students (24%) classified themselves as a minority (61 African American, 50 Hispanic, 186 Asian, 23 Native American, 8 Pacific Islander, and 82 other non-Caucasian). There was no difference in response rate by minority status. Among responders, minority and nonminority students were similar with respect to sex, age, and year in school. Minority students were less likely to be married (P < .001) and to have children (P = .05).

As shown in Table 2, 47% (754 of 1620) of students met the criteria for burnout, with 37% (600 of 1603) having high EE and 28% (435 of 1555) having high DP or a low sense of personal accomplishment (PA). As there were no statistically significant differences in mean scores for EE, DP, or PA among different minority groups (EE, P = .44; DP, P = .55; and PA, P = .66 [data not shown]), minority student data were pooled for additional analysis. When the individual domains of burnout were evaluated, nonminority students had a higher DP score (mean ± SD, 7.0 ± 5.35 vs 6.2 ± 5.11; P = .01) than minority students but a similar EE score (mean ± SD, 23.5 ± 9.94 vs 22.6 ± 9.94; P = .10). There was no difference in PA scores by minority status (mean ± SD, 37.1 ± 7.21 vs 36.2 ± 7.80; P = .14). Overall, using the published thresholds to classify burnout,²⁹ nonminority students were more likely to be burned out (P = .03) and have higher EE (P = .03) and DP (P = .01) scores. The differences in burnout by race/ethnicity persisted on multivariate analysis after age (≥25 years), sex, and parenting and marital status were controlled for (all P < .01).

Nearly half of all students (820 of 1691) screened positive for symptoms of depression (Table 2). There were no statistically significant differences in screening positive for symptoms of depression among minority groups (P = .49) or between minority students as a group and nonminority students (P = .12).

The scores of the students' mental and physical QOL are shown in Table 2. When age-matched individuals and the general US population were compared, medical students had lower mental QOL scores but higher physical QOL scores than both of these reference groups (all P ≤ .001). There were no statistically significant differences in mental or physical QOL scores between different minority groups (P = .41 and P = .85 for mental and physical QOL scores, respectively). When minority students were collectively compared with nonminority students, the minority students had lower physical QOL scores (P = .02) but similar mental QOL scores. Although small, this difference persisted on multivariate analysis after age (≥25 years), sex, and parenting and marital status were controlled for (parameter estimate, 1.06; P = .02).

As normative scores on the SF-8 are available for Caucasian, African American, Asian/Pacific Islander, Native American, and Spanish/Hispanic Americans, comparisons were made between students' SF-8 scores in these racial/ethnic groups and the corresponding normative racial/ethnic group in the general US population. Mental QOL scores were consistently lower among medical students than among normative individuals of the same racial/ethnic group in the general US population (all P < .01), while physical QOL scores did not differ substantially except among Caucasians and Native American medical students, who had higher physical QOL scores than their respective normative samples (P < .001 and P = .02, respectively).

Minority students (46 of 406) were more likely than nonminority students (28 of 1278) to report that their race had adversely affected their medical school experience (P < .001). Among those students who indicated that their race had negatively affected their experience, 70% (52 of 74) provided a text answer describing how it had done so. Four major themes were identified by analysis of these comments: (1) racial discrimination (eg, unfair treatment based on race); (2) racial prejudice (eg, an unfavorable preformed opinion based on race); (3) feelings of isolation; and (4) interpersonal and communication differences owing to cultural upbringing (eg, consequences of differences in socialization such as cultural norms, expectations, and upbringing). Both minority and nonminority students made comments that were categorized within the first 3 themes. With respect to racial discrimination, minority students wrote about feeling harassed, experiencing bigotry, and receiving inequitable performance evaluations, while nonminority students expressed the view that their needs were considered inconsequential by the medical school, particularly with respect to personal needs and career opportunities. Comments about experiences of racial prejudice included being considered less intelligent or less qualified to be a medical student owing to race among minority students and being unfairly stereotyped (eg, assumptions about economic status based on race/ethnicity) among both minority and nonminority students. Both minority and nonminority students wrote about difficulty connecting with their peers and, as a result, feeling isolated. In addition to such comments about isolation from peers, some minority students reported feeling isolated from faculty and their family. Unique to minority students were comments about the impact of their cultural upbringing on interactions with faculty. Differences in cultural norms also affected relationships with patients, as did English as a second language. Representative comments for each theme are shown in Table 3.

Of the 74 students (46 minority and 28 nonminority) who indicated that their race had negatively affected their experience, only 12 (16%) had discussed their concerns with someone at the OMA. More minority students (10 of 46) than nonminority students (2 of 28) had accessed the OMA. Of those who had gone to the OMA, 67% (8 of 12) of students felt supported and thought that the experience had been positive. Few students commented that the OMA was unhelpful. Examples of comments are shown in Table 4.

Of the remaining 62 students who reported that their race adversely influenced their medical school experience, 69% (25 of 36) of minority students and 65% (17 of 26) of nonminority students explained why they had not discussed the matter with their OMA. Common barriers included (1) a perception that the OMA was unable to effectively address racial issues, (2) an unawareness of the resource (although all schools participating in the study had an OMA at the time of the survey), (3) concerns about adverse personal consequences of reporting about racial issues, (4) problems with accessing the OMA, and (5) a belief among nonminority students that the office was not for them. Among minority students, the predominant reason cited was the belief that racial issues are simply too pervasive as a societal issue for the OMA to effectively address, whereas the predominant reason among nonminority students was their perception that the office was a resource for minority students only. Other access barriers were also cited by minority students, including reports that office hours were not conducive to their schedule and that the OMA staff members were not available at times when the students could meet. A final barrier mentioned by both minority and nonminority students was a concern about adverse personal consequences. Typical comments can be found in Table 4.

Minority students who perceived that their race had adversely affected their medical school experience (n = 46) were more likely than minority students who reported no such experience (n = 360) to have high EE (58% [26 of 45] vs 29% [98 of 335]; P = .001) and high DP (39% [17 of 44] vs 20% [62 of 316]; P = .01) scores and to meet criteria for burnout (64% [29 of 45] vs 38% [130 of 338]; P = .001). These minority students were also more likely to screen positive for depression (72% [33 of 46] vs 49% [176 of 359]; P = .004) and to have lower mental QOL scores (mean ± SD, 36.5 ± 12.71 vs 44.2 ± 9.79; P = .001). In contrast, no difference in burnout, depression, or QOL was identified among nonminority students who reported that race had adversely affected their experience (n = 29) relative to their peers without such complaints (n = 1249).

Attention to the well-being of learners is necessary to provide effective medical education.³⁷ Lack of well-being and the presence of psychological distress may have adverse effects on both professional development^38- 40 and personal life.^41- 43 Consistent with other studies, we found a high prevalence of symptoms of depression and lower mental QOL scores among medical students than among national samples of age-comparable individuals.^15,17,20 Furthermore, we confirm our earlier finding that burnout is common (47%) among medical students,²⁰ providing further support to the hypothesis that the origins of physician burnout begin during the earliest phase of physician training.^44- 45 While we and others have previously explored the role of personal life events, curricular factors, family or personal history of depression, demographic variables (eg, sex, marriage, and parenting), personality, and stress on medical student distress,^15,20 to our knowledge this is the largest study to explore the relationship between race/ethnicity and distress.

In both the present study and our previous study of medical students from a single Midwestern state,²¹ we found no difference in the prevalence of symptoms of depression by race/ethnicity. Consistent with Maslach’s²⁶ historic studies of human service professionals, we also found burnout to be less common among minority students, a finding that persisted on multivariate analysis. Why nonminority students have higher degrees of burnout is unknown. Maslach speculated that the life experiences of minorities have made them more resilient to overcoming obstacles. This may be particularly true of minority medical students, who may have overcome substantial educational and/or professional challenges as well as cultural/racial challenges before matriculation. While this hypothesis provides a possible explanation, its validity is unknown, and further research is needed to provide insight into this phenomenon.

Although the overall frequency of burnout among minority students was lower than among nonminority students, minority students continue to face unique challenges. Our study suggests that race is profoundly related to distress for individual minority students who experience discrimination, prejudice, or isolation related to the color of their skin. Minority students were almost 5 times more likely than nonminority students to report that race had adversely affected their training experience. Those who felt that their race had adversely affected their medical school experience were twice as likely to have high EE and DP scores as minority students who reported no such experience. These students were also more likely to be burned out, screen positive for depression, and have lower mental QOL scores. While it is initially tempting to speculate that these findings are attributable to biased reporting among burned out individuals, several facts argue against this assumption. First, minority students who reported that race adversely affected their medical school experience had worse emotional distress (ie, burnout, symptoms of depression, and lower mental QOL scores) than, but similar physical QOL scores to those of, their minority peers without such experience. Second, the prevalence of burnout, the symptoms of depression, and the mental and physical QOL scores among nonminority students who reported that race had adversely affected their experience were similar to those among the remaining nonminority students without such complaint. These observations argue against generalized “complaining” by burned out medical students as the cause of these findings and suggest a specific relationship between adverse experiences related to race and minority student distress.

First, medical schools need to be aware of the prevalence of student distress among the entire student body and to support efforts, including further research studies, to identify the causes and consequences and to provide solutions to this problem. Second, medical schools need to recognize the impact of race/ethnicity on students' experiences as well as the higher prevalence of distress among students who have adverse experiences because of race/ethnicity. Third, medical schools need to carefully review the performance evaluations of minority medical students for evidence of covert discrimination (as reported not only in our study but also on Association of American Medical Colleges graduation surveys).⁴⁶ Fourth, medical schools need to address the barriers to students' discussing issues regarding race/ethnicity with faculty or staff members who are empowered to handle such complaints (eg, OMA). Efforts are needed to mitigate students' fear of reprisal and their perception that issues pertaining to race cannot be dealt with in a constructive manner within the institution. Fifth, medical schools need to prepare all of their graduates to be resilient through the course of their career. The impact of race continues well beyond medical school, into practice, with ongoing discrimination, prejudice, and isolation.⁴⁷ Given the commonality of experience reported by students in the current and other series,^33,47- 48 institutions need to more effectively respond to and address racial issues. Sixth, we suggest that medical schools need to (1) promote collaboration among students of various racial/ethnic backgrounds, (2) provide a venue for students to openly discuss racial issues, (3) avidly promote nontolerance of racism among faculty, and (4) develop effective support systems for students who feel that race has detrimentally affected their experience. Additional approaches have also been suggested.³⁴

Our study is limited by several factors. First, its cross-sectional design precludes determination if minority status is causally related to student distress and QOL. Second, although our response rate is typical of physician³⁵ and medical student¹⁵ surveys, response bias is also a possibility. How symptoms of depression, QOL, and burnout affect the response rate is unknown. While some investigators may speculate that distressed students might lack the motivation to fill out a survey, others may hypothesize that distressed students will be more interested in the topic and be more likely to participate. We do not know whether nonresponders are more or less likely to have reported distress. Third, our sample was biased toward female students and younger students, who were more likely to respond to the survey. Whether the distress level of nonresponders differs by sex or age is also unknown.

Our study also has several important strengths. First, to our knowledge, it is the largest multicenter study conducted on student distress¹⁵; we enrolled nearly twice as many minority medical students as all previous studies combined. All but 2 other studies^36,49 on medical student distress have been conducted at a single center. Second, the students in our study were from diverse private and public medical schools spread across the United States and are representative of US medical students with respect to sex, relationship status, and parenting status,^50- 51 lending generalizability to the results. Third, there was no variation in response rate by race/ethnicity, which was the primary focus of this analysis. Fourth, our survey used the MBI, PRIME MD, and SF-8, all of which are standardized and validated instruments that allow comparison with the general population and other samples of medical students, residents, and physicians. Fifth, the prevalence of positive depression screen results and burnout among students in this survey is similar to that reported in other studies of medical students,¹⁵ suggesting that the distress we observed is typical of students in the United States. Sixth, the qualitative analysis was performed with rigor, and the verbatim quotations convey face validity to the themes.³² In light of the anonymous nature of the survey, free-text responses to open-ended questions about the impact of race/ethnicity on students' experiences were unlikely affected by response bias or social desirability.

In conclusion, developing a racially/ethnically diverse physician population is necessary to meet the health care needs of the nation's multicultural population and to address health care disparities between racial groups. Adverse experiences related to race appear to relate strongly to burnout among minority students and may be related to the increased attrition rates of minority medical students. Given the frequency of distress among medical students of all races, much work remains to ensure that medical students find a safe, supportive, and mentoring environment in which to train. Well-designed, multicenter studies exploring the role of distress and QOL on student learning and competency, attrition from medical school, and the diversity of tomorrow's physician workforce would benefit our schools, our students, and, ultimately, our patients.

Correspondence: Liselotte N. Dyrbye, MD, Department of Internal Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55906 (dyrbye.liselotte@mayo.edu).

Accepted for Publication: May 16, 2007.

Author Contributions: The primary author had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Dyrbye, Thomas, and Shanafelt. Acquisition of data: Dyrbye, Thomas, Eacker, Harper, Massie, and Power. Analysis and interpretation of data: Dyrbye, Thomas, Shanafelt, Huschka, Novotny, and Sloan. Drafting of the manuscript: Dyrbye, Thomas, and Shanafelt. Critical revision of the manuscript for important intellectual content: Dyrbye, Thomas, Eacker, Harper, Massie, Power, Huschka, Sloan, and Shanafelt. Statistical analysis: Huschka, Novotny, and Sloan. Obtained funding: Dyrbye, Thomas, and Shanafelt. Study supervision: Dyrbye, Sloan, and Shanafelt.

Financial Disclosure: None reported.

Funding/Support: This study was funded through an intramural grant from the Mayo Clinic College of Medicine.

Role of the Sponsor: The sponsor had no role in the design or conduct of the study; the collection, management, analysis, or interpretation of the data; or the preparation, review, or approval of the manuscript.

Additional Contributions: Janice L. Hanson, PhD, Uniformed Services University of the Health Sciences, and Ilene Harris, PhD, University of Illinois–Chicago College of Medicine, reviewed the manuscript.