Predicting domains and rates of change in borderline personality disorder.

What changes and how quickly these changes occur as a result of therapy in borderline personality disorder (BPD) is an important ongoing question. The features of BPD patients that are most predictive of rates of change in such patients remain largely unknown. Using the Cornell Personality Disorders Institute (CPDI) randomized controlled trial data, we sought to determine (a) the number and nature of broad domains underlying a large number of rate of change (slope) measures across many psychological, psychiatric, and psychosocial indexes, and (b) which baseline individual difference psychological features of the BPD patients correlated with these rate of change domains. We examined the latent structure of slope (rate of change) measures gleaned from individual growth curves for each subject, studied in multiwave perspective, on separate measures of anger, aggression, impulsivity, depression, global functioning, and social adjustment. Three broad domains of change rate could be discerned. These domains were reflected in factors that are described as (a) anger/aggression change ("aggressive dyscontrol"), (b) global functioning/social adjustment change ("social adjustment/self-acceptance"), and (c) anxiety/depression/impulsivity change ("conflict tolerance/behavioral control"). Factor scores were computed for each change domain and baseline measures of personality and psychodynamic features, selected a priori, were correlated with these factor scores. Multiple regression analyses revealed (a) baseline negative affectivity and aggression predicted the aggressive dyscontrol change domain, (b) baseline identity diffusion predicted the social adjustment/self-acceptance change domain, and (c) baseline social potency predicted the conflict tolerance/behavioral control change domain. These baseline predictors suggest potential research foci for understanding those aspects of BPD that change at comparable rates over time.