Machine learning model distinguishes levels of psychological resilience in health care workers with 75% accuracy

A study published in Discover Artificial Intelligence has found that machine learning models can accurately predict levels of psychosocial resilience in healthcare workers, with 75.6% accuracy. Researchers from the University of Chicago trained logistic regression, random forest, and support vector machine models using data from the How Right Now Mental Health & Coping dataset.

The models classified workers into high and low resilience categories based on four psychometric variables—resilience, ability to bounce back, control, and confidence. Stress, depression, and anxiety emerged as the strongest predictors of low resilience, while coping strategies such as seeking social support, engaging in hobbies, prayer, and meditation were associated with higher resilience.

However, the data represent a single point in time, are based on self-reported responses from a U.S. population, and may not be generalizable to other healthcare systems. The authors suggest validating the model with Latin American healthcare workers and incorporating additional measures such as sleep data or physiological indicators.

They also propose that such interpretable machine learning models could be used by occupational health departments to identify at-risk individuals at an early stage, without relying on complex artificial intelligence systems.