For nearly half a century, stress in medical training has been a topic of concern.1–3 Trainees' stress during medical school3 and residency training2,4 has been well documented in the literature. Common stressors include heavy workload, sleep deprivation, difficult patients, poor learning environments, financial concerns, information overload, and career planning.2 These stressors often exert negative effects on students' and residents' academic performance, physical health, and psychological well-being, making them more susceptible to depression.2,3,5,6–12
Most of these studies of depressive and suicidal indicators among medical trainees (medical students and residents) were conducted years ago or involved only one medical school or program. Few minorities and women were represented in these studies. Variable prevalence rates for depression among medical students and residents have been reported ranging from 2% to 35%, with the highest rates among residents.1,3,13–18 Despite this range, there is more evidence to support that medical students and residents experience depression at higher rates than graduate students or young adults in the general public (8%–15%).16,19–22 Furthermore, in a recent study of medical students at three medical schools, levels of depressive symptoms varied by year of training, with the highest reported during the second year.21 Suicidal ideation among medical trainees has been investigated less often and only periodically. Clark and colleagues1 reported that 25% of interns have had suicidal ideation at one time. Although no studies have reported rates of suicidal ideation among medical students, in 2000, the National College Health Assessment Survey found that 9.5% of students preparing for health-related fields reported that they had experienced suicidal ideation.23
Women in the general population have a higher lifetime risk of depression than men. Similarly, studies of residents have consistently shown higher rates among females.10,14,18 However, studies comparing depressive symptoms by gender among medical students have been inconclusive, showing either no difference by gender or higher rates among female medical students.24 The few studies examining depressive symptoms among minority medical trainees have had mixed results, reporting no differences, lower levels, and higher levels. Clearly, more research is needed in this area. The purpose of our multisite, anonymous study was to assess depressive symptoms and suicidal ideation among medical students and residents throughout the United States using the Center for Epidemiologic Studies–Depression scale (CES-D). Differences by trainee level, gender, and ethnicity are examined herein. We hypothesized that rates of depression and suicidal ideation would be highest among women and minorities.
Medical students and residents from all programs (including, but not limited to, internal medicine, obstetrics–gynecology, pathology, pediatrics, psychiatry, and surgery) at six sites were recruited. The sites were the University of Hawaii School of Medicine (D.G., D.T.), University of Iowa College of Medicine (K.E., J.T.), University of Cincinnati College of Medicine (C.B.), University of Southern California School of Medicine (J.S.), University of Texas Southwestern Medical School (M.K.), and University of Washington School of Medicine (P.B., A.K.). These sites were chosen because they reflected diverse populations and were willing to take part in the study. Our intent was to include schools across the country with various program sizes and ethnically diverse trainees.
In spring/summer 2003 or spring 2004 (April though August, depending on IRB approval), trainees were asked to complete the CES-D (a widely used self-report instrument designed to measure current depressive symptomatology in community populations25 and also to screen patients26), the Primary Care Evaluation of Mental Disorders (PRIME-MD, based on the DSM IV criteria), and social/demographic questions. The study investigators or research assistants distributed the surveys to the classes. The surveys were administered before mandatory lectures or other forums for students or residents where attendance was expected at each of the sites. First- through fourth-year students and PGY1- through PGY4-level residents were included at each of the six sites. Cover letters described the project and solicited participation. All responses were anonymous. Participants completed the forms, placed them in a separate box, and received a snack item in appreciation for returning the survey. Between 350 and 625 surveys were distributed per site. A small fraction of participants (fewer than 10 per site) anonymously left the completed survey for the site contact. Envelopes containing the forms were sent to the primary investigator (D.G.). The study was approved by the IRBs of all participating sites. Because of the anonymous nature of the study, written consent was not required. Agreement was assumed if the survey was returned.
The demographic section included site, trainee level, ethnicity, gender, call schedule, average number of hours of sleep in the last week, rotation, and difficulty of rotation (ranked from zero = not at all to three = extremely difficult). Participants were asked whether they had a history of depression and, if so, whether they had received treatment. They were asked whether they were currently receiving treatment for a mental health issue. They were also asked whether they had a family history of depression; these were yes/no questions.
The suicidal ideation question was taken from the PRIME-MD (although students completed the PRIME-MD in its entirety, just one question was used for this study). Medical trainees were asked how often they were bothered by thoughts that they would be better off dead or in hurting themselves in some way in the last two weeks; the four possible answers were not at all, several times, half the time, and nearly every day.
Participants completed the CES-D. Participants rated 20 items on a scale ranging from zero (rarely or none of the time) to three (most or all of the time). Based on positive predictive value, sensitivity, and specificity for depression, cutoff scores have been proposed for the CES-D. We applied these CES-D cutoff scores to assess three levels of depression—no depression (CES-D total < 16), mild to moderate depression (CES-D total between 16 and 21), and probable major depression (CES-D total > 21).27
CES-D cutoff categories were calculated. Using χ2, CES-D categories and suicidal ideation were compared by levels of training, gender, ethnicity, depression history, and training-related variables. ANOVA was used to assess mean differences in amount of call and sleep by CES-D categories. A two-sided significance level of .05 was used to determine the statistical significance of observed differences. SPSS version 16 was used for all analyses.
Out of a possible 2,475 surveys, a total of 2,193 surveys (89%) were completed. Medical students returned 1,343 surveys, and residents returned 679 surveys. The estimated response rate for medical students was 95%, of which 88% had CES-D data and 90% had answered the suicidal ideation question, and the estimated response rate for residents was 64%, of which 78% had CES-D data and 83% had answered the suicidal ideation question. Participants were not excluded if they did not answer all questions; for each analysis, participants were included if they responded to the related questions. One hundred seventy-one respondents (8%) did not indicate their level of training. The response rate was calculated based on the estimated number of potential respondents by training level.
Fifty-two percent of respondents (1,059) were female. With regards to ethnicity, 51.0% (1,118) identified themselves as Caucasian, 29.5% (648) Asian, 5.6% (122) Hispanic, 3.1% (67) African American/black, 1.3% (29) Pacific Islander, 0.8% (18) and Native American/Alaskan Native; 9% (191) did not indicate their ethnicity.
Seven percent of respondents (159) indicated that they were currently receiving mental health treatment. Seventeen percent of respondents (370) stated they had a history of depression. Of these, 69.1% (256) responded that they had received treatment. Nearly 30% (590) marked that they had a family history of depression.
Of those completing the CES-D, 12% (226) had probable major depression (CES-D > 21), with an additional 9.2% (173) reporting symptom levels consistent with mild to moderate depression (CES-D = 16–21). Respondents with a history of depression had higher rates of probable mild to moderate depression and probable major depression compared with respondents without a history of depression (15.4% and 30.3% versus 8.8% and 9.6%, respectively; χ2 = 76.94, df = 2, P < .001). Respondents with a family history of depression had higher rates of probable mild to moderate depression and probable major depression compared with respondents without a family history of depression (13.4% and 19.4% versus 8.4% and 9.9%, respectively; χ2 = 33.73, df = 2, P < .001).
Table 1 provides prevalence rates for probable mild/moderate depression, probable major depression, and suicidal ideation by trainee type, gender, and ethnicity. Medical students were more likely to be classified as having depression when compared with residents (χ2 = 10.42, df = 2, P = .005). Nearly 25% of medical students (292) were classified as having probable depression compared with 11.9% of residents (63). There were also significant differences by year of training among medical students (χ2 = 18.18, df = 6, P = .006), with first-, second-, and third-year students more likely than fourth-year students to report experiencing depression. There were no statistically significant differences by year of residency training. Women had a significantly higher rate of probable major depression compared with men for both students and residents (χ2 = 22.1, df = 2, P < .001). There were no statistically significant differences in the rates of depression by ethnicity.
Respondents averaged 6.57 hours sleep per night in the week preceding the survey, with no statistically significant differences based on trainee type (6.61 hours for medical students, 6.51 hours for residents). There were statistically significant differences in the hours of sleep by depression category (F = 10.98, df = 2, 1,417, P < .001). Those with probable major depression slept an average of 6.22 hours per night, those with mild to moderate depression slept an average of 6.45 hours, and those without depression slept an average of 6.64 hours. There were no statistically significant differences based on the amount of call (1.07 times in the last week for those without depression, 1.06 times for those with mild and moderate depression, and 0.99 times for those with probable major depression). However, there were significant differences based on the perceived difficulty of the rotation (F = 13.80, df = 2, 676, P < .001). Those with probable major depression and those with probable mild and moderate depression ranked their current rotations as more difficult than those without depression (1.88, 1.77 versus 1.41, respectively).
Of those responding, 102 (5.7%) reported suicidal ideation. Suicidal ideation occurred significantly more often among those with probable major depression (68.5% of 226) compared with those with probable minor or moderate depression (20.4% of 173) and those without probable depression (11.1%; χ2 = 165.83, df = 2, P < .001). The rates of depression among nonrespondents to the suicidal ideation question were not significantly different from the rates of depression among those responding to the question.
Respondents with a history of depression were 3.7 times more likely to report suicidal ideation compared with those without a history of depression (13.7% versus 3.7%; χ2 = 50.70, df = 1, P < .001). Respondents with a family history of depression were 2.3 times more likely to report suicidal ideation compared with those without a family history of depression (9.1% versus 4.0%; χ2 = 36.84, df = 1, P < .001).
Medical students were more likely to report suicidal ideation compared with residents (6.6% versus 3.9%; χ2 = 5.19, df = 1, P = .023) (Table 1). There were no statistically significant differences by year of medical student or residency training. There were no statistically significant gender differences. There were significant differences in suicidal ideation by ethnic group (χ2 = 10.42, df = 3, P = .015), with the highest rate among black/African American respondents (13%) and the lowest among Caucasian respondents (4.5%).
To our knowledge, this is the first large-scale study examining rates of probable depression among medical students and residents. It shows that depression remains a significant issue for medical students and residents despite the major changes in medical and resident training parameters, such as the number of duty hours, as well as changes in trainee demographics. The overall depression reported by our respondents rate was 21.2%, suggesting that medical trainees have higher rates than other graduate students and young adults in the general public (8%–15%).16,19–22 Furthermore, unlike other studies, rates for medical students were more than double those for residents. Previous reports of prevalence rates for depression among medical students have ranged from 15%14 to 23%.3 Generally, higher rates for depression among residents have been reported, ranging between 27% and 35%.1,15 However, among our respondents, suicidal ideation was highest among fourth-year medical students. The reasons why our findings contradict previous studies are unclear, but the differences may be attributable, in part, to cohort effects and study limitations. The response rate from the survey was good, especially for medical students. It is, of course, unknown whether participants who refused to answer the questionnaire had differences in depressive symptoms or suicidal ideation. It is possible that some of the residents who did not respond to multiple questions may be experiencing more sleep deprivation, depressive symptoms, and suicidal ideation than those who responded.
The discrepancy we found in depressive symptom levels by trainee type may also be a reflection of system changes that have been implemented to risks among residents. In 2003, the Accreditation Council for Graduate Medical Education (ACGME) approved program duty hours restrictions, involving a reduction in hours for work. In response to this change, residents reported between “moderate agreement” and “feelings of” neutrality regarding the effect of lowered stress and improved work performance.28 Unlike residents, however, medical students are not protected by work hours limits. Following ACGME changes, the Liaison Committee on Medical Education suggested that similar regulation of work hours be introduced for medical students. However, in our institutions, medical student schedules and workloads may exceed the 80-hour week.
Differences in sleep duration, with less sleep being reported by respondents who had major or mild depression, were expected, because affective disorders result in poor quality and quantity of sleep. Although differences in trainee levels were found, it seems that trainees, including residents, are getting adequate rest with the work hours reduction, because average sleep was greater than six hours for all groups. Interestingly, this number is essentially unchanged from previous studies done well before the ACGME changes.29 However, time on call seemed to have no bearing in our study, a finding that is corroborated by Kirsling and colleagues.30 However, our analyses indicated that amount of sleep and difficulty of rotation were related to levels of depressive symptoms, suggesting that further study is warranted in this area. We did not examine disruption in the sleep cycle, quality of sleep, and consistency of sleep duration, which may increase trainees' risk for depression, nor did we measure age, marital status, financial strains, parental status, or social support. For example, Dyrbye and colleagues demonstrated that both work-related stress and personal life events were related to professional burnout, a measure of distress.24 Although the ACGME changes have resulted in a reduction in the number of residents' duty hours, the other important trainee factors may be similar pre- and postresidency hour changes. It will be interesting to see whether, over time, trainees have lower rates of depression.
Nearly 6% of medical trainees in our sample reported current suicidal ideation. Given that 18% of respondents did not answer this question, actual rates may be higher. That rates of suicidal ideation among medical students were much higher than for residents was not surprising. Vitaliano and colleagues31 noted a doubling of depression between the beginning and end of the medical student year. Medical school also has a greater number of hurdles such as board examinations, applications to residency, and the match process. Medical students often lack a sense of control over these and other events, which may be a factor in the increasing depression and suicidal ideations. By the spring of the fourth year of medical school, board examinations and residency issues are completed. Not surprisingly, depressive symptoms are low during this time. We cannot assume that the suicide risk disappears. Although depressive symptoms may decline during the course of medical training, suicide remains one of the top causes of early death in practicing physicians. In a meta-analysis of 25 studies from 1960 to 2003, rates of male-physician suicide were 1.4 times higher than for the average population and 2.3 times higher for female physicians.32
The number of women training to become physicians has more than doubled since 1980, and women currently make up 49% of medical school classes in the United States.33 We found that rates of depression among women were more than two times higher than for men. In a recent study, Tjia and colleagues14 found slightly higher rates among female medical students. Hsu and Marshall10 noted significantly more depressive symptoms in female residents, with 40% of female interns having at least mild depression. In a large Norwegian study, higher rates of depression were found among women during both medical school training and residency (41% versus 27%).18 Although depressive symptoms significantly decreased for men during residency, symptoms remained high for women.13,18 Although these differences are also seen in the general population, they have implications on designing models of care for trainees. In addition to the stressors and potential burnout associated with training, women may be more vulnerable to pressures of family life and work or not being partnered, as well as succeeding in a traditionally male-dominated profession.25 They may also be the targets of sex-based or sexual harassment. Accordingly, programs should consider not only the mental health of trainees but also the coping skills they will need.
The percentage of minority participants was extremely high in our study (61.5%) compared with previous studies, which have involved predominantly Caucasian males. Participants in our study reflect the ethnic distribution across the six sites, which is higher than the ratio of minorities enrolled in medical trainee programs nationally.33 In our multisite study, we did not find significant ethnic differences in rates of depression. However, the rates of suicidal ideation showed marked differences among minority groups and levels of training despite similar rates of depression, suggesting that suicidal ideation is not synonymous with depression. Given the relatively small number of black/African American students, the high rate can easily be skewed by a few individuals with significant illness. These high rates may be attributable to increasing feelings of isolation from family and communities. Although there were insufficient numbers to report rates for depression and suicidal ideation among Pacific Islanders and Native American/Alaska Natives, there is a need to examine this group further. Given our findings, help-seeking behavior among minority trainees is an important area for further research.
There are a number of limitations to this study. The CES-D asks about depressive symptoms, which are distinct from syndromal depression. The CES-D cannot be used to distinguish between depressive and burnout symptoms. The survey was done in the spring and summer and did not ask about seasonal depressive effects. Only a minimal history was collected. The study did not ask about illicit substances, which may be used to self-treat or which may mask affective symptoms. Self-treatment and self-medication issues are not clear. Previous studies have shown that illicit and substance use was associated with depressed mood.34 By the nature of our study design and excused absences, there was variability in participation by trainee level. Finally, this survey occurred during implementation of the ACGME restrictions in duty hours. Many programs were under a great deal of stress with changes in night float, use of hospitalists, and other systems, which may have added to resident trainees' stress.
This study highlights the importance of continual assessment of the mental health of medical students and residents. Access to mental health care deserves further consideration. Students are concerned that treatment for depression could jeopardize their career.35 Few trainees utilize psychiatric services because of issues of stigma, cost, and accessibility, despite previous studies showing the desirability of such programs.36 In a consensus statement, Center and colleagues19 highlight the low priority given to physician mental health within the culture of medicine, despite evidence of untreated mood disorders and an increased incidence of suicide. Discrimination in medical licensing, hospital privileges, and professional advancement serve as punitive barriers to seeking help. At the same time, medical schools and residency programs continue to implement new ways to teach students to recognize the importance of their own health and to seek help when they need it, such as through anonymous forums, workshops to teach mindfulness, support groups, and “stress” rounds.31 As ongoing assessment of, treatment of, and education about depression become commonplace for trainees, it is our hope that these new physicians will continue to integrate psychological well-being into routine care for themselves, their colleagues, and their patients.
The authors thank Robert Sabalis, PhD, at the Association of American Medical Colleges, who served as a gateway to medical school contacts; Peter Katsufrakis, MD, at USC, for his help in planning the study; the research support staff at the UH Department of Psychiatry for their assistance in obtaining references; and Lanelle Dullong, MD, for her assistance with the pilot work for this study.
The authors are grateful for the funding received from the Queen Emma Research Fund at the Queen's Medical Center in support of this project.
All authors have full access to all of the data in the study, and Dr. Goebert takes responsibility for the integrity of the data and the accuracy of the data analysis.
1 Clark DC, Salazar-Grueso E, Grabler P, Fawcett J. Predictors of depression during the first 6 months of internship. Am J Psychiatry. 1984;141:1095–1098.
2 Levey R. Sources of stress for residents and recommendations for programs to assist them. Acad Med. 2001;76:142–150.
3 Mosley TH Jr, Perrin SG, Neral SM, Dubbert PM, Grothues CA, Pinto BM. Stress, coping, and well-being among third year medical students. Acad Med. 1994;69:765–767.
4 Bellini LM, Baime M, Shea JA. Variation of mood and empathy during internship. JAMA. 2002;287:3143–3146.
5 Firth-Cozens J. Emotional distress in junior house officers. BMJ. 1987;295:533–536.
6 Reuben DB. Psychologic effects of residency. Southern Med J. 1983;76:380–383.
7 Caplan RP. Stress, anxiety, and depression in hospital consultants, general practitioners, and senior health service managers. BMJ. 1994;309:1261–1263.
8 Klamen DL, Grossman LS, Kopacz D. Posttraumatic stress disorder symptoms in resident physicians related to their internship. Acad Psychiatry. 1995;19:142–149.
9 Brewin CR, Firth-Cozens J. Dependency and self-criticism as predictors of depression in young doctors. J Occup Health Psychol. 1997;2:242–246.
10 Hsu K, Marshall V. Prevalence of depression and distress in a large sample of Canadian residents, interns, and fellows. Am J Psychiatry. 1987;144:1561–1566.
11 Maslach C. Burnout: The Cost of Caring. Englewood Cliffs, NJ: Prentice-Hall; 1982.
12 Pines A, Aronson E. Career Burnout: Causes and Cures. New York, NY: Free Press; 1988.
13 Rosal MC, Ockene IS, Ockene JK, Barrett SV, Ma Y, Hebert JR. A longitudinal study of students' depression at one medical school. Acad Med. 1997;72:542–546.
14 Tjia J, Givens JL, Shea JA. Factors associated with undertreatment of medical student depression. J Am Coll Health. 2005;53:219–224.
15 Schneider SE, Phillips WM. Depression and anxiety in medical, surgical and pediatric interns. Psychol Rep. 1993;72:1145–1146.
16 Katz ED, Sharp L, Ferguson E. Depression among emergency medicine residents over an academic year. Acad Emerg Med. 2006;13:284–287.
17 Hainer BL, Palesch Y. Symptoms of depression in residents: A South Carolina family practice research consortium study. Acad Med. 1998;73:1305–1310.
18 Hendrie HC, Clair DK, Brittain HM, Fadul PE. A study of anxiety/depressive symptoms of medical students, house staff, and their spouses/partners. J Nerv Ment Dis. 1990;178:204–207.
19 Center C, Davis M, Detre T, et al. Confronting depression and suicide in physicians: A consensus statement. JAMA. 2003;289:3161–3166.
20 Levine RE, Bryant SG. The depressed physician: A different kind of impairment. Hosp Physician. 2000;36:67–73.
21 Dyrbye LN, Thomas MR, Shanafelt TD. Systematic review of depression, anxiety, and other indicators of psychological distress among U.S. and Canadian medical students. Acad Med. 2006;81:354–373.
22 Kessler RC, Chiu WT, Demler O, Merikangas KR, Walters EE. Prevalence, severity, and comorbidity of twelve-month DSM-IV disorders in the National Comorbidity Survey Replication (NCS-R). Arch Gen Psychiatry. 2005;62:617–627.
23 Kisch J, Leino EV, Silverman MM. Aspects of suicidal behavior, depression, and treatment in college students: results from the spring 2000 national college health assessment survey. Suicide Life Threat Behav. 2005;35:3–13.
24 Dyrbye LN, Thomas MR, Huntington JL, et al. Personal life events and medical burnout: A multicenter study. Acad Med. 2006;81:374–384.
25 Radloff LS. The CES-D scale: A self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1:385–401.
26 Schuulberg HC, Saul M, McClelland M, Gaguli M, Christy W, Frank R. Assessing depression in primary medical and psychiatric practices. Arch Gen Psychiatry. 1985;42:1164–1170.
27 Weissman MM, Sholomskas D, Pottenger M, Prusoff BA, Locke BZ. Assessing depressive symptoms in five psychiatric populations: A validation study. Am J Epidemiol. 1977;106:203–214.
28 Golub JS, Weiss PS, Ramesh AK, Ossoff RH, Johns MM III. Burnout in residents of otolaryngology–head and neck surgery: A national inquiry into the health of residency training. Acad Med. 2007;82:596–601.
29 Schernhammer E. Taking their own lives—The high rate of physician suicide. N Engl J Med. 2005;352:2473–2476.
30 Kirsling RA, Kochar MS, Chan CH. An evaluation of mood states among first year residents. Psychol Rep. 1989;65:355–366.
31 Vitaliano PP, Maiuro RD, Russo J, Mitchell ES, Carr JE, Van Citters RL. A biopsychosocial model of medical student distress. J Behav Med. 1988;11:311–331.
32 Schernhammer E, Colditz G. Suicide rates among physicians: A quantitative and gender assessment (meta-analysis). Am J Psychaitry. 2004;161:2295–2302.
34 Maddux JF, Hope SK, Costello RM. Psychoactive substance use among medical students. Am J Psychiatry. 1986;143:187–191.
35 Rosenthal JM, Okie S. White coat, mood indigo—Depression in medical school. N Engl J Med. 2005;353:1085–1088.
36 Borenstein DB. Should physician training centers offer formal psychiatric assistance to house officers? A report on the major findings of a prototype program. Am J Psychiatry. 1985;142:1053–1057