Multidomain correlates of burnout: A population-based study using supervised machine learning

Area of Science:

• Occupational Health
• Psychology
• Sociology

Background:

• Burnout is a significant occupational health concern with increasing societal impact.
• Understanding burnout risk factors is crucial for developing effective prevention strategies.

Purpose of the Study:

• To investigate the relationship between burnout and organizational factors, psychosocial employment conditions, sociodemographic variables, and mental health.
• To identify key predictors of burnout severity using advanced machine learning techniques.

Main Methods:

• Analysis of cross-sectional survey data from the German population-based DigiHero cohort (n=27,020).
• Utilized linear associations and XGBoost machine learning to explore multidomain variable relationships with burnout, measured by the Maslach Burnout Inventory.
• Interpreted machine learning results using SHAP values and compared with univariate statistics, adjusting for sample bias.

Main Results:

• Effort-reward imbalance, work-life interference, overcommitment, and poor general mental health (depression, anxiety symptoms) were the strongest burnout predictors.
• Psychosocial employment conditions and individual mental health were more influential than organizational and sociodemographic factors.
• Occupation, extended remote work, age, and income showed smaller but notable effects on burnout.

Conclusions:

• Psychosocial employment conditions and individual mental health are paramount in burnout etiology, overshadowing organizational and sociodemographic factors.
• Positive social work environments are suggested for burnout prevention, while pre-existing mental health issues heighten vulnerability.
• Prospective longitudinal studies are recommended for deeper causal insights into burnout development.