Abstract
We aimed to determine and compare the longitudinal predictive power of Diagnostic and Statistical Manual of Mental Disorders, fifth edition's (DSM-5) two models of personality disorder (PD) for multiple clinically relevant outcomes. A sample of 600 community-dwelling adults-half recruited by calling randomly selected phone numbers and screening-in for high-risk for personality pathology and half in treatment for mental health problems-completed an extensive battery of self-report and interview measures of personality pathology, clinical symptoms, and psychosocial functioning. Of these, 503 returned for retesting on the same measures an average of 8 months later. We used Time 1 interview data to assess DSM-5 personality pathology, both the Section-II PDs and the alternative (DSM-5) model of personality disorder's (AMPD) Criterion A (impairment) and Criterion B (adaptive-to-maladaptive-range trait domains and facets). We used these measures to predict 20 Time 2 functioning outcomes. Both PD models significantly predicted functioning-outcome variance, albeit modestly-averaging 12.6% and 17.9% (Section-II diagnoses and criterion counts, respectively) and 15.2% and 23.2% (AMPD domains and facets, respectively). Each model significantly augmented the other in hierarchical regressions, but the AMPD domains (6.30%) and facets (8.62%) predicted more incremental variance than the Section-II diagnoses (3.74%) and criterion counts (3.31%), respectively. Borderline PD accounted for just over half of Section II's predictive power, whereas the AMPD's predictive power was more evenly distributed across components. We note the predictive advantages of dimensional models and articulate the theoretical and clinical advantages of the AMPD's separation of personality functioning impairment from how this is manifested in personality traits. (PsycInfo Database Record (c) 2024 APA, all rights reserved).