Effectiveness of Psychological Treatments for Borderline Personality Disorder and Predictors of Treatment Outcomes: A Multivariate Multilevel Meta-Analysis of Data from All Design Types

Abstract

We examined the effectiveness of psychotherapies for adult Borderline Personality Disorder (BPD) in a multilevel meta-analysis, including all trial types (PROSPERO ID: CRD42020111351). We tested several predictors, including trial- and outcome type (continuous or dichotomous), setting, BPD symptom domain and mean age. We included 87 studies (N = 5881) from searches between 2013 and 2019 in four databases. We controlled for differing treatment lengths and a logarithmic relationship between treatment duration and effectiveness. Sensitivity analyses were conducted by excluding outliers and by prioritizing total scale scores when both subscale and total scores were reported. Schema Therapy, Mentalization-Based Treatment and reduced Dialectical Behavior Therapy were associated with higher effect sizes than average, and treatment-as-usual with lower effect sizes. General severity and affective instability showed the strongest improvement, dissociation, anger, impulsivity and suicidality/self-injury the least. Treatment effectiveness decreased as the age of participants increased. Dichotomous outcomes were associated to larger effects, and analyses based on last observation carried forward to smaller effects. Compared to the average, the highest reductions were found for certain specialized psychotherapies. All BPD domains improved, though not equally. These findings have a high generalizability. However, causal conclusions cannot be drawn, although the design type did not influence the results.