Eighteen studies were included: 4 RCTs (n=384), 4 non-randomised trials (n=230), 7 before-and-after studies (n=220) and 3 case studies (n=6). The sample sizes given represent baseline and do not reflect the reported high rates of loss to follow-up.
Overall, the methodological quality and the quality of treatment fidelity assessment in the included studies were poor. Completeness of follow-up was inadequate or unclear in 12 out of 15 studies, excluding the case studies.
Nursing only interventions (5 studies).
One case study reported a positive change in social functioning and symptoms, whereas an RCT, a non-randomised trial, a before-and-after and another case study provided insufficient data to determine an effect on social functioning.
Mixed discipline interventions (13 studies).
Although most of the mixed discipline intervention studies showed positive changes in social functioning, symptoms and service use, overall the results were inconsistent. Two non-randomised and 2 before-and-after studies that reported personality disorder status showed significant improvements.
Group interventions (5 studies).
The results from group intervention studies that reported change in social functioning and symptoms were inconsistent. Service use was reported in one RCT that found no change. Improvement in personality disorder status was reported in one before-and-after study.
Interventions with individuals (2 studies).
A non-randomised study showed improvement in personality disorder status. A before-and-after study reported positive changes in social functioning, service use, symptoms and personality disorder.
Group and individual mixed interventions (5 studies).
An RCT, a non-randomised study and a before-and-after study reported change in social functioning and symptoms in favour of the intervention. The non-randomised study also reported personality disorder status, showing a change in favour of the intervention. A reduction in service use was reported in the RCT and in one non-randomised study.
Cognitive-behavioural interventions (3 studies).
Two case studies reported a positive change in social functioning but the one before-and-after study did not. The before-and-after study showed no change in symptoms, whereas one case study reported an improvement.
Nursing management interventions (3 studies).
These studies provided insufficient data to determine any effects.