In 2017, the American Psychiatric Association (APA) Board of Trustees Workgroup on Psychiatrist Well-Being and Burnout was charged by Anita Everett, M.D., then president of APA, with studying psychiatrist well-being and burnout. The workgroup surveyed APA members to determine the prevalence of symptoms of burnout and depression and to identify demographic and practice characteristics that contribute to higher levels of burnout and depression.
Burnout is notoriously difficult to characterize (
1), but Maslach’s definition points to emotional exhaustion, depersonalization, and decreased personal efficacy as key elements and conceptualizes burnout as a response of individuals to a stressful workplace (
2). Burnout is regarded as an experience rather than as a diagnostic entity, and it is better understood dimensionally rather than categorically. By contrast, depression reflects specific psychiatric diagnostic characteristics and is conceptualized as an individual problem arising in response to a unique set of biological, psychological, and social vulnerabilities. There is a lack of consensus on the relationship between burnout and depression, with different authors viewing them as either the same or as different constructs (
3,
4).
The prevalence of burnout among physicians ranges from 0% to 80.5% (
1) because of differences in sample characteristics, response rates, and instruments used to measure burnout; a frequently cited study found a burnout rate among physicians of 54.4% (
5). A recent report of medical school faculty, across all disciplines, showed that 29% of respondents experience burnout (
6). Depression among physicians has been more difficult to study. A variety of screening tools have yielded screen positive rates of 11.3% of 1,145 physicians in Michigan (
7), 10.8% of physicians in the United Kingdom (
8), and 23.2% of 3,213 physicians in Canada (
9). Rates of depression appear to be higher among resident physicians (residents), and a meta-analysis of 54 studies of residents suggests a prevalence rate of 28.8% (
10). Suicide rates for physicians are reported to be higher than those in the general population, with a strikingly higher relative risk among female physicians (
11). Shanafelt and colleagues found that 6.4% of physicians reported having suicidal ideation in the past 12 months (
5).
Proposed drivers of burnout include the electronic health record, loss of the sense of community among doctors, distance from control over workflow, shame related to medical errors, and inefficiencies in the workplace (
12,
13). Burnout takes its toll on physicians’ joyous commitment to a meaningful calling and is also associated with decreased quality of care, increased medical errors, decreased patient satisfaction, decreased academic productivity, earlier retirement, and greater financial burden for already strapped health care systems (
14–
17).
Psychiatrists share fundamental similarities with colleagues in other specialties but may have some unique characteristics. For example, psychiatrists are seen as particularly interested in the human condition, may desire a less demanding professional lifestyle, and are more likely to practice out-of-network and continue working to an older age (
18). The limited data available about psychiatrist well-being and burnout indicate that psychiatrists are less burned out than physicians in other specialties and that they have slightly greater satisfaction with work-life balance (
19). In addition, some correlates of burnout (e.g., trauma exposure) have been identified in this group (
20). Work hours and lack of supervision were found to be correlated with burnout among psychiatric trainees (
21). Several small studies on the prevalence of depression point to higher rates of depression among psychiatrists compared with other physician specialties (
22). Yet, to our knowledge, there are no studies to date that examine the factors associated with both burnout and depression among psychiatrists.
This study aims to assess the prevalence of burnout and depression among psychiatrists, identify demographic and work factors associated with burnout and depression, and examine the correlation between burnout and depression among psychiatrists.
Methods
Sample
Psychiatrists, including APA members and nonmembers, were invited to participate in an online survey (
https://www.psychiatry.org/wellbeing) on well-being and burnout through multiple modalities, including APA media outreach, direct e-mail, electronic announcements, newsletters, electronic mailing lists, and solicitations at a variety of town halls and conferences. The online survey tool was designed for easy use and provided respondents immediate feedback on their scores on measures of burnout and depression (see below), with a graphical comparison to all other survey respondents upon completion of both measures. The survey was open from October 30, 2017, through December 10, 2018.
A total of 2,588 individuals completed the online survey, which included measures of demographic data, depression, and burnout. A total of 504 physicians from other specialties and nonphysicians were excluded, leaving a sample of 2,084 psychiatrists. No identifying information was collected, and an APA institutional review board approved the study and provided a waiver of both written and verbal consent.
Burnout Measure
Respondents completed the Oldenburg Burnout Inventory (OLBI), a 16-item scale designed to measure burnout across a wide variety of occupations that has been extensively used internationally. The OLBI was chosen because it includes questions that reflect both burnout and well-being and is highly correlated with the Maslach Burnout Inventory (
23). A score of 35 (out of 64) or above has been used in the literature as the cutoff score to indicate an increased risk for burnout (
24).
Depression Measure
The Patient Health Questionnaire–9 (PHQ-9) is a 9-item self-report scale that has been widely utilized. A cutoff score of 10 and above is used to screen for the presence of moderate or severe depression (
25).
Demographic and Work-Related Factors
Demographic factors included age, gender, race/ethnicity, and years since medical school graduation. Respondents also provided data on their work setting (academic, community, government, etc.), work type (outpatient, inpatient, combination of outpatient and inpatient), perceived ability to control work schedule (yes or no), work schedule (full time or part time), hours per week of direct clinical face time, career status (medical student, resident, early-career professional [defined as being within 7 years from graduation or residency-fellowship training], and mid and late career), and geographic location (New England and eastern Canada, New York, mid-Atlantic, etc.). These factors were chosen because they were hypothesized to be associated with burnout and depression.
Statistical Analysis
Data were analyzed using linear regression analysis, with the OLBI score as the dependent variable and with demographic characteristics, psychiatry work variables, and depression score, as measured by the PHQ-9, as independent variables. Independent variables that were associated with the dependent variable at an alpha of 0.10 were included in the adjusted linear regression.
Similarly, a linear regression analysis was conducted with the PHQ-9 score as the dependent variable and with demographic characteristics, psychiatry work variables, and burnout scores, as measured by the OLBI, as independent variables. Independent variables that were associated with the dependent variable at an alpha of 0.10 were included in the adjusted linear regression. Finally, a linear regression analysis was conducted with the OLBI score as the dependent variable, to compare psychiatrists who endorsed suicidal ideation on the PHQ-9 with those who did not.
For all analyses, we used linear regression with log link function and an alpha of 0.05. SAS, version 9.4, was used to carry out analyses (SAS Institute, Cary, N.C.).
Results
Sample Characteristics
Participants’ demographic characteristics are summarized in
Table 1. Of note, our sample had a higher proportion of women and was younger than APA members in general. The sample was 58.5% female (41% of APA members are female), and the age breakdown reveals a higher proportion of resident-fellow and early-career psychiatrists in our sample than in the APA membership.
OLBI and PHQ-9 Scores
The mean OLBI score of the sample was 40.4 (SD=7.9) (
Table 1); 78% of respondents had scores that were above the cutoff score of 35, indicating a positive screen for burnout. The mean PHQ-9 score of the sample was 5.1 (SD=4.9); 16% of respondents had a PHQ-9 score ≥10, consistent with moderate to severe depression.
Demographic and Work Characteristics Associated With Burnout
In the unadjusted analysis, depression, younger age, female gender, African American race, lack of control over schedule, earlier career phase, full-time worker, nonacademic settings (e.g., multispecialty clinic, inpatient, community, or government setting), and more than 20 hours per week of direct clinical face time were associated with higher levels of burnout (
Table 2). Age older than 61 and part-time schedule were associated with lower levels of burnout.
After accounting for each of the factors found to be significantly associated with burnout in the unadjusted model, the adjusted model found that the following factors remained significantly associated with high burnout scores: depression, including mild depression (relative risk=1.16, 95% CI=1.14–1.17), moderate depression (relative risk=1.24, 95% CI=1.21–1.26), and severe depression (relative risk=1.31, 95% CI=1.27–1.34); female gender (relative risk=1.02, 95% CI=1.01–1.04); self-report of having no control over schedule (relative risk=1.07, 95% CI=1.06–1.09); and practice in inpatient (relative risk=1.03, 95% CI=1.00–1.06), community (relative risk=1.05, 95% CI=1.02–1.07), and government settings (relative risk=1.04, 95% CI=1.01–1.07). Age was inversely associated with burnout, showing a significant negative association with burnout for respondents ages 61–70 (relative risk=0.96, 95% CI=0.93–0.99) and those older than 70 (relative risk=0.86, 95% CI=0.82–0.91).
Demographic and Work Characteristics Associated With Depression
In the unadjusted analysis, burnout, younger age, female gender, African American race, lack of control over schedule, earlier career phase, multispecialty clinic, inpatient, community, and government settings, and more than 20 hours per week of direct clinical face time were associated with higher levels of depression (
Table 3). Age older than 61 and Asian and Asian American race were associated with lower levels of depression.
After accounting for each of the factors that were found to be significantly associated with depression in the unadjusted model, the adjusted model found that the following factors remained significantly associated with high depression scores: an OLBI score ≥35 (relative risk=2.98, 95% CI=2.38–3.73); female gender (relative risk=1.09, 95% CI=1.01–1.18); resident (relative risk=1.25, 95% CI=1.08–1.46) and early-career professional (relative risk=1.14, 95% CI=1.01–1.28) status; and practice in nonacademic settings, such as outpatient (relative risk=1.19, 95% CI=1.03–1.37), inpatient (relative risk=1.23, 95% CI=1.07–1.42), community (relative risk=1.21, 95% CI=1.04–1.40), and government settings (relative risk=1.21, 95% CI=1.02–1.45). There was a significant inverse association with depression for Asian and Asian American race (relative risk=0.83, 95% CI=0.73–0.93).
Relationship Between the OLBI and PHQ-9 Scores
The increased relative risk between OLBI scores among nondepressed, mildly depressed (score of 5–9 on the PHQ-9), moderately depressed (score of 10–14 on the PHQ-9), and severely depressed (score ≥15 on the PHQ-9) respondents is shown in
Table 2. The prevalence of mild, moderate, and severe depression for respondents with and without burnout is shown in
Table 4, which indicates that 20.2% of those who screened positive for burnout had moderate, moderate to severe, or severe symptoms of depression. The unadjusted Pearson correlation coefficient between the PHQ-9 and OLBI scores was 0.566 (p<0.001).
Suicidal Ideation
Compared with psychiatrists without suicidal ideation, those who endorsed suicidal ideation on the PHQ-9 had a higher burnout score (relative risk=1.17, 95% CI=1.14–1.20, based on the unadjusted model). After adjusting for the PHQ-9 (excluding the suicidal ideation component) and other significant factors, this association was no longer statistically significant (relative risk=0.98, 95% CI=0.96–1.01) (
Table 5).
Discussion
These data replicate the finding that psychiatrists, like other physicians, have substantial burnout. Although categorical cutoff points in burnout scales have questionable significance because burnout is best understood as a dimensional phenomenon, 78% of psychiatrists in our sample had OLBI scores above the cutoff score of 35. Although this is higher than the rate of 47% that Shanafelt et al. (
5) found using the Maslach Burnout Inventory, the rates cannot be directly compared even though the OLBI and the Maslach scale are highly correlated (
26). Sixteen percent of our respondents reported moderate or severe depression, consistent with the range of estimates in the literature (
7–
10).
Our findings suggest that female psychiatrists appear to be at risk for high levels of both burnout and depression. This may in part explain the significant reduction in female physician work hours and attrition from medicine (
27). Increased vulnerability to burnout and depression among female physicians (
6) has been hypothesized to be related to greater challenges with work-home balance, gender inequality in pay, sexual harassment, and frustration regarding greater aspirations for the extent of care they want to provide relative to male physicians and the constraints on their ability to do so (
13,
28–
30). Previous research demonstrates that more than one-third of the relationship between gender differences in depression among medical interns is explained by work-family conflict (
31). Programs that have the potential to ease the competing and incompatible demands of work and family have the potential to reduce the risk for depression and, potentially, burnout among female psychiatrists. In addition to targeted work-family programs, creating schedules and positions with increased autonomy and control over one’s schedule is likely to benefit both male and female psychiatrists.
Residents and early-career physicians have repeatedly been found to have high rates of burnout (
32–
34) and depression (
10), and these findings are consistent with ours. Potential explanations that have been suggested for these findings include greater medical school debt, greater productivity demands, work-home balance for those with young families, less supervision and/or time with colleagues, insufficient time in the profession to have found the best career niche, and exposure to burned-out mentors and senior colleagues. Newer psychiatrists are also more exposed to the demands of the changing health care system, while mid- and late-career psychiatrists may have had the opportunity to find positions that minimize some of these stressors.
There is a substantial literature in medicine and in occupational health that suggests that autonomy promotes health and well-being (
35), and our finding that a lack of control over one’s schedule is associated with an increased relative risk of burnout is consistent with this. Finally, work settings such as inpatient, community, and government settings were associated with an increased risk for both burnout and depression, and outpatient and multispecialty clinics were associated with increased risk for depression. Relatively little is known about the relationship between workplace setting and burnout.
Burnout and Depression
Although our data do not allow us to test causal relationships, they add weight to the understanding that there is significant overlap between burnout and depression. In our sample, virtually all psychiatrists (98%) who rated themselves as moderately or severely depressed also rated themselves as having significant burnout. The clinical evaluation of the burned-out and depressed physician presents significant diagnostic and treatment dilemmas. Psychiatrists are well poised to provide input to the rest of medicine on well-being and burnout given their expertise in psychiatric, stress-related, painful, and/or traumatic human experiences. At a minimum, our finding raises an important clinical consideration for all psychiatrists treating physicians with burnout: major depressive disorder must be part of the differential diagnosis of any evaluation for burnout, as some 20% of those with burnout are likely experiencing a depressive episode.
Although previous studies have demonstrated an association between suicidal ideation and burnout (
36), our findings are consistent with other studies that have not found a correlation between burnout and suicidal ideation after controlling for other depressive symptoms (
4). Our study highlights the importance of accounting for other factors that are associated with burnout when examining the relationship between burnout and suicidal ideation and may explain the inconsistent findings in previous research.
Burnout, Psychiatric Workforce, and Policy Implications
The psychiatrists most at risk for burnout and depression—women, younger psychiatrists, those on the front lines of inpatient and community work, and possibly minority psychiatrists—are those whose participation is arguably most critical for the long-term health of the psychiatric workforce. The proportion of psychiatrists who are female is increasing—in 2017, 50.5% of residents and 39.1% of active psychiatrists were female (
37)—a long-term trend that is expected to continue. Our specialty is aging (
18), and a robust flow of young psychiatrists will be necessary to replenish our ranks. Schedule control and other elements of autonomy are at the center of many discussions about the changing identity of psychiatrists and the very significant trend toward employment and system change. Finally, younger psychiatrists are significantly more diverse, and increasing the minority psychiatric workforce is critical to addressing the significant disparities in care and providing effective care to marginalized populations.
We believe that these findings have an important workforce policy implication. Because burnout has been associated with a move to part-time status, increased leaves of absence, job change, and early retirement (
15,
16), interventions to decrease burnout are also interventions to enhance the psychiatric workforce. Systemic interventions for burnout should target especially those at highest risk, who appear to be female psychiatrists, psychiatrists early in their careers or in training, and those with less scheduling autonomy. Health care organizations can easily identify those at higher risk by screening for burnout and preferentially allocating their scarce resources toward these individuals. Health care organizations should also consider devoting resources to interventions focused on decreasing burnout among minority psychiatrists because of their potential increased risk and essential participation in the workforce. Staff health programs designed to meet the needs of physicians—confidential, accessible, and unstigmatized—will help detect and treat depressed psychiatrists.
Strengths and Weaknesses of Sample and Methodology
To our knowledge, this is the largest sample of psychiatrist well-being and burnout data reported in the literature. However, it is not possible to estimate the participation rate given our sampling methods; thus, our sample may not be representative of psychiatrists nationally and is subject to selection bias. For example, it is possible that those psychiatrists struggling with burnout and/or depression might be more likely to be interested in participating in the survey compared with those not experiencing depression and/or burnout. However, it is also possible that those struggling with burnout and/or depression may be more unlikely to participate in research because of the additional time and energy required for participation, relative to those not experiencing depression and/or burnout. It is difficult to obtain data on the demographic characteristics of psychiatrists in the United States to assess the similarity of our sample to psychiatrists in general, but in comparison to the demographics of APA members, our sample is younger and has a higher proportion of women.
We utilized self-report measures of depression and burnout. Although the PHQ-9 is a well-validated measure of depressive symptoms, there is much discussion in the literature about the validity of the scales used to assess burnout. The OLBI is highly correlated with the Maslach Burnout Inventory, and we report our main findings from the OLBI as a continuous measure (
24). The question of the validity of self-report for burnout is important, as self-report could be subject to context and may fluctuate over time. However, no objective measures of burnout are yet available.
Next Steps
We propose five directions for further work in the area of psychiatrist well-being and burnout, several of which are already under way among physicians in general.
1.
The limitations of our sample may have precluded findings about the relationship between burnout and depression and race/ethnicity, and more work needs to be done to collect data among minority groups underrepresented in medicine. In addition, studying other subgroups, such as part-time psychiatrists or those in particular workplaces, may yield valuable information that would allow additional targeting of interventions to decrease burnout.
2.
An important area of investigation is identifying individual and workplace factors that could predict burnout, depression, and suicide risk among psychiatrists so that preventive interventions can be developed and employed. Potentially important individual and workplace factors include work-home balance, such as family status and caregiver burden for family members (
38), sense of belonging to the physician community, efficiency of the workplace, ease of use of electronic health records, perception of response to medical errors, and other psychiatry-specific factors (e.g., out-of-network practice).
3.
There is a gap between the quantitative data gathered from studies like ours and the reports of individual physicians and psychiatrists with lived experience. Systematic qualitative data would allow for a more nuanced analysis of some of the phenomena associated with burnout and depression.
4.
To our knowledge, there are no data on the effect of interventions for burnout among psychiatrists, except for one study of an intervention in several psychiatry residency training programs (
39) focused on building resilience among individuals. In addition to individually targeted interventions, there needs to be a greater focus on health care system–level interventions, as burnout is fundamentally an organizational problem that affects individuals.
5.
The relationship between burnout and depression deserves much more study. Factor analysis of survey data could be helpful in defining the relationship, but a longitudinal study of physicians and psychiatrists as they progress through training and practice would be the most valuable method for clarifying the temporal relationship and overlap between burnout and depression, as well as the impact of interventions on each.
Acknowledgments
The authors thank their colleagues on the APA Board of Trustees Workgroup on Psychiatrist Well-Being and Burnout, including Carol Bernstein, M.D., Deanna Chaukos, M.D., Julie Chilton, M.D., Matthew Goldman, M.D., James Lomax, M.D., Theresa Miskimen, M.D., M. Steven Moffic, M.D., David Pollack, M.D., Tony Rostain, M.D., Linda Worley, M.D., and Glenda Wrenn, M.D., for the workgroup’s participation in and support for this study; Anita Everett, M.D., for convening the group and highlighting the need for psychiatrist well-being; and numerous APA staff members for helping to implement the online survey tool and analyze the data.