Physicians have long considered burnout (i.e., emotional exhaustion, depersonalization, and decreased feelings of personal accomplishment1) to be an occupational hazard. Estimates of physician burnout range from 25% to 70%,2–7 and the onset is linked to residency training.8 A longitudinal study of first-year residents documented that residents began training programs with high levels of well-being but that, by the end of the intern year, they were experiencing dramatic levels of depression and burnout.8 Physicians have a higher rate of depression, substance abuse, anxiety, and suicide when compared with the general population,9–13 and burnout may be the precursor to these disorders. Social isolation, self-blame for negative outcomes, lack of attention to emotional needs, inadequate medical care, and strong emotional responses to the care of complex patients may also be causal factors.9–10,13–15 Psychological disorders are underdiagnosed and undertreated among physicians because of a hesitance to admit to personal problems and seek help.12,16
Physician burnout also has a societal cost. Evidence exists linking physician stress and burnout to lower-quality patient care.6–7,17–18 The bulk of research examining physician stress and patient care centers on studies of resident physicians. Multiple studies have shown dramatic increases in the rates of depression, anxiety, and drug use among residents throughout their training with resultant increases in cynicism, decreased empathy, and poorer performance overall.19–21 Residency stressors range from long work hours and sleep deprivation to lack of knowledge and self-doubt.8,11,17,22–27 In addition to financial pressures, there is little opportunity to balance work and home and little time to seek social support.3,11,17,18 Unfortunately, the habit of ignoring self-care is often perpetuated in life after residency.11,28
Clever6 insists that it is not acceptable for physicians to experience this level of psychological pain, and Riley10 states that prevention must be a top priority. Physicians must be taught to recognize and respond to precursors of depression, such as burnout, and there must be a collective effort to draw attention to the issue.10 Residency programs have attempted to address the issue of resident wellness through workshops, retreats, and education, but adequate outcome data are not available. There is initial evidence that the 80-hour-workweek restriction has led to decreases in emotional exhaustion and improved quality of life, but additional data are needed before definitive conclusions can be drawn.22,29–31 Programs that offer more comprehensive and available mental health services report significant gains in utilization since inception.22 Although such programs are costly, if counseling services help prevent one medical error, the program has paid for itself.22 Although experts intuitively believe counseling is effective, no evidence suggests what type of counseling works best. In fact, a major obstacle is that programs do not know what works in the prevention and treatment of physician burnout.32
Based on available data, the promotion of physician wellness must occur early, and intervention programs should be widespread, normalized, and evidence based. Unfortunately, academic medicine knows very little about physicians' wellness needs and even less about protective factors. The community needs to study strengths, virtues, and protective factors so that, collectively, we may nurture what is best, enhance buffers, and not just fix what is broken.11,33
Medical educators need key pieces of missing information to effectively prevent and treat resident burnout. The goals of this study are (1) to identify which resident self-reported stressors are associated with the actual presence of burnout, and (2) to identify which resident self-reported protective factors are associated with the absence of burnout, as a proxy measure to wellness.
We obtained approval from institutional review boards from two hospital systems in Michigan. In October and November, 2006, we sent a letter of introduction and data collection instrumentation to 17 multispecialty residency program directors from these two Michigan hospital systems. The letter asked program directors or residency coordinators to inform 395 residents about a confidential and voluntary opportunity to participate in a study of resident burnout. We asked program directors or residency coordinators to distribute the data collection measures, which could then be placed in individually sealed envelopes for return to the lead investigator. The data collection measures included a demographic questionnaire, a survey of factors that promote burnout and wellness (Appendix), and the Maslach Burnout Inventory (MBI). The survey included the most common stressors encountered in residency training based on available literature3–4,8,11,17,22–25,27 and local faculty experience. The survey also included potential factors that might be protective against burnout based on stress and coping literature.34 We asked residents to rate (on a 5-point, Likert-type scale where 1 = “does not apply” and 5 = “strongly applies”) factors that contribute to their burnout and factors that protect from burnout. The MBI is a well-established measure of burnout with three scales: emotional exhaustion, depersonalization, and personal accomplishment.1 Emotional exhaustion measures the extent to which a person is feeling emotionally overwrought. Depersonalization measures the degree to which a person feels detached toward or cynical about patients. Personal accomplishment measures the level of pride or satisfaction with one's accomplishments. A high level of burnout is defined as high emotional exhaustion, high depersonalization, and low personal accomplishment. The authors of the MBI describe burnout as a qualitative combination of scale scores along a spectrum of medium to high. Medium to high elevations on emotional exhaustion and depersonalization with lower scores on personal accomplishment indicate elevated levels of burnout. There is no particular order to the scales, which are mutually exclusive.
For the purpose of this study, we defined wellness as lower emotional exhaustion, lower depersonalization, and higher personal accomplishment. We considered this formula a proxy measure of wellness because there were no known validated measures of wellness at the time of the study.
We used three different step-down, multivariable general linear predictive models fitted with the factors that promote burnout and wellness to evaluate their association with emotional exhaustion, depersonalization, and personal accomplishment as continuous factors. The first step of each model included all of the burnout factors as continuous factors. The second step of the model included all of the wellness factors as continuous factors. The least significant variable was dropped at each step until only those factors with significant P values (P ≤ .05) remained in the final model. Because of low numbers, we did not, for statistical purposes, attempt to categorize respondents into high, medium, and low on each burnout scale. We completed all statistics using SAS version 9.1.3 (SAS Inc., Cary, North Carolina).
A total of 168 residents, from a pool of 395, completed the research questionnaires. One set of questionnaires was determined to be invalid because the respondent provided the same response to every question, even where inappropriate. We excluded an additional 16 questionnaires because of incomplete data. Our usable response rate was 38%. Specialties that chose not to participate in the study included colon and rectal surgery, emergency medicine, physical medicine and rehabilitation, and radiation oncology. We sent a follow-up to these residency directors, but we received no responses. Because participation was strictly anonymous, we did not collect information on nonresponders.
Table 1 reports participants by hospital, specialty, sex, and postgraduate year. It is important to keep in mind that different specialties require differing numbers of years in residency training. An analysis of the variables in Table 1 revealed no significant relationships between these and emotional exhaustion, depersonalization, and personal accomplishment.
Perceived social support was the only demographic variable significantly associated with levels of emotional exhaustion, depersonalization, and personal accomplishment. As perceived social support increased, emotional exhaustion decreased (P < .03), depersonalization decreased (P < .03), and personal accomplishment increased (P ≤ .001). Where a resident completed medical school (e.g., within the United States or outside the United States) was significantly associated with emotional exhaustion and depersonalization; those physicians who completed medical school outside of the United States had significantly lower emotional exhaustion (P < .01) and lower depersonalization (P ≤ .003), but there were no significant differences in personal accomplishment based on medical school location.
Because there were very few respondents in some residency specialties, we created physician specialty groups based on similarities among the groups. We examined specialty differences, based on the review of Thomas,35 which indicated potential differences among physicians in different specialties. The specialty groups comprised (1) specialist (diagnostic radiology, pathology, ophthalmology, transitional year, urology, nuclear medicine), (2) primary care physician (obstetrics–gynecology, family medicine, internal medicine, pediatrics, medicine–pediatrics), and (3) surgery (general surgery, orthopedic surgery). The specialty group was significantly associated with levels of depersonalization (P < .01). The surgery group had the highest depersonalization scores, and the primary care group had the lowest depersonalization scores. There were no significant differences for emotional exhaustion or personal accomplishment.
Twenty-seven of 32 total burnout factors were significantly associated with at least one burnout scale (Table 2). Only pessimism was significantly associated with all three scales (indicative of higher burnout). The following 11 other burnout factors were significantly associated with two of the three burnout scales:
- lack of coping skills for stress;
- personal bad habits (smoking, recreational drug use);
- lack of control over office processes;
- lack of control over schedule;
- poor relationships with colleagues;
- lack of time for self-care;
- difficult and complicated patients;
- not enough time in the day;
- excessive paperwork; and
- regret over chosen career.
As reports of each of these burnout factors increased, a higher level of burnout was measured.
Twenty-five of the 29 wellness factors were significantly associated with at least one wellness scale (Table 3). Only one wellness factor, use of prescription medications, was significantly associated with all three scales—lower exhaustion, lower depersonalization, higher personal accomplishment—indicative of lower burnout. Thirteen other wellness factors were significantly associated with at least two of the three burnout scales, again (except for one factor) indicating lower burnout. These wellness factors comprised
- use of meditation, relaxation, massage, or other alternatives;
- use of alcohol or illicit drugs;
- use of a support group for physicians;
- talking about feelings;
- use of professional counseling;
- feeling like one has a say in the training program;
- feeling like one has some control over one's schedule;
- having a plan for the future;
- having enough money;
- having a supportive work environment;
- feeling connected to and compassionate toward patients;
- having good coping skills; and
- being very happy with child care.
The remaining 11 wellness factors predicted lower scores—less burnout—on a single burnout scale. As reported use of each wellness factor increased, measured burnout decreased, except for use of alcohol and illicit drugs. Reported use of alcohol or illicit drugs as a wellness tool was significantly associated with higher burnout.
This study determined which resident-identified factors were associated with either the presence or absence of burnout. By comparing our questionnaire of common burnout and wellness factors with a well-studied and validated measure of burnout (the MBI), we were able to describe these associations. The study was exploratory because there is little research regarding the relationship of stressors to physician burnout, and no identified information exists about protective factors. Other investigators may use the information gleaned from this study to guide additional research regarding specific interventions.
The relationship of social support to burnout is expected. Social support is a key factor that protects individuals from many different kinds of stressors.34 The incorporation of social support into any wellness program is integral, but information about what other sorts of interventions are the most useful and effective is lacking. More data are needed.
The finding that surgical specialties were more detached from patients than primary care specialties is also not surprising. Relationships in primary care are ongoing and patient centered, whereas surgical relationships tend to be more short term and problem focused.36 A certain level of detachment may be considered necessary to enable the accomplishment of invasive procedures.
A surprising finding is that physicians who attended medical school outside the United States report significantly lower emotional exhaustion and depersonalization compared with physicians who attended medical school in the United States. There were no significant differences in personal accomplishment between the groups. Regardless of where physicians receive their education, they seem to feel similar levels of pride regarding their accomplishments. However, if a physician attends medical school in the United States, he or she is more likely to feel emotionally exhausted and cynical about patients. This is consistent with the work of Newton et al37 that demonstrates decreases in empathy among U.S. medical students as training progresses. One question is how medical education provided in the United States contributes to burnout. What role do cultural value differences play in burnout levels? For example, how does the American value of accomplishing more in less time compare with medical training in other countries? There is little research comparing use of the MBI between U.S. medical graduates and international medical graduates (IMGs) with which to correlate these findings. We do not know whether IMGs are less likely to endorse symptoms of burnout. One thought is that cultural differences might make it taboo for internationally trained physicians to publicly admit symptoms of burnout. More research on the differences is needed.
Multiple burnout factors were associated with the presence of different aspects of burnout. Pessimism was associated with all three scales, which is not surprising. However, this personality variable is difficult to alter and may be difficult for programs to address. One possibility to consider is measuring pessimism and raising awareness among residents that it contributes to burnout. Another possibility is to teach residents cognitive-behavioral skills to alter negative thinking. The relationship of pessimism to burnout is important for both residents and faculty to understand. It is surprising, however, that more burnout factors did not correlate with all three aspects of burnout.
The relationship between using alcohol and/or drugs and emotional exhaustion is expected. Clearly, medical education needs specialized identification of, and intervention for, at-risk students, yet addressing physician impairment may still feel taboo in some programs. Evidence from this study may be important in consciousness raising. Lack of sleep as a burnout factor was associated with only lower personal accomplishment. This was an unexpected finding because we anticipated that this factor would have a stronger association with burnout. It may be that physicians in training expect a lack of sleep and adjust or adapt to it.
Five burnout factors had no relationship with the burnout scales: not having enough money, poor support staff, job boredom, too many work demands, and worry about malpractice. It may be that physicians-in-training expect and accept these particular demands of the job. For example, residents expect a lot of work in residency, and they expect low pay. It may be that these factors are not troublesome because they are anticipated to be temporary. In addition, there is little time for residents to become bored, and medical education tends to shield residents from malpractice concerns.
There are many opportunities for residency program directors to address resident-identified stressors associated with burnout. Education programs and workshops may help residents explore issues of control, time management, difficult patients, career regret, team building, and paperwork. Mentoring or advising residents while keeping these issues in mind may assist residents with coping skills, provide them with recognition, and identify drug and alcohol use, depression, anxiety, and pessimism. If programs are not financially able to offer specialized counseling services, directors should consider developing a list of local resources for counseling and primary care where confidentiality is ensured. However, we do not know which, if any, of these interventions are effective.
As with the burnout factors, different wellness factors were associated with different aspects of low burnout (a proxy measure of wellness). The use of prescription medications was the only wellness factor associated with lower emotional exhaustion, lower depersonalization, and higher personal accomplishment (i.e., absence of burnout). The wording of the question did not ascertain the type of medication. It may be that individuals who recognize burnout or other psychological conditions and seek medication to manage them may be more likely to be well. Yet, this is not clear and needs to be explored in further research.
Notably, four wellness factors (being an optimist, having good work relationships, spiritual beliefs, and saying no when one needs to say no) were not significantly associated with any of the burnout scales. It is not clear why these factors were not significant. One would expect that optimism, social support at work, spirituality, and boundary setting would be key protective factors. Interestingly, some respondents reported using alcohol or illicit drugs as a wellness tool (rather than seeing it as a sign of burnout); however, use of these was associated with higher emotional exhaustion and higher depersonalization. Activities that physicians tend to avoid,31 such as talking about feelings and seeking professional counseling, were associated with lower emotional exhaustion and lower depersonalization. Many self-care activities, including exercise, meditation and other alternatives, and good coping skills (as defined by the respondent) were also associated with the absence of burnout. Multiple work-related factors, control over schedule, input into training, satisfaction with child care, and a supportive work environment were associated with lower burnout scores on the three scales. These work-related factors may be used to create a better environment in which residents work.
Interestingly, adequate sleep was associated with lower emotional exhaustion, but lack of sleep was not associated with higher emotional exhaustion. The same was true for financial concerns. Having a plan for the future, having a financial plan, and having enough money were all associated with the absence of burnout, yet a lack of these was not associated with the presence of burnout. This means that financial security and a plan for the future protect from burnout, but, if they are not present, individuals do not seem to be at higher risk of burnout. Again, this may be attributable to residents' expectations of medical training. Most residents do not expect to get adequate sleep or have financial security. However, if these things are achievable, they are protective.
Our findings support the idea that wellness encompasses much more than simply the absence of burnout. Unfortunately, there is no clear definition of wellness in the literature, and there is little guidance on how to define and measure it. Program directors may try to bolster self-care activities, a sense of control, and financial planning, but, again, there is no evidence to suggest such interventions are effective. The domain of wellness has largely been ignored in the literature, and this is part of the problem. At this time, we know more about pathology than prevention.
Physicians are at high risk of burnout, which predisposes them to mental health problems and less effective patient care. What we do not know is how to decrease the risk of burnout and its effects. Some components of burnout (i.e., higher emotional exhaustion) are probably inevitable. On the other hand, other components (i.e., lower personal accomplishment) are likely preventable. Educators need to identify and nurture signs of wellness (e.g., participation in groups that facilitate self-reflection) in residents early in the education process. Directors may design interventions to promote low levels of emotional exhaustion and depersonalization and high levels of personal accomplishment. Program directors and other medical educators should also begin to identify those residents at highest risk of burnout, and faculty must role model healthy behaviors (e.g., talking about feelings and showing compassion for patients) in their own lives. Clearly, a multipronged approach that includes consideration of the many factors associated with burnout is needed, and further study of possible interventions is critical. In addition, much more research is needed to more completely define what makes a physician well.
There are significant limitations to this study. The purpose was exploratory, and the data are cross-sectional. The response rate was low, approximately 40%, with some specialties not participating at all. The data represent only those residents who chose to participate at the time of the study. Unfortunately, data are not available on nonresponders, and we cannot make comparisons with the participants. Generalizability to all residents may, therefore, be limited.
Regarding instrumentation, we measured wellness as lower levels of burnout because no validated measures of wellness are known. This method does not adequately measure all of the dimensions of wellness, which is more than the simple absence of burnout. We developed the survey of factors that promote burnout and wellness based on the available literature and expert knowledge. As such, reliability has not been established. Furthermore, we gathered self-reported data, which may not be accurate. Nonetheless, the data presented in this study are consistent with previous research and provide many directions for future research.
1 Maslach C, Jackson SE, Leiter MP. Maslach Burnout Inventory. 3rd ed. Mountainview, Calif: Consulting Psychologists Press; 1996.
2 Castelo-Branco C, Figueras F, Eixarch E, et al. Stress symptoms and burnout in obstetric and gynaecology residents. BJOG. 2007;114:94–98.
3 Panagopoulou E, Montgomery A, Benos A. Burnout in internal medicine physicians: Differences between residents and specialists. Eur J Intern Med. 2006;17:195–200.
4 Goehring C, Gallacchi Bouvier M, Künzi B, Bovier P. Psychosocial and professional characteristics of burnout in Swiss primary care practitioners: A cross-sectional survey. Swiss Med Wkly. 2005;135:101–108.
5 Chopra SS, Sotile WM, Sotile MO. STUDENTJAMA. Physician burnout. JAMA. 2004;291:633.
6 Clever LH. Who is sicker: Patients—or residents? Residents' distress and the care of patients. Ann Intern Med. 2002;136:391–393.
7 Shanafelt TD, Bradley KA, Wipf JE, Back AL. Burnout and self-reported patient care in an internal medicine residency program. Ann Intern Med. 2002;136:358–367.
8 Rosen IM, Gimotty PA, Shea JA, Bellini LM. Evolution of sleep quantity, sleep deprivation, mood disturbances, empathy, and burnout among interns. Acad Med. 2006;81:82–85.
9 Schernhammer E. Taking their own lives: The high rate of physician suicide. N Engl J Med. 2005;352:2473–2476.
10 Riley GJ. On being a doctor—Commentary: Understanding the stresses and strains of being a doctor. Med J Aust. 2004;181:350–353.
11 Shanafelt TD, Sloan JA, Habermann TM. The well-being of physicians. Am J Med. 2003;114:513–519.
12 Stanton J, Caan W. How many doctors are sick? BMJ. 2003;326:S97.
13 Arnetz BB. Psychosocial challenges facing physicians of today. Soc Sci Med. 2001;42:203–213.
14 Meier DE, Back AL, Morrison RS. The inner life of physicians and care of the seriously ill. JAMA. 2001;286:3007–3014.
15 West CP, Huschka MM, Novotny RJ, et al. Association of perceived medical errors with resident distress and empathy. A prospective longitudinal study. JAMA. 2006;296:1071–1078.
16 Center C, Davis M, Detre T, et al. Confronting depression and suicide in physicians: A consensus statement. JAMA. 2003;289:3161–3167.
17 Spickard A Jr, Gabbe SG, Christensen JF. Mid-career burnout in generalist and specialist physicians. JAMA. 2002;288:1447–1450.
18 Firth-Cozens J, Greenhalgh J. Doctors' perceptions of the links between stress and lowered clinical care. Soc Sci Med. 1997;44:1017–1022.
19 Bellini LM, Shea JA. Mood change and empathy decline persist during three years of internal medicine training. Acad Med. 2005;80:164–167.
20 Collier VU, McCue JD, Markus A, Smith L. Stress in medical residency: Status quo after a decade of reform? Ann Intern Med. 2002;136:384–390.
21 Koran LM, Litt IF. House staff well-being. West J Med. 1988;148:97–101.
22 Dabrow S, Russell S, Ackley K, Anderson E, Fabri PJ. Combating the stress of residency: One school's approach. Acad Med. 2006;81:436–439.
23 Bruce SM, Conaglen HM, Conaglen JV. Burnout in physicians: A case for peer support. Intern Med J. 2005;35:272–278.
24 Gopal R, Glasheen JJ, Miyoshi TJ, Prochazka AV. Burnout and internal medicine resident work-hour restrictions. Arch Intern Med. 2005;165:2595–2600.
25 Linzer M, Visser MR, Oort FJ, et al. Predicting and preventing physician burnout: Results from the United States and Netherlands. Am J Med. 2001;111:170–175.
26 Geurts S, Rutte C, Peeters M. Antecedents and consequences of work-home interference among medical residents. Soc Sci Med. 1999;48:1135–1148.
27 Colford JM Jr, McPhee SJ. The ravelled sleeve of care. Managing the stresses of residency training. JAMA. 1989;261:889–893.
28 Gabbe SG, Webb LE, Moore DE, Harrell FE Jr, Spickard A Jr, Powell R Jr. Burnout in medical school deans: An uncommon problem. Acad Med. 2008;83:476–482.
29 Hutter MM, Kellogg KC, Ferguson CM, Abbott WM, Warshaw AL. The impact of the 80-hour resident workweek on surgical residents and attending surgeons. Ann Surg. 2006;243:864–875.
30 Goiten L, Shanafelt TD, Wipf JE, Slatore CG, Back AL. The effects of work-hour limitations on resident well-being, patient care, and education in an internal medicine residency program. Arch Intern Med. 2005;165:2601–2606.
31 Pitt E, Rosenthal MM, Gay TL, Lewton E. Mental health services for residents: More important than ever. Acad Med. 2004;79:840–844.
32 McCray LW, Cronholm PF, Bogner HR, Gallo JJ, Neill RA. Resident physician burnout: Is there hope? Fam Med. 2008;40:626–632.
33 Seligman ME, Csikszentmihalyi M. Positive psychology. An introduction. Am Psychol. 2000;55:5–14.
34 Carpenter BN, Steffen PR. Stress. In: Haas LJ, ed. Handbook of Primary Care Psychology. New York, NY: Oxford University Press; 2004:563–577.
35 Thomas NK. Resident burnout. JAMA. 2004;292:2880–2889.
36 Haas LJ, DeGruy FV. Primary care, psychology, and primary care psychology. In: Haas LJ, ed. New York, NY: Oxford University Press; 2004:5–19.
37 Newton BW, Barber L, Clardy J, Cleveland E, O'Sullivan P. Is there hardening of the heart during medical school? Acad Med. 2008;83:244–249.