High Incidence of Burnout in Academic Chairpersons of Anesthesiology: Should We Be Taking Better Care of Our Leaders?

• Burnout among academic chairs has been studied in other specialties and, when present, diminishes leadership

• In a survey of 102 chairs of anesthesiology departments, more than half met the criteria for high or moderate burnout

• Risk factors for burnout were low job satisfaction and reduced self-reported spousal/significant other support

Academic anesthesiology chairs in the United States contend with job attributes that can be frustrating and eventually lead to emotional exhaustion and burnout. Examples of these attributes include reporting conflicting relationships, responsibility for things over which they have no control, and human resource challenges. To understand the emotional impact of these frustrations, it is important to define precisely the vocabulary that describes the results of workplace frustration and unhappiness. Burnout is a work-related psychologic syndrome characterized by emotional exhaustion, low personal accomplishment, and depersonalization.1Emotional exhaustion is the subjective sense of fatigue or stupor related to one's work. Low personal accomplishment is a feeling of frustration with work-related achievements. Depersonalization is a person's attempt to separate himself or herself from his or her work as a defense mechanism. Burnout syndrome differs from depression because it is specific to the work environment, whereas depression extends to both one's professional and personal life.

Burnout syndrome was first characterized in the early 1970s.1Clinical manifestations are often nonspecific and include fatigue, sleep and eating disorders, headache, and emotional instability. The validated instrument most commonly used to study burnout is the MBI-HHS Burnout Inventory (MBI),1which uses a composite score that takes into consideration the three subscales based on questions relating to emotional exhaustion, depersonalization, and reduced personal accomplishment. Burnout syndrome is considered present when the responder demonstrates high scores in emotional exhaustion and depersonalization and a low score in personal accomplishment.

The prevalence of burnout is higher among individuals whose job involves interactions with people (e.g. , physicians, nurses, and social workers).2Evolving changes in health care, including decrease in physician services reimbursement, challenges with the Accreditation Council for Graduate Medical Education, and difficulties with faculty retention, generate stress for the department chair that can potentially predispose him or her to develop burnout.3Burnout has been studied in academic chairs of other specialties.3–5Gabbe et al.  3concluded that the psychologic well-being of the chairs of academic departments of obstetrics and gynecology affected the quality of leadership they provided in teaching, patient care, and administration.

The purpose of this study was to evaluate work-related stress and personal factors associated with professional burnout in chairs of anesthesiology departments. We designed a cross-sectional survey that was adapted from those used in the previously mentioned studies of other medical subspecialties3–5to evaluate whether the trends observed in other departments would also apply to chairpersons of anesthesiology departments.

The present study was approved by the Northwestern University Institutional Review Board, Chicago, Illinois. A cross-sectional nationwide survey was sent to 117 chairs of academic anesthesiology departments in the United States ( appendix). The mailing list was developed from the 2009–2010 directory of the American Medical Association section of Graduate Medical Education. The initial search yielded 132 programs. Seven programs did not have a chairperson when the survey was conducted, and eight programs did not provide the electronic address of the chairperson. The survey was created using software (Survey Monkey; SurveyMonkey Inc., Portland, OR). To ensure confidentiality of the participants, the survey was set up to delink the responses to the respondents' e-mail addresses. The participants who did not respond to the electronic questionnaire were mailed a copy of the survey with a self-addressed return envelope addressed to the primary investigator.

The questionnaire was divided into five parts. Open-ended and multiple choice questions were used. Likert scales were used to quantify respondents' level of agreement with a statement. The first section consisted of 11 questions designed to capture demographic information about the chairperson: age, sex, time of service as a chair, size of department by number of faculty, division chiefs, residents and fellows, amount of work hours per week, percentage of time dedicated to patient care, administrative duties, research, and whether his or her medical school had a support group for chairs. The second part of the survey required the chairperson to select from among 15 potential stressors those that have affected the department; also, one question assessed the degree to which these factors affected the chairs. They were asked to rank the factors on a five-point scale from “not at all” to “extreme amount.” Current job satisfaction was assessed, as were job satisfaction perceived 1 and 5 yr prior (if appropriate) and likelihood the chair would resign in the next 1–2 yr (ranging from “not likely” to “extremely likely” using a five-point scale). Chairs were also asked to rate their job satisfaction in their position at 1 and 5 yr before the survey but were not expected to answer these questions if they had not been in the position for that length of time. The chairpersons were then questioned regarding satisfaction relating to the balance of personal and professional life using a five-point scale, ranging from “very satisfied” to “very dissatisfied.” The third portion of the survey assessed the chairpersons' opinion regarding their professional life using a modified self-efficacy scale.6The sum of these responses was calculated, ranging between a low of 7 and a maximum of 35. Respondents also ranked their effectiveness as a chair on a scale ranging from 0 (representing “least effective”) to 100 (representing “most effective”).

The fourth part of the survey included 12 questions from the MBI–Human Services Survey (HSS).1–7The full MBI–HSS involves 22 questions: 5 assessing depersonalization, 9 assessing emotional exhaustion, and 8 assessing personal accomplishment. A score is given to each part of the MBI–HSS based on a frequency scale of 0 (“never”) to 6 (“every day”). The questionnaire evaluates depersonalization with statements such as “I feel I have become more callous toward people,” emotional exhaustion with statements such as “I feel emotionally drained from my work” and “I feel used up at the end of the workday,” and personal accomplishment with statements such as “I feel I am positively influencing people's life through my work.” The MBI-HHS survey was shortened to 12 questions to facilitate comparison with other specialties that were evaluated for burnout in their academic chairs.3–5The 12 questions selected were identified by Gabbe et al.  3using factor analysis of the subscales in the original MBI-HHS questionnaire. These questions included three evaluating depersonalization, five examining emotional exhaustion, and four assessing personal accomplishment. From the subscale values, the original MBI–HSS was calculated using proportional scoring. The ranges of the subgroup scoring are as follows: emotional exhaustion, 0–16 (low), 17–26 (moderate), and >26 (high); depersonalization, 0–6 (low), 7–12 (moderate), and >12 (high); and personal accomplishment, 0–31 (high), 32–38 (moderate), and >39 (low). A high risk of burnout was considered present when the respondent scored high in both emotional exhaustion and depersonalization and low in personal accomplishment. A moderate risk was considered when two or more of the previously mentioned criteria were met.

The fifth part of the survey included six questions that evaluated the chair's support from his or her spouse/significant other and family. These questions were adapted from previous investigations3from the marital support questionnaires developed by Spanier,8Pearlin and Schooler,9Penkower et al. ,10and Phelan et al.  11A five-point scale, ranging from “never” to “always,” was applied to questions as follows: “How often do you disagree with your spouse/significant other or other family members about the amount of time you spend on work?” and “How often does your spouse/significant other encourage you to take advantage of professional opportunities?” A score ranging from a low of 6 (indicating minimal support) to a maximum of 30 (indicating considerable support) was calculated based on the sum of the individual responses.

Characteristics of the respondents by sex, length of time as chair, and hours worked were compared using the Fisher exact test. Current self-evaluation of the level of satisfaction was compared with responses regarding satisfaction and 5 yr prior using the sign test. The associations of the likelihood of stepping down with characteristics of the chairs were estimated using a 10,000-sample bootstrap and the Spearρrgr;. Respondents whose scores indicated a high risk of burnout on the MBI-HHS scale were compared with those with a low to moderate risk using the Fisher exact test statistic or the Mann–Whitney U test. MBI-HHS subscale scores among the risk of burnout-indexed groups were compared using the Kruskal–Wallis H test and the Mann–Whitney U test with Bonferroni correction. Estimates of exact P  values were determined for the Sign, Kruskal–Wallis H, and Mann–Whitney tests using a Monte Carlo method with 10,000 samples and confidence limits of 99%. Factors associated with a high risk for burnout (P < 0.1) were entered into a binary logistic regression model. The model was fitted using stepwise backward elimination with removal testing (P > 0.1) based on the probability of the likelihood-ratio statistic. Confidence intervals (95%) for the variables in the model were estimated using a 10,000-sample bootstrap. The overall predictive value of the model was assessed as the area under the receiver operating characteristic curve of burnout risk predicted by the model versus  that predicted by the MBI-HHS scale. Sensitivity, specificity, and positive likelihood of a positive test result were calculated using a standard formula. Missing data were handled listwise for grouped comparisons and pairwise for assessment of current and prior satisfaction. P < 0.05 was required to reject the null hypothesis. Nominal and ordinal are presented as counts and percentages of respondents. Interval data are presented as medians with interquartile ranges (IQRs). All reported P  values are two-tailed. Data were analyzed using computer software (NCSS 2007 version 7.1.20, release date February 19, 2010 [NCSS LLC, Kaysville, UT]; and PASW Statistics 18.0.2, release date April 2, 2010 [SPSS Inc, Chicago, IL]).

A total of 102 chairpersons responded to the survey, 67 via  the electronic version and 35 via  the mail, corresponding to a response rate of 87%. Nine surveys had insufficient responses for calculation of the burnout index. There were 43 data elements used for analysis on each survey and a total of 4,386 data elements in the 102 returned questionnaires. In total, 4,096 data points were obtained (93%). Eighty-eight surveys had all possible responses entered. Nine respondents completed between 10 and 36 elements but did not complete the MBI–HSS section; therefore, the MBI-HHS score could not be computed. Three surveys that contained an MBI–HSS section had one or more elements missing for inclusion in the logistic regression model.

The median (IRQ) age for the chairs (n = 98) was 55 (52–61) yr, and 51 of 101 respondents had been chair for less than 5 yr. Of the respondents, 82 (82%) were men and 18 (18%) were women. Of 98 responding chairs, 63 (64%) reported that they worked more than 60 h/wk, with 14 (78%) of female chairs reporting a work week of more than 60 h compared with 49 (61%) of male chairs (P =0.28). Duration on the job did not have an impact on the number of hours worked; 32 (67%) of 48 chairs with more than 5 yr in their current position reported working more than 60 h/wk versus  32 (63%) of 51 chairs with less than 5 yr in their position as chair (P = 0.83). Of 100 respondents, 82 (82%) reported spending less than 40% of their time involved with patient care. Administrative duties accounted for more than 40% of the chairs' time in 81 of 100 respondents; 83 (83%) of 100 spent less than 20% of their time in research-related activities. Of 101 respondents, 51 (50%) reported that their department included more than 50 faculty members and 31 (31%) had more than 60 faculty members. Of 101 chairs, 31 (31%) oversaw residency programs with more than 60 residents. Most of the medical schools did not have a support group for chairpersons (79 of 98 respondents).

Responses to issues that occurred in the year before the survey that created stress for the responding chairs are shown in table 1. Stressful problems that affected many chairs were faculty retention and department finances. Of 94 chairpersons, 32 (34%) noted that these issues had affected them only to a slight degree, 34 (36%) reported being moderately affected, and 28 (30%) reported being largely to extremely affected. Nevertheless, 32 (34%) of 93 respondents reported current high job satisfaction; however, this represented a significant decline compared with that reported for 1 yr (P = 0.009) and 5 yr (P = 0.001) before the survey (fig. 1). Regarding the balance between personal and professional life, 41 (44%) of 94 respondents reported dissatisfaction to high dissatisfaction with balance, but only 12 (13%) of 94 reported this same level of dissatisfaction with their salary. The median (IQR) composite score of the modified efficacy scale of 21 (18–24) (n = 93) corresponds to a feeling of a moderate level of control by the chairpersons over their professional life, and the median (IQR) self-assessment of effectiveness of 85 (75–90) (n = 90) suggests that the chairpersons view their impact in a predominantly favorable manner. Of 93 chairs, 43 (46%) reported that it is moderate to extremely likely that they will step down as a chair within 1–2 yr, with 26 (28%) suggesting that the chance of stepping down was very or extremely likely. There was no association of the age (P = 0.16) or sex (P = 0.82) of the respondents and their anticipated likelihood of stepping down or their rating of their effectiveness (P = 0.63); however, respondents who reported a higher likelihood of stepping down scored lower on the modified efficacy scaρêgr; =−0.303, P = 0.003) than those reporting a low likelihood of stepping down.

The distribution of respondents at risk for burnout based on the MBI-HHS criteria is shown in figure 2. Of 93 anesthesiology chairs, 26 (28%) met the criteria for high burnout, with an additional 29 (31%) in the moderate to high burnout category. Median (IQR) subscale scores for emotional exhaustion, personal accomplishment, and depersonalization were 32 (21–41), 36 (24–44), and 10 (7–16), respectively. The breakdown of the MBI-HHS subscale scores for the levels of the burnout index is shown in figure 3. Emotional exhaustion scores were more likely to be increased in chairs with a moderate-risk index of burnout, and depersonalization scores increased as the risk reached the moderately high level; scores for personal accomplishment remained consistent until the respondents reached the high-risk category.

A comparison of respondent characteristics in chairpersons with a high risk of burnout compared with those with a low to moderate risk is shown in table 2. Age, sex, time as a chair, time worked weekly, and perceived effectiveness did not differ between chairs in the high-risk compared with the lower-risk categories. Interestingly, chairs of larger departments appear to be at lower risk of burnout compared with chairs with medium (range, 40–50) size faculty. High-risk chairpersons reported a greater likelihood of stepping down within 2 yr, demonstrated lower personal efficacy scores, had low current job satisfaction, and were more affected by stressors facing the department. Faculty retention and departmental budgetary issues were more frequently rated as high to extremely high stressors by chairpersons who were at high risk for burnout.

The median (IRQ) score for support from a spouse/significant other was 20 (17–24) and was not different between male and female chairpersons. Support scores were lower in the high risk of burnout chairpersons compared with those at low to moderate risk of burnout (table 2). Chairpersons reporting a high likelihood of stepping down in the next 2 yr also reported lower median (IQR) scores of 17 (14–22) for spousal support compared with those unlikely to step down (21 [18–25]) (P = 0.006). Of 26 respondents, 17 (65%) in the high risk of burnout group reported their spouse/significant other was not understanding of extra hours worked compared with 5 (8%) of 65 in the low- to moderate-risk group (P < 0.005), although the median number of hours per week that each of these groups worked was not different.

Multivariate analysis identified decreased current job satisfaction and low spousal/significant other support as independent predictors of a high burnout risk. The risk ratios (95% confidence intervals) for high burnout for respondents with moderate (10.9 [2.6–84.1]) (P = 0.001) and high (8.2 [2.6–79.3]) (P = 0.008) dissatisfaction were greater than those with high satisfaction. Chairpersons with a support index of less than 20 had a 5.2 (1.6–27.9) times greater likelihood of scoring in the high burnout category (P = 0.007). The area under the receiver operating characteristics curve for predicted high risk of burnout and actual risk was 0.80. The sensitivity and specificity (95% confidence intervals) of the model for predicting high burnout risk were 72% (52–86%) and 88% (78–94%), respectively. The positive likelihood ratio (95% confidence interval) for a predicted high risk of burnout from the model was 6.0 (3.1–12.9).

The important finding of this study is the high incidence of risk of burnout (28%) among academic chairs of anesthesiology departments in the United States. When considering the chairs who are at risk for developing the syndrome (defined by a MBI-HHS score of moderate–high burnout), the percentage reaches 59%. Chairs of academic anesthesiology departments constantly deal with challenges of the ever-evolving healthcare environment, including providing high-quality clinical services in the face of decreasing reimbursements, nurturing research programs with limited career funding12and high competition and underperformance by faculty in National Institutes of Health grant submissions and awards,13and ensuring accreditation of educational programs while adapting to changing requirements of the Accreditation Council for Graduate Medical Education and the American Board of Anesthesiologists. In addition, chairs are frequently involved in strategic planning at the institutional and/or hospital level and substantial university committee service that may limit the amount of time that they are able to devote to their own departmental issues. Finally, all serious disciplinary action must run through the chair (e.g. , questions of clinical competence and appropriate interpersonal interactions), which can be extremely time-consuming and frequently include difficult conversations. Among the stress factors assessed that were identified in chairs with a high risk of burnout, departmental budgetary issues and faculty retention concerns were most frequent.

Many chairs of academic anesthesiology departments exhibited at least one component of the MBI-HHS scoring system, with 69% reporting high emotional exhaustion, 60% reporting high depersonalization, and 39% reporting low personal accomplishment. Chairs with moderate burnout risk generally reported increased emotional exhaustion and increased depersonalization. Personal accomplishment generally remained strong until subjects reached the highest-risk category, suggesting that stress at that stage had reached a point at which individuals began to question their value to the department and to themselves. A comparison of the MBI-HHS subscales with those reported among chairpersons of other academic specialties is shown in table 3. Based on our findings overall, anesthesiology chairs exhibit a higher rate of burnout compared with chairs of obstetrics/gynecology,3otolaryngology,4and ophthalmology.5The degree of depersonalization and emotional exhaustion was higher in anesthesiology chairs than in chairs of these departments.

Physicians who cultivate their personal and professional well-being are less likely to develop burnout or will at least diminish its impact on their lives.14The development of well-being should be stimulated throughout one's career, always being careful to minimize the delayed gratification mechanism used so frequently by physicians.15McCue and Sachs16demonstrated that resident physicians who have learned stress management techniques decreased their subscale score on depersonalization and emotional exhaustion. Another method of better preparing faculty for the stresses related to the chairperson's position is through mentorship. According to Bates and Blackhurst,17mentors have an important role in guiding new chairs through administrative challenges and in introducing them to other leaders. The presence of support in the form of a mentor(s) might have a greater impact on younger chairs who are new to the demands of the position. Although changes on an individual level might be part of the answer,18the fact that an organization recognizes the potential for burnout as a problem decreases the chance that individuals will blame themselves or the recipients of their care as a cause of job-related stress. Support groups have also been suggested as a potential method to reduce physician burnout.19,20In our sample, only 19% of medical schools had a support group for chairs, yet the rate of burnout among chairs in institutions with support groups was 13% lower than among those without support groups. However, our study was not designed to assess the effect of support groups on the risk of burnout.

Burnout can have significant health implications for the individuals who are affected by the syndrome. McCall21suggested that substance abuse is more common among healthcare workers affected by burnout. There is also growing evidence that burnout might substantially increase the risk of cardiovascular disease due to sympathetic system activation, sleep disturbances, immune function compromise, and poor health behaviors.22The high incidence of burnout among anesthesiology chairs calls for preventive measures and early interventional modalities of treatment. Unfortunately, therapeutic interventions to decrease burnout have not been well studied. The fact that 87% of the academic anesthesiology chairs in the United States responded to this survey also suggests that the chairs as a group perceive that there is a problem.

Twenty-six chairs stated that they were very to extremely likely to step down within 1–2 yr. Age, sex, and perceived effectiveness did not appear to be predictive of this response because none of these variables was different between chairs who believed that they were likely to step down compared with chairs who believed that they were unlikely to step down. Departmental chairs are an extremely valuable resource to universities, and turnover can cause emotional distress to faculty members and can have significant financial implications for the institutions.23 

We did not find a sex difference in the frequency of burnout in male and female chairs of anesthesiology, although the few female chairs may have limited our ability to detect an effect. In a study of 5,704 male and female physicians in primary and specialty nonsurgical care, McMurray et al.  24found that women have a 1.6 times higher rate of burnout than their male counterparts. This risk was increased by 12–15% as the work hours per week increased from 40 h to 45 h. More recently, this same group compared the burnout risk between US physicians and those in the Netherlands and found that there was a sex difference in burnout risk in the United States but not in the Netherlands, where the difference in work hours was more similar between men and women.25Our study did confirm the findings of previously mentioned researchers25who determined that the odds of burnout were 40% less when workers have a high amount of support from spouses or significant others. In our study, we found that there is a 15% less chance of experiencing burnout or being at high risk of burnout in chairs who had greater support from family members. Linn et al.  26showed that among academic internists, those who did not have spousal/significant other support were more depressed and dissatisfied with their work than those who did receive support. Chairs of anesthesiology should make their families aware that they value their support and they should create an environment that allows them to have more control over their time because both factors may be protective against burnout.

This study has several limitations. The surveys were self-reported and might not represent actual behaviors. The questionnaires were not completed in a controlled setting. We also did not use the full 22 questions of the MBI–HSS; instead, we used the same 12 questions as Gabbe et al.  3because those questions were shown to have a better correlation with the burnout subcomponents. Thus, we reduced the survey burden to the participants. There were missing data elements, and the reason for the missing data was not apparent. The most frequently omitted section included the questions related to spousal support (only 91 respondents completed the questions). In addition, the results of our study may not be generalizable beyond the United States because of differences in healthcare practices and demands placed on academic chairs in other countries. We did survey 15 academic chairs from countries other than the United States and found that most agreed that burnout was an important issue in their country, despite low rates of faculty and chair turnover (table 4).

In conclusion, we report a high incidence and risk of burnout among anesthesiology chairs. We found that age, sex, time as a chair, time worked, and perceived effectiveness were not associated with high burnout, but low job satisfaction and reduced spousal/significant other support significantly increased the risk. Because of the high financial and emotional cost to institutions and because of the paramount role these leaders have on shaping the future of anesthesiology, academic institutions and professional societies should be encouraged to develop strategies and perform studies evaluating methods of reducing burnout in chairs.

The authors would like to thank the following individuals for reviewing the data in this article and commenting on their thoughts regarding burnout in academic anesthesiology chairs in their country: A. Barry Baker, M.B.B.S., D.Phil., F.A.N.Z.C.A., F.J.F.I.C.M., Emeritus Professor, University of Sydney, Executive Director of Professional Affairs, Australian and New Zealand College of Anaesthestists, Sydney, Australia; Neville M. Gibbs, M.B.B.S., M.D., F.A.N. Z.C.A., Clinical Professor, School of Medicine and Pharmacology, University of Western Australia, Head of Department of Anaesthesia, Sir Charles Gaidner Hospital, Nedlands, Australia; José Reinaldo Cerqueira Braz, M.D., Ph.D., Professor of Anesthesiology, School of Medicine, UNESP, Botucatu, SP, Brazil; Davy Cheng, M.D., M.Sc., F.R.C.P.C., F.C.A.H.S., Distinguished University Professor, Chair/Chief, Department of Anesthesia and Perioperative Medicine, London Health Sciences Centre and St. Joseph Health Care, University of Western Ontario, London, Ontario, Canada; Homer Yang, M.D., C.C.F.P., F.R.C.P.C., Professor and Chair and Anesthesia-in-Chief, Department of Anesthesiology, The Ottawa Hospital, University of Ottawa, Ottawa, Canada; Lars S. Rasmussen, M.D., Ph.D., Professor of Anesthesiology, Centre of Head and Orthopaedics, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark; Hans-Joachim Priebe, M.D., F.R.C.A., F.F.A.R.S.C.I., Professor of Anesthesia, Department of Anesthesia, University Hospital Freiburg, Freiburg, Germany; Tiberiu Ezri, M.D., Professor of Anesthesiology, Department of Anesthesiology, Wolfson Medical Center, Holon (affiliated with Sackler Faculty of Medicine, Tel Aviv University), Israel; Kazuyoski Hirota, M.D., F.R.C.A., Professor and Chairman, Department of Anesthesiology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan; Masahiko Kawaguchi, M.D., Professor of Anesthesiology, Department of Anesthesiology, Nara Medical Center, Nara, Japan; Michel MRF Struys, M.D., Ph.D., Professor and Chair, Department of Anesthesiology, University Medical Center Groningen, Groningen, the Netherlands; Alex T. Sia, M.MED., Head and Senior Consultant in Anaesthesia, Associate Professor of Anesthesiology, Obstetrics and Gynaecology, KK Women's and Children's Hospital, Singapore; James Murray, M.D., F.R.C.A., F.C.A.R.C.S.I., Consultant Anesthetist, Senior Lecturer, Clinical Skills Education Centre, Belfast, Northern Ireland; Jennifer M. Hunter, M.B., Professor of Anaesthesia, Department of Anaesthesia, University Clinical Department, University of Liverpool, Liverpool, United Kingdom; and John Kinsella, M.B., B.S., M.D., F.R.C.A., Head of Section of Anaesthesia, Pain and Critical Care Medicine, University of Glasgow, Glasgow Royal Infirmary, Glasgow, United Kingdom.