If effective, less intensive treatments for people with personality disorder have the potential to serve more people.To compare the clinical effectiveness and cost-effectiveness of psychoeducation with problem-solving (PEPS) therapy plus usual treatment against usual treatment alone in improving social problem-solving with adults with personality disorder.Multisite two-arm, parallel-group, pragmatic randomised controlled superiority trial.Community mental health services in three NHS trusts in England and Wales.Community-dwelling adults with any personality disorder recruited from community mental health services.Up to four individual sessions of psychoeducation, a collaborative dialogue about personality disorder, followed by 12 group sessions of problem-solving therapy to help participants learn a process for solving interpersonal problems.The primary outcome was measured by the Social Functioning Questionnaire (SFQ). Secondary outcomes were service use (general practitioner records), mood (measured via the Hospital Anxiety and Depression Scale) and client-specified three main problems rated by severity. We studied the mechanism of change using the Social Problem-Solving Inventory. Costs were identified using the Client Service Receipt Inventory and quality of life was identified by the European Quality of Life-5 Dimensions questionnaire. Research assistants blinded to treatment allocation collected follow-up information.There were 739 people referred for the trial and 444 were eligible. More adverse events in the PEPS arm led to a halt to recruitment after 306 people were randomised (90% of planned sample size); 154 participants received PEPS and 152 received usual treatment. The mean age was 38 years and 67% were women. Follow-up at 72 weeks after randomisation was completed for 62% of participants in the usual-treatment arm and 73% in the PEPS arm. Intention-to-treat analyses compared individuals as randomised, regardless of treatment received or availability of 72-week follow-up SFQ data. Median attendance at psychoeducation sessions was approximately 90% and for problem-solving sessions was approximately 50%. PEPS therapy plus usual treatment was no more effective than usual treatment alone for the primary outcome [adjusted difference in means for SFQ -0.73 points, 95% confidence interval (CI) -1.83 to 0.38 points; p = 0.19], any of the secondary outcomes or social problem-solving. Over the follow-up, PEPS costs were, on average, £182 less than for usual treatment. It also resulted in 0.0148 more quality-adjusted life-years. Neither difference was statistically significant. At the National Institute for Health and Care Excellence thresholds, the intervention had a 64% likelihood of being the more cost-effective option. More adverse events, mainly incidents of self-harm, occurred in the PEPS arm, but the difference was not significant (adjusted incidence rate ratio 1.24, 95% CI 0.93 to 1.64).There was possible bias in adverse event recording because of dependence on self-disclosure or reporting by the clini...