Predictive modeling of burnout dimensions based on basic socio-economic determinants in health service managers and support personnel in a resource-limited health center.

Background: Burnout is a prevalent condition in the healthcare sector, and although it has been extensively studied among healthcare professionals, less is known about its impact on non-professional workers, particularly in low-resource settings. This study aimed to test a preliminary predictive model based on basic socioeconomic and sociodemographic determinants to predict symptoms of burnout among support personnel and health services managers in a resource-limited health center. Methods: A prospective cross-sectional study was conducted. Using simple random sampling, symptoms of burnout were surveyed among health service managers and support personnel using the Maslach Burnout Inventory (MBI). Statistical analyses included correlation tests and predictive models using random forest models to identify significant associations and cast predictions. Results: A total of 76 participants were included. Of these, 34.21% exhibited high levels of emotional exhaustion (EE), 42.11% showed elevated depersonalization (DP), and 7.89% reported low personal accomplishment (PA). Significant negative correlations were observed between household income and the EE and DP dimensions. The predictive models demonstrated acceptable performance in identifying socioeconomic factors associated with burnout, with prediction errors ranging from 7.68% to 20.31%. Conclusions: Burnout is common among support personnel and health services managers in resource-limited settings, particularly among those with lower incomes. The findings underscore the importance of implementing policies that address both working conditions and economic well-being to mitigate the risk of burnout. More robust predictive models could serve as a valuable tool for early identification and prevention of burnout in this type of setting.