We thank Drs. Ong, Lim, and Ong1 for their interest in our publication2 and appreciate the opportunity to discuss burnout, an issue that is relevant well beyond anesthesiology residents. Questions are raised about the poor characterization of burnout, inaccuracy of burnout assessment tools, and the lack of diagnostic criteria to identify burnout. We agree that burnout is a complex issue and that estimates of its prevalence should be interpreted in appropriate consideration of the context.
As stated by Ong et al.,1 burnout is classified as an “occupational phenomenon” by the World Health Organization (Geneva, Switzerland). The manifestation of burnout, as a psychologic syndrome, may depend on personal characteristics, working environment, and even social, political, and economic factors.3 Both theoretical models and empirical evidence have guided the characterization of burnout, and qualitative work from social, clinical, and industrial-organizational psychologists has identified different dimensions of burnout, including exhaustion, cynicism, and inefficacy.3,4 These dimensions are reasonably captured by how the International Classification of Diseases, 11th edition, characterizes burnout.
Ong et al.1 correctly indicate that there are various proposed scales of burnout based on different conceptualizations. For example, the Maslach Burnout Inventory– Human Services Survey assesses emotional exhaustion, personal accomplishment, and depersonalization, and the Maslach Burnout Inventory–General Survey assesses exhaustion, cynicism, and professional efficacy.5 Other measures focus on exhaustion alone, including subtypes. The Shirom-Melamed Burnout Measure, for example, distinguishes among physical fatigue, emotional exhaustion, and cognitive weariness.6 The existence of multiple proposed scales and the availability of thousands of peer-reviewed articles related to burnout speak to the importance—and difficulty—of burnout assessment.
We recognize that richness of information is lost when continuous scores of burnout dimensions must be translated into a dichotomous classification of burnout, and the sensitivity and specificity levels associated with cutoff criteria may not always be provided to inform the reader. Studies of psychologic and somatic symptoms of burnout and associated biomarkers might be helpful in searching for the optimal cutoff criteria.7 We concur with the recommendation to assess the degree of burnout on a continuous scale,5 and join the call for establishing consistent cutoff criteria when a classification is deemed necessary, especially for the same measurement tool.
Ong et al.1 provide examples of discrepancies in the estimates of burnout prevalence, specifically with the use of the abbreviated Maslach Burnout Inventory. They note that de Oliveira et al.8 estimated 41% of anesthesiology residents to be at high risk for burnout in 2013, and our study reported an estimate of 51% among anesthesiology residents and first-year residency graduates from 2013 to 2016.2 We suspect that different compositions of subgroups and the timing of the studies contributed to the difference in the estimates, although both demonstrate alarmingly high rates. Lim et al.9 reported strikingly different estimates of burnout prevalence among the same group of anesthesiology residents in Singapore when different cutoff criteria were applied—22.4% based on Maslach Burnout Inventory–Human Services Survey and 62.1% based on its abbreviated version. Had the same Maslach-recommended criteria been applied, however, the prevalence of burnout in Lim et al.’s study would be estimated at 20.7% based on the abbreviated Maslach Burnout Inventory, which would be close to the 22.4% identified based on the full scale.9 In addition, the correlation coefficients for the three subscales ranged from 0.92 to 0.96 between the two versions. We argue that Lim et al.’s study actually provides some assurance that the abbreviated version offers a reasonable alternative for brevity. Regarding the prevalence of 51% of burnout among U.S. anesthesiology residents2 versus 22% among Singapore anesthesiology residents,9 the limited generalizability of conclusions in the latter study due to small sample size (N = 58) and imbalance of males (N = 17) and females (N = 41) must be recognized. We also suggest that there are a multitude of sociocultural factors that might impact burnout beyond language and training system.
In summary, we concur with Ong et al. that burnout could be better defined, more precisely characterized and measured, and compared with more consistency. Nonetheless, we also want to acknowledge that progress in burnout characterization and assessment has been made since it was first described in the 1970s, and we welcome a continuation of the discussion about its relevance to anesthesiologists.
Drs. Sun and Zhou are staff members of the American Board of Anesthesiology (ABA); Drs. Culley, Keegan, Macario, and Warner are ABA Directors and receive a stipend for their participation in ABA activities.