Lifetime and prospective associations among personality trait domains and suicide-related behaviors in patients with severe mental illness

Background: Despite extensive research and clinical efforts, the suicide rate in the United States continues to rise, driving the need for more research to identify latent factors that increase risk for suicide and to guide treatment decision-making. Methods: The current study examined a large cohort (N = 1,219) of high-risk psychiatric inpatients to explore associations between personality traits and suicide-related variables measured retrospectively (lifetime history prior to hospital admission) and prospectively (at discharge and 12-month follow-up). Results: Lifetime suicide-related behavior (SRB: combination of ideational severity, aborted, interrupted, actual attempts, and non-suicidal self-injury) was associated with age (younger), gender (female), and elevated scores on the Personality Inventory for DSM-5 (PID-5) negative affectivity, borderline trait composite score, and five-factor model traits of conscientiousness and neuroticism. Patients who manifested persistent suicidal ideation throughout a 6–8 week inpatient treatment (n = 162; 16.9%) tended to be younger, female, and to have elevated PID-5 borderline trait composite scores. Twelve-month post-discharge SRB was predicted by elevated PID-5 borderline trait composite scores. Limitations: Personality traits accounted for a small amount of variance in the overall model, thus limiting prediction based on individual traits. Conclusions: This large sample of high-risk inpatients with longitudinal outcomes provides a rare assessment of proximal personality traits in predicting lifetime SRB, persistent suicidal ideation observed during the course of a 6–8-week intensive inpatient treatment, and SRB outcomes within 12 months after discharge from hospitalization. Personality traits should be included in future attempts to create predictive algorithms that include relevant biological data (neuroimaging, genetic, microbiome).