A total of 480 patients were treated in a large, multicenter randomized trial of a brief form of cognitive therapy, manual-assisted cognitive behavior therapy (MACT) versus treatment as usual (TAU) for recurrent deliberate self-harm. Each patient was randomized after a self-harm episode assessed at an accident and emergency center and followed up over 1 year. The main hypothesis tested was that those allocated to MACT would have a lower proportion of self-harm episodes in the succeeding year. A total of 60% of those allocated to MACT had face-to-face treatment and 430 (90%) of all patients had self-harm data recorded after 1 year. Although the results showed no significant difference between those repeating self-harm in the MACT group (39%) compared with the TAU group (46%) (P = 0.20), the treatment was cost effective (10% cheaper than TAU) and the frequency of self-harm episodes was fewer (50%) in the MACT group. A total of nine of 10 patients had some personality disturbance (42% of these with disorder), and for those where information on parasuicide events was collected, the proportion having a repeat episode ranged from 33% to 63% for different personality disorders. Those with BPD were most likely to repeat episodes quickly (mean 89 days for 25% to repeat) with dissocial personality disorder (equivalent mean 384 days) the slowest to repeat. Total costs were significantly greater in those with personality disorder and were reduced in those allocated to MACT; this saving was reversed in those with borderline disorder. On average, MACT appeared to increase the cost of those patients with BPD (BPD) and reduce the cost of those with other personality disorders. It is concluded that MACT has value in preventing self-harm cost effectively but this appears to be confined mainly to those who do not have BPD.