Medical students’ mental health is a topic of great interest, given that several multi-institutional studies have revealed that they have a substantially lower mental quality of life than similar-age individuals in the general population (1–3). These are issues that have implications for medical student professionalism (4), dropout rates (5), and the health of future physicians (6). Medical school pressures, including overwhelming amounts of material to learn in short periods of time, unclear expectations, and financial stressors (7), contribute to excessive work demands. Burnout, stress, and depression are three possible adverse consequences of these stressors.
Burnout is a pathological syndrome with three dimensions: emotional exhaustion, depersonalization, and loss of sense of personal achievement (8). Burnout and depression are also associated with suicidal behaviors among medical students (9). Only in the past 6 years has the evaluation of burnout among medical students become of interest to the medical community (3). We found nine recently published studies evaluating burnout (1, 2, 4, 5, 9–13), only five of which measured stress, burnout, and depression simultaneously among United States medical students (1, 4, 5, 9, 13). These five studies were limited to the same seven medical schools. Moreover, only one study of burnout investigated resources or coping mechanisms used by students (10).
In our medical school, two medical student suicides within an 11-month span heightened concerns about the emotional well-being of our students. Also, our institution is relatively unique in that it teaches the basic sciences within 1.5 years, rather than 2 years, and this practice may result in different patterns of psychological morbidity throughout the medical student career. Our objective in this study, therefore, was to investigate the prevalences of stress, depression, and burnout at this school in the first-, second-, and third-year classes, as well as the resources used by students to alleviate psychological distress.
The online survey included three scales designed to assess burnout, depression, and stress; one open-ended question on institutional programs that had helped them cope; and questions on demographic background.
We utilized the Maslach Burnout Inventory, a 22-item instrument that is the gold standard, and most frequently used measure of burnout in the medical literature (9), to assess burnout. Each item asks participants to indicate how often they have certain work-related feelings. The questions were adapted with permission from the publisher to address attitudes toward medical school (rather than work). Three subscales evaluate three domains of burnout: emotional exhaustion (a sense of overwhelming work demands that deplete emotional resources and the ability to help people), depersonalization (negative, cynical feelings toward patients or colleagues), and lack of personal achievement (a sense of inefficacy and low personal accomplishment) (10). It has been validated among human-service workers (14, 15) and used extensively in studies of medical students (1, 2, 4, 5, 9–13). Although the original burnout scale utilized a 7-point Likert scale (0–6; “never” to “every day”) for response options to each item, we inadvertently excluded the “never” option in the response options for each item. This converted the scaling for each item to a 6-point scale (“a few times a year or less” to “every day”). Reliability of the scale among our sample was 0.90, with reliabilities for each subscale at 0.90, 0.76, and 0.78 (for emotional exhaustion, depersonalization, and personal achievement, respectively).
We used the two-item Primary Care Evaluation of Mental Disorders (16), a screening tool that performs as well as longer instruments (17) for assessing potential depression. Respondents screen positive for possible depression if they answer positively to either “During the past month, have you often been bothered by feeling down, depressed, or hopeless?” or “During the past month, have you often been bothered by little interest or pleasure in doing things?” This instrument has a sensitivity of 86%–96% and a specificity of 57%–75% for major depressive disorder (16, 17).
We selected the Perceived Medical School Stress Scale (7) to assess stress. Each of the 13 items assessed students’ agreement (on a 5-point Likert scale, from Strongly Disagree to Strongly Agree) with a statement about medical school stress. Previous research has provided validation information and documented the reliability of the scale at 0.83 (7). Reliability in our data was 0.87. Finally, we asked one open-ended question: “What (medical school) program(s) have helped you cope with stress, anxiety, or burnout? Please explain.”
Approval for the study was obtained from the school’s Institutional Review Board. All 526 first- (MS1s), second- (MS2s), and third-year (MS3s) students were e-mailed three times to ask for their participation in the online survey. Fourth-year students were invited to participate as well, but results from this cohort were not included in this study because of low participation rates. The e-mail included a description of the compensation, which was a $5 “reward” to be added to class funds for each student who participated. The survey was open from March 14 until April 2, 2010. Participation was elective, and responses were anonymous. Information about help available for concerns relating to the survey was provided in the e-mail.
For data from the three scales, both descriptive (mean, median, standard deviation [SD], percent), and inferential statistics were used. To determine differences among classes for burnout and stress, we used analysis-of-variance tests with eta-squared (η2) for effect-size estimation. To determine differences in classes for depression, we used chi-square tests (χ2, with Cramer’s V [φ] for effect-size estimation. We used SPSS Version 18.0 to conduct our analyses, with alpha (α) set at 0.05. We adjusted for the inadvertent truncation of the burnout scale by converting responses of “1” (“a few times a year or less”) to “0” (“never”) for the emotional exhaustion and depersonalization subscales, while keeping the personal achievement subscale a 6-item scale. Responses greater than 1 were not converted. This change underestimated the prevalence of burnout, as burnout is calculated on the basis of cumulative scores.
To analyze qualitative data from the open-ended question, two team members (TH and FE) used the constant-comparison method (18) to review a random subset of 40 responses to develop a coding scheme of 15 possible coping approaches. Two other team members (EC and JC) independently coded the responses. Because responses were multifaceted (with multiple approaches to coping identified in one response), coders could assign up to four themes for each response. Initial inter-coder reliability was measured, with kappa (κ) at 0.84. Coders resolved disagreements, establishing consensus for each response through discussion.
A total of 336 students in the first 3 years of medical school returned the survey (69.3% response rate). This comprised 150 MS1s (response rate: 82.0%), 118 MS2s (68.2%), and 98 MS3s (57.7%); 56% were women. Almost half (45.4%) reported their race or ethnicity as Caucasian, 33.1% as Asian, 10.9% as Hispanic, and 4.1% as African American. The demographic characteristics of our sample were consistent with the demographic characteristics of our full student body.
The percentage of students reporting high levels of emotional exhaustion, depersonalization, and low personal achievement were 51.7%, 44.0%, and 52.0%, respectively, with varying percentages of students in each class (see Table 1). Analysis of variance revealed significant differences between MS1s, MS2s, and MS3s for depersonalization and personal achievement, but not for emotional exhaustion. Post-hoc analyses revealed that MS1s reported significantly less depersonalization than MS2s or MS3s, although MS2s and MS3s were not different from each other. MS1s also reported significantly less personal achievement than MS2s, although there was no difference between MS3s and either MS1s or MS2s. About 60% of students (59.6%) were “symptom-screen positive” for depression, with no significant differences between classes. Conversely, there was a significant difference between the classes in perception of medical school stress, with MS1s reporting less stress than either MS2s or MS3s. However, all effect sizes for all of these differences were small.
TABLE 1.Percentages of Medical Students Experiencing Burnout,aa Stress,bb and Depression,cc and Mean Scores Among Class Cohorts
| Add to My POL
|Maslach Burnout Inventory|
| Emotional Exhaustion|
| % With High Burnout||53.8%||51.3%||49.0%||51.7%|
| Mean (SD)||26.6 (9.8)||25.9 (11.1)||27.0 (10.6)||26.5 (10.4)||NS|
| % With High Burnout||34.3%||46.0%||55.8%||44.0%|
| Mean (SD)||7.2 (5.7)||9.5 (5.4)||10.8 (6.0)||9.0 (5.9)||<0.001dd|
| Personal Achievement|
| % With High Burnout||57.6%||45.1%||52.1%||52.0%|
| Mean (SD)||30.9 (7.0)||33.5 (6.6)||32.9 (6.7)||32.3 (6.8)||<0.01ee|
| % Met Criteria||99 (66.4%)||99 (66.4%)||65 (55.1%)||53 (54.6)||217 (59.6%)|
|Medical School Stress Scale|
| Mean (SD)||38.4 (9.5)||41.5 (9.5)||42.4 (8.7)||40.5 (9.4)|
Of the 13 items in the Perceived Medical School Stress Scale, 5 stressors received responses of “Agree” or “Strongly Agree” from the majority of students: “Decisions regarding electives and clerkships are made on the basis of information obtained from fellow students and not from the faculty.” (72%); “I am concerned that I will be unable to master the entire pool of medical knowledge.” (70%); “Personal finances are a source of concern to me.” (61%);“Medical training controls my life and leaves too little time for other activities.” (57%);“I am concerned that I will not be able to endure the long hours and responsibilities associated with clinical training and practice.” (51%).
A total of 189 students (56.3%) responded to the open-ended question concerning which school programs helped students cope with stress, anxiety, or burnout. About one-third (35.4%) regarded nothing offered by the institution as helpful or had not used school resources. Of the 133 students who reported something that helped them cope, 99 of these found at least one institutional program helpful. Among these 99, 38 students reported other students and student-oriented programs as helpful, with the “PRN” program, a student-led, peers-as-resource networking program, cited by most; 28 students regarded faculty and mentors as helpful in coping, whereas 25 found structured events such as courses and student-affairs activities helpful, and 28 found counseling helpful. Of the 32 who cited noninstitutional coping approaches, responses included religious organization or practice, family and friends, extracurricular activities such as volunteering and exercise, and time away from coursework.
The prevalences of burnout, depression, and stress were higher in this sample of first- through third-year medical students, as compared with other medical-student groups (1, 2, 4, 5, 7, 9–13). Approximately 50% of students met criteria for high burnout in each of the three domains. In contrast, approximately 30%–40% of students met criteria for high burnout in each of the three domains in a seven-institution study (9). These findings indicate that previous studies may not have captured the entire range of experiences of burnout across medical schools. Administrators, deans and medical educators should therefore appreciate this possibility of higher levels of burnout within their own medical student groups.
Similarly, the prevalence of depression and stress, even when using the same measurement tools, were also higher than those previously reported (1, 2, 4, 5, 7, 9–13); 60% of students answered affirmatively to either of the two depression screening questions, whereas 47% of students in the earlier multisite study answered affirmatively to either of these two questions (9). One earlier study also found a lower level of stress than that which was reported here (7). Earlier studies have found lower rates of burnout among minorities, as compared with non-minorities, and no differences in depression by race/ethnicity. Minorities who experience discrimination or adverse experiences involving race, however, have higher rates of burnout (1). Mixed findings are reported in comparisons of depressive symptoms between genders, with at least half of studies reporting no difference (19–25). Longitudinal studies consistently show higher rates of depression and suicidal ideation in second-, third-, and fourth-year students than in first-year students (9, 25, 26).
This study did not compare burnout and depression among race/ethnicity or between gender groups, but analyzed results by class because of the relatively unique 1.5-year curriculum. We had wondered whether burnout and depression levels would be lower among students in the clinical years because of a more flexible schedule during this time as compared with other schools. Although our results cannot be compared directly with those of other schools, and timing during the school year and closeness to stressful academic or personal events can be a factor in the varying results between studies, it has been suggested that mental health worsens as students progress through medical school (9, 25, 26). In our study, however, there was no significant difference in the level of depression among class years.
The first-, second-, and third-year medical students surveyed identified several factors that were beneficial in improving wellness and relieving stress. Students identified support from faculty, peers, and relationships outside of medical school as helpful. Counseling services and extracurricular activities were also perceived to be helpful. Respondents valued the sense of personal achievement gained from extracurricular activities, and this may be especially important in the preclinical years of medical school. Previous studies have noted that medical students develop a sense of personal accomplishment as they progress through medical school, which appears to counterbalance the burnout resulting from emotional exhaustion and depersonalization (10). Also, because emotional exhaustion was highest among the first-year students, intervention programs may be of most benefit if initiated in the first year.
Although this institution had been involved in activities identified by the literature and described above to promote student health and well-being, recent student suicides at this school prompted the medical college administration to initiate formal work groups comprising faculty and students in order to identify additional areas of student well-being in need of improvement. The time interval between the most recent adjustments made and this survey, however, was only approximately 6 months. Ongoing research is underway to evaluate the efficacy of these interventions. First-year students, who experienced greater interventions through the school, including pass-fail grading and the availability of a new stress-reduction elective, experienced slightly less medical-school stress.
Limitations of the study included its cross-sectional design, which did not allow the identification of factors that might cause burnout. Second, focus groups might be warranted to help determine what contributes to stress and what helps students cope. Moreover, we reported here on only the first 3 years of medical school, and there may be a response bias due to the lack of fourth-year participation and the 30% of students who did not respond, and the findings may not be generalizable from this one school. Another limitation is the lack of baseline data obtained on mental health measures for these students before their entry into this medical school. This limits our ability to evaluate for differences in vulnerability to burnout, depression, and stress present in typical accepted applicants, which may be higher in our institution, based on our criteria for acceptance. Finally, the PRIME-MD tool has limited specificity, and even for those correctly identified as depressed, it does not indicate the severity of depression. The PRIME-MD is limited to use as a screening tool that detects those at risk for major depression.
In conclusion, our findings underscore those of earlier surveys by emphasizing the importance of vigilance for burnout, stress, and depression in medical students. The high rates of burnout, stress, and depression found here suggest that further study into this area is indicated. In particular, longitudinal studies to delineate the relationship between these variables and efficacy studies are needed to determine the potential benefits of wellness programs. Multisite studies measuring levels of psychological distress at schools with shorter preclinical curricula, as compared with longer ones, would also be helpful. A medical school culture that fosters resiliency through peer-support, mentorship, extracurricular activities, and teaching self-care from the start of the first year may be beneficial in improving the mental health of students. More work needs to be done in identifying, preventing, and reducing the rate of these psychologically distressing symptoms in medical students.
We acknowledge Cayla Teal, Ph.D., and Britta Thompson, Ph.D., for their contributions in study design, statistical analysis, and editing, as well as Toi Harris, M.D., for her contribution in study design, coding, and editing.