Employment in Veterans With Mental Illness: An Exploratory Penalized-Regression and Random-Forest Analysis

ImportanceReturning to civilian and community life after military service represents a critical and often challenging transition. Employment can contribute to social identity, financial security, social participation and community reintegration, and overall quality of life among veterans.ObjectiveTo examine candidate factors of employment among veterans who reported a current mental health diagnosis.DesignCross-sectional survey study.SettingOnline data collection.ParticipantsA total of 156 veterans who reported having a current mental illness were recruited; 151 with complete data (mean age = 37.85 years) were analyzed. Most participants identified as male (67%), White (70%), and married or living with a partner (62%).Outcomes and MeasuresEmployment status was the outcome, and candidate predictors included demographic and military characteristics, mental health, physical health and functioning, and psychosocial and behavioral factors.ResultsThe BIC-LASSO model revealed that greater disability severity and pain intensity were each associated with lower odds of employment, with moderate bootstrap selection recurrence (57.1% for disability severity and 56.9% for pain intensity) that remained robust after sensitivity analysis. The random forest model did not exceed majority-class accuracy.Conclusions and RelevanceGreater disability severity and greater pain intensity emerged as plausible, though not yet confirmed, correlates of employment status among veterans with mental illness. Because these exploratory findings require replication in larger, more diverse, and longitudinal samples, they do not yet justify changes to practice; in the interim, they support attending to everyday functioning and pain, alongside psychiatric symptoms, when occupational therapy practitioners evaluate work participation with this population.