LAUSR

Background Mental health services lack a strong evidence base on the most effective interventions to reduce compulsory admissions. However, some research suggests a positive impact of crisis-planning interventions in which patients are involved in planning for their future care during a mental health crisis. Aims This review aimed to synthesise randomised controlled trial (RCT) evidence on the effectiveness of crisis-planning interventions (for example advance statements and joint crisis plans) in reducing rates of compulsory hospital admissions for people with psychotic illness or bipolar disorder, compared with usual care (PROSPERO registration number: CRD42018084808). Method Six online databases were searched in October 2018. The primary outcome was compulsory psychiatric admissions and secondary outcomes included other psychiatric admissions, therapeutic alliance, perceived coercion and cost-effectiveness. Bias was assessed using the Cochrane collaboration tool. Results The search identified 1428 studies and 5 RCTs were eligible. One study had high risk of bias because of incomplete primary outcome data. Random-effects meta-analysis showed a 25% reduction in compulsory admissions for those receiving crisis-planning interventions compared with usual care (risk ratio 0.75, 95% CI 0.61–0.93, P = 0.008; from five studies). There was no statistical evidence that the intervention reduced the risk of voluntary or combined voluntary and compulsory psychiatric admissions. Few studies assessed other secondary outcomes. Conclusions Our meta-analysis suggests that crisis-planning interventions substantially reduce the risk of compulsory admissions among individuals with psychotic illness or bipolar disorder. Despite common components, interventions varied in their content and intensity across the trials. The optimal models and implementation of these interventions require further investigation. Declaration of interest E.M., S.L., S.J. and B.L.-E. received funding from the National Institute for Health Research during the conduct of the study.