Keith Hawton, Ellen Townsend, Ella Arensman, David Gunnell, Philip Hazell, Allan House, Kees van Heeringen

Deliberate self-harm is a major health problem associated with considerable risk of subsequent self-harm, including completed suicide. To identify and synthesise the findings from all randomised controlled trials that have examined the effectiveness of treatments of patients who have deliberately harmed themselves. Electronic databases screened: MEDLINE (from 1966-February 1999); PsycLit (from 1974-March 1999); Embase (from 1980-January 1999); The Cochrane Controlled Trials Register (CCTR) No.1 1999. Ten journals in the field of psychiatry and psychology were hand searched for the first version of this review. We have updated the hand search of three specialist journals in the field of suicidal research until the end of 1998. Reference lists of papers were checked and trialists contacted. All RCTs of psychosocial and/or psychopharmacological treatment versus standard or less intensive types of aftercare for patients who shortly before entering a study engaged in any type of deliberately initiated self-poisoning or self-injury, both of which are generally subsumed under the term deliberate self-harm. Data were extracted from the original reports independently by two reviewers. Studies were categorized according to type of treatment. The outcome measure used to assess the efficacy of treatment interventions for deliberate self-harm was the rate of repeated suicidal behaviour. We have been unable to examine other outcome measures as originally planned (e.g. compliance with treatment, depression, hopelessness, suicidal ideation/thoughts, change in problems/problem resolution). A total of 23 trials were identified in which repetition of deliberate self-harm was reported as an outcome variable. The trials were classified into 11 categories. The summary odds ratio indicated a trend towards reduced repetition of deliberate self-harm for problem-solving therapy compared with standard aftercare (0.70; 0.45 to 1.11) and for provision of an emergency contact card in addition to standard care compared with standard aftercare alone (0.45; 0.19 to 1.07). The summary odds ratio for trials of intensive aftercare plus outreach compared with standard aftercare was 0.83 (0.61 to 1.14), and for antidepressant treatment compared with placebo was 0.83 (0.47 to 1.48). The remainder of the comparisons were in single small trials. Significantly reduced rates of further self-harm were observed for depot flupenthixol vs. placebo in multiple repeaters (0.09; 0.02 to 0.50), and for dialectical behaviour therapy vs. standard aftercare (0.24; 0.06 to 0.93). There still remains considerable uncertainty about which forms of psychosocial and physical treatments of self-harm patients are most effective, inclusion of insufficient numbers of patients in trials being the main limiting factor. There is a need for larger trials of treatments associated with trends towards reduced rates of repetition of deliberate self-harm. The results of small single trials which have been associated with statistically significant reductions in repetition must be interpreted with caution and it is desirable that such trials are also replicated.