Machine learning analysis of the association between psychosocial risks and burnout syndrome in mining workers

Ramirez J.K.; Candia-Puma M.A.; Arratia-Corrales M.A.; Peña-Pinto Y.M.; Corzo Salas de Valdivia A.; Gutiérrez-Araníbar R.J.C.; Llahuilla-Quea J.A.; García-Torres I.; Chuquimbalqui-Maslucán R.; López-Alvarado R.L.; Tapia-Manrique E.R.; Yuli-Posadas R.A. · Frontiers in Public Health (2026) · DOI: 10.3389/fpubh.2026.1871339

Resumen

Introduction – Burnout syndrome represents a critical issue in occupational health, particularly in high-demand contexts such as mining, where physical, environmental, and psychosocial risks converge and affect workers' wellbeing and job performance. In this context, the study objective is to analyze the association between psychosocial risk factors and burnout syndrome among mining workers in Moquegua, Peru. Methods – A quantitative, analytical cross-sectional study with a non-experimental design was conducted. The study population consisted of 65 workers from a mining unit in Moquegua, Peru. Given the complete accessibility of the target population, a census approach was adopted, and all eligible workers were included in the study (N = 65). Validated instruments were used, including the SUSESO/ISTAS21 questionnaire for psychosocial risks and the Maslach Burnout Inventory. Data analysis involved descriptive and correlational statistics, multiple linear regression, and machine learning techniques. Results – The findings revealed significant associations between psychosocial factors particularly social support, leadership, and work–family conflict (double presence) and burnout. All analyzed factors demonstrated significant associations capacity, with double presence emerging as the most influential predictor. Furthermore, machine learning analyses identified relevant burnout-related patterns within the analyzed dataset, highlighting their effectiveness in identifying burnout-related patterns. Discussion – Burnout in mining is a multifactorial phenomenon influenced by organizational and psychosocial conditions. The results support the use of machine learning as an useful tool for identifying psychosocial risk patterns that may support prevention strategies, contributing to improved occupational health strategies in high-risk industrial settings. Copyright

Instituciones: Universidad Nacional Mayor de San Marcos; Universidad Nacional Daniel Alcides Carrión; Universidad Católica de Santa María