Personality Markers for Common Mental Disorders: A Multi-Trait, Multi-Diagnosis, Multi-Rater Approach

Personality science has been slow to translate into clinical prediction, largely because research has focused on broad self-reported trait domains. Although these domains are established risk factors for common mental disorders (CMDs), including anxiety and mood disorders, they overlook more specific facet-level risk markers and limit the overall predictive power. Using data from the Estonian Biobank (N = 3,221; 1,918 females; ages 18–75), we linked baseline multi-informant NEO-PI-3 personality domains and facets with six future CMD diagnoses recorded in national health insurance data within 1–15 years after assessment (Mdn = 7 years), controlling for prior diagnostic history. While prior diagnosis was the strongest individual predictor (β = .32–.64), personality traits accounted for additional unique variance. Facet-level models explained 17–47% of variance across diagnoses and outperformed domain-level models (R²mean = .34 vs. .31; nearly doubling explanatory power when diagnostic history was excluded: R²mean = .15 vs. .08), exceeding well-established risk factors (R²max ~ .13) by up to 2.6-fold. We identified six transdiagnostic and three disorder-specific personality markers that were uniquely associated with clinical CMDs. Multi-informant design reduced single-method bias and enhanced prediction, particularly when prior diagnostic information was unavailable (facet models: R² = .08, .06, and .15 for self-, informant-, and multi-informant reports, respectively). This facet-level, multi-informant approach represents a comprehensive advance in personality-based clinical prediction, refining how lower-level personality facets map onto future real-world psychopathology and underscoring the untapped potential of personality-based risk assessment and fine-grained intervention targets in clinical practice.