Personality improves prediction of the onset of common mental disorders

Common mental disorders (CMDs), including mood and anxiety disorders, are prevalent and place a substantial burden on individuals and societies. Identifying individuals at risk of first-onset CMDs is crucial for screening and early prevention efforts. Because personality traits are stable and conceptually close to psychopathology, they may outperform existing established risk factors such as demographic, environmental, and biomarker screens. Using data from the Estonian Biobank (N ≈ 77,000) linked with medical registry records over a 20-year period, we tested whether Big Five domains and their single-item "personality nuances" predict first-onset depressive, anxiety, stress/adjustment, somatoform and non-organic sleep disorders. Elastic-net logistic regression compared models based on demography, domains, nuances, and their combinations, assessed by area-under-the-curve (AUC) and cross-disorder predictions. Nuances gave the highest discrimination (mean AUC = 0.71), outperforming demographics (0.59) and domains (0.66); performance peaked for recurrent depressive disorder (AUC = 0.78). Neuroticism dominated domain-level prediction, followed by conscientiousness, while key nuances reflected low fun-seeking, poor relaxation, and intense emotions. Similar accuracy across diagnoses indicated that personality captures a largely transdiagnostic risk signal rather than disorder-specific liability. Brief item-level personality screens could therefore provide a simple, quick, low-cost tool for flagging high-risk individuals. These screens could guide preventive mental health interventions.