I read the paper by Rath et al with interest. The authors determined the burnout rate and associated factors among gynecologic oncologists using a cross-sectional study design. A total of 1086 subjects were invited, and 369 subjects completed the survey. The percentages of burnout, depression, history of suicidal ideation, alcohol abuse, and impaired quality of life were 32%, 33%, 13%, 15%, and 34%, respectively. Using univariate logistic regression analyses, 10 significant odds ratios of independent variables for burnout were recognized, including a female sex and an age ≤50 years. The authors concluded that burnout was associated significantly with psychosocial distress and lower levels of career satisfaction among gynecologic oncologists. I have some concerns regarding their study.

First, the authors used the Maslach Burnout Inventory (MBI) to measure burnout, but they defined burnout syndrome as a high score for either emotional exhaustion or depersonalization. Among 369 subjects, 32% were identified as exhibiting burnout in their study. The MBI has another subscale for “lack of personal accomplishment,” but the authors did not consider this scale in their logistic regression analysis. Guveli et al evaluated the relationships among burnout, job satisfaction, psychologic wellbeing, and stress coping among 159 oncology staff members in combination with their sociodemographic and occupational characteristics. They also used the MBI, and the prevalence of emotional exhaustion, depersonalization, and lack of personal accomplishment were 30.2%, 8.2%, and 44%, respectively. With the use of a logistic regression analysis, the adjusted odds ratio of depersonalization for emotional exhaustion was 1.7 (95% confidence interval, 1.3–2.2). In contrast, emotional exhaustion was not associated significantly with a lack of personal accomplishment or job satisfaction. Rath et al described the prevalence of a lack of personal accomplishment among gynecologic oncologists as 11%, although the prevalence of 2 other MBI subscales were the same as those reported by Guveli et al. Rath et al studied subjects with high academic ranks; 96.3% of their subjects declared their overall health as being excellent, very good, or good. I suspect that the low prevalence of a lack of personal accomplishment should be considered with caution.

Second, the authors did not use a multivariate logistic regression analysis, and an adjusted odds ratio was not calculated. Because >100 subjects reported burnout, >10 independent variables can be used in analyses to obtain stable estimates with the use of a logistic regression analysis. The prevalence of burnout and associated factors should be clarified by understanding the characteristics of the target subjects.