To the Editor:
We read with interest the recent article on burnout among anesthesiology residents by Sun et al.1 Burnout within anesthesiology is of growing concern and quite rightly so; unaddressed burnout can lead to suboptimal patient care and clinical practice, mental health issues, and long-term physical disease in clinicians.2 Although an important issue to tackle, the accurate estimation of burnout in large-scale surveys is difficult and poses a significant challenge to academics. Some of these reasons are discussed below and should be considered when interpreting the prevalence of burnout reported in the literature.
Burnout is poorly characterized. Burnout is classified as an “occupational phenomenon” by the World Health Organization (Geneva, Switzerland) and not a medical condition. As such, a diagnostic criterion does not exist. The 11th revision of the International Classification of Disease (ICD-11) characterizes burnout by the presence of (1) feelings of energy depletion or exhaustion, (2) increased mental distance from one’s job or feelings of negativism or cynicism related to one’s job, and (3) reduced professional efficacy. This description reflects a shift in the field to recognize broader areas of “cynicism” and “professional efficacy” as part of the dimensions of “depersonalization” and a perceived sense of “lack of personal achievement” respectively.2 Nonetheless, these dimensions vary over time and exist on a scale of varying severity, not as dichotomous variables. That said, how does one measure such dimensions reliably and then decide universally what and when is it problematic? Such fundamental questions are part of ongoing debates because moderate or severe symptoms can be present in clinicians who are not burnt out. These uncertainties reflect our limited understanding of the syndrome.
Multiple tools detect burnout, but they can be very inaccurate. The Maslach Burnout Inventory and its variations (e.g., the abbreviated Maslach Burnout Inventory, the Oldenburg Burnout Inventory, and the Copenhagen Burnout Inventory) are some examples of burnout detection tools. Not all tools assess every dimension of burnout, and none are recommended by the World Health Organization. The Maslach Burnout Inventory for humans services is a 22-question survey which is by far the most commonly used and validated tool in clinicians. It assesses all dimensions of burnout but can be time-consuming to complete and costly to administer, especially over several time points. Therefore, academics may find that abbreviated versions can improve response rates in large population studies. One such version, the 12-question survey used by Sun et al.,1 is increasingly used within anesthesiology. However, it has been recently shown to have a poor positive predictive value which can lead to the overestimation of burnout prevalence.3
Many criteria for defining burnout and the severity of its symptoms exist, even for the same tool. Because different detection tools assess different dimensions of burnout, the readouts of these tools are expectedly heterogeneous and not always appropriate to compare. Even if comparisons are limited to studies using the Maslach Burnout Inventory, different cut-off values for symptom severity and burnout criteria have been used in literature. These can also vary significantly between the complete Maslach Burnout Inventory and its abbreviated versions. For example, Sun et al.1 reported that burnout prevalence in U.S. anesthesiology residents was 51%, but de Oliveira et al.4 reported that it was 41% using the same tool but different cut-offs of symptom severity. Without doubt, the lack of a universal standard accounts for significant heterogeneity in burnout prevalence reported by systematic reviews and meta-analyses.5,6
Considering the above, caution is needed when interpreting burnout prevalence reported in literature. In our experience, we have found that screening for burnout in our cohort of anesthesiology residents was best done without the sole reliance on detection tools because the false-positive rates were high (62.1% detected burnout vs. 22.4% actual burnout). This would have unnecessarily strained resources, prolonged training times, and negatively impacted service provision. Being a developed country that uses English as its main language, and having an anesthesiology program that was modeled closely after the United States and accredited by the Accreditation Council for the Graduate Medical Education-International, we believe our experience has relevance to U.S. anesthesiology residencies. Through the use of the full Maslach Burnout Inventory and its abbreviated version, both of which have been validated in non-U.S. and U.S. populations, burnout symptoms in Singapore and U.S. anesthesiology residents were found to be similar.3 However, we have determined clinically that actual burnout prevalence was low and corresponded more closely to burnout rates reported in United Kingdom anesthesiology trainees and intensive care staff (approximately 25%).7,8
Although further research is needed in syndrome characterization and the development of more accurate screening tools, we are of the opinion that greater awareness and trainee self-reporting of burnout is a more practical way forward. Through the education of trainees and trainers on its signs and symptoms, trainees who feel burnt out can be offered a confidential platform to self-report without negative implications. Further validation and clinical correlation through the use of detection tools and multi-source data could then validate findings and improve the accuracy of detection. In so doing, finite support services could then be channeled selectively and efficiently to trainees who require support the most. Screening large populations en masse through voluntary surveys is inefficient for providing intervention because it only captures data from responders and further efforts would then be needed to distinguish the true positives from the false positives.
In summary, the sole use and reliance of detection tools may limit the accurate detection of burnout including its prevalence. Better screening tools are needed, and clinical correlation is advised.
The authors declare no competing interests.