DISCUSSION
Burnout prevalence and workplace factors that influence burnout were measured among clinicians who provide pediatric care in a hematology-oncology setting. Burnout levels were substantial across all 4 role-based cohorts in all 3 MBI dimensions, similar to previous studies.10,11 To contextualize part of our results, respondents from all 4 cohorts reported feeling emotionally exhausted or experiencing reduced personal accomplishment from their work an average of twice per month. Depersonalization (e.g., treating a patient as an immaterial object) rates were substantially lower compared to emotional exhaustion and reduced personal accomplishment, which is encouraging as this construct represents a lack of presenteeism and may reflect a more severe expression of burnout.26
Among workplace factors associated with burnout, the NASA-TLX dimension, “frustration”, was the most significant contributor across all 4 cohorts and 3 MBI dimensions. Temporal demand was also associated with burnout among MDs and nurses, but to much less a degree than frustration. Similarly, PSE involvement was significantly associated with burnout only among nurses.
Our post-hoc review of open-response comments elucidates the strong associations we observed between burnout rates and frustration levels. Comments addressing time constraints parallel our temporal demand results, and perceptions of feeling rushed or having inadequate time to complete tasks effectively are known predictors of burnout.28-30 Respondents expressed perceptions of insufficient administrator support in terms of feeling unable to garner managerial assistance when experiencing confrontations with colleagues, families, or patients. They also noted a general disconnect between the leadership’s perspective and the realistic needs of front-line staff. Insufficient institutional support for self-care reflected a desire for time away from work after experiencing patient loss, including both an immediate reprieve from clinical duties and paid time off to grieve and recover. Finally, perceptions of inadequate staffing and frequent turnover were identified. Inadequate staffing can result in increased individual task demands (a commonly cited predictor of burnout),31 and restrict employees’ ability to utilize paid time off. Turnover, especially voluntary turnover, can affect unit morale.32,33 When team members quit, remaining members may face a paucity of expertise and need to mentor new colleagues. Hence, inadequate staffing and turnover may create and perpetuate burnout in those perceiving a lack of adequate resources as an increase in individual demands is required to accomplish the necessary tasks.
PSE involvement can be detrimental to a clinician’s psychological well-being decades after the event, and there are unfortunate examples of clinician suicide after medical errors.34 Effects can mirror post-traumatic stress disorder and include lingering negative perceptions of professional self-efficacy.35 Our study is one of the first to identify a connection between involvement in PSEs and burnout among inpatient and ambulatory nurses in a pediatric hematology-oncology setting. Given the known consequences of burnout that include job performance deficits and potential turnover, our study provides further evidence for the development and implementation of resources to preserve the psychological well-being of staff involved in medical errors and serious safety events.
One strength of our study was using RWA, which addresses concerns raised when researchers use unique variance accounted for in regression models to rank predictors by significance.36 The principal challenge is that predictor variables are often intercorrelated, which leads to flawed interpretations of traditional metrics to partition unique predictor variance. We treated the 3 MBI dimensions as separate outcomes and used RWA to determine unique contributions of 6 predictors (NASA-TLX dimensions and PSE involvement).
As for limitations, St. Jude has a unique care model treating only a specialized population of patients. Our results may not generalize well to other pediatric hospitals caring for children with varied acute and chronic issues or to smaller pediatric oncology settings. Cross-sectional designs have many limitations, including causality, directionality, and biases originating from the study sample (e.g., responder bias, recall bias, and selection bias). Lastly, the MBI is the most widely used burnout survey instrument but does not include clinically validated cutoff scores. Publishers of the MBI do not advocate using results to determine burnout rates or support cutoff scores for low, moderate, and high levels of MBI dimensions. Researchers have developed their own MBI cutoff scores, but they vary widely.37 Lack of standardized MBI cutoff scores limits the ability to compare burnout rates across studies, which hinders its capability in expanding knowledge on causes, effects, and protective resources to reduce and prevent burnout.