An assertive outreach intervention does not reduce repeat suicide attempts compared with usual care
Question: Is an assertive outreach intervention better than usual care for reducing subsequent suicide attempts in a group of people who have attempted suicide?
Patients: In total, 243 people (76% female, mean age 31) older than 12 years admitted in 14 days following suicide attempt to six regional hospitals in the catchment area of Copenhagen. Exclusion criteria: patients diagnosed with schizophrenia spectrum disorders, severe depression, severe bipolar disorder and severe dementia; those receiving outreach services from social services or living in institutions; patients admitted to a psychiatric ward more than 14 days after the initial suicide attempt.
Setting: Outpatient department, Copenhagen University Hospital, Denmark; time period not stated.
Intervention: Assertive intervention for deliberate self-harm (AID) or usual care. The AID intervention consisted of 8–20 outreach consultations delivered by psychiatric nurses over a 6-month period. The intervention included case management with crisis intervention and flexible problem solving. Motivational support was provided and patients were actively assisted to and from scheduled appointments to improve compliance. Usual care consisted of referral to a range of different treatment modalities dependent on the diagnosis and clinical and social circumstances of the patient. As part of usual care, patients from both groups not receiving other treatments and not abusing substances were offered 6–8 sessions with a psychologist. Pharmacological support was continued or prescribed as appropriate.
Outcomes: Repeated suicide attempt and death by suicide (identified by self-report via telephonic interview, medical records and death register). Suicide attempt was defined as per the WHO definition.
Patient follow-up: Seventy per cent provided self-report data, 100% had register data and were included in intention to treat analysis (multiple imputation was used for missing values).
Design: Randomised controlled trial.
Blinding: Single blind (assessors of medical records blinded; assessors of self-report of subsequent suicide attempt not blinded).
Follow-up period: One year.
The AID intervention did not reduce the risk of a subsequent suicide attempt identified by registered hospital records compared with usual care at 12-month follow-up (AR: 16% with AID vs 11% with usual care; OR 1.60, 95% CI 0.76–3.38; p=0.22). In a completer analysis, there was also no significant difference between AID and usual care in self-reported new suicide attempts (12% with AID vs 18% with usual care; OR 0.61, 95% CI 0.26–1.46; p=0.27). If multiple imputation was used to estimate missing data there was still no significant difference between AID and usual care (AR: 12% with AID vs 19% with usual care; OR 1.69, 95% CI 0.34–1.43; p=0.32). Seven patients (four in the intervention group) were identified with a subsequent suicide attempt in the hospital record despite the patients not reporting this and nine patients self-reported a suicide attempt which was not listed in hospital records. One patient in the intervention group died by suicide within 2 weeks of the initial attempt.
An assertive outreach intervention delivered to individuals following a suicide attempt did not reduce subsequent suicide attempts when compared with usual care.
Sources of funding: The Danish Ministry of Health and Internal Affairs, the Danish National Board of Social Services TrygFonden and Aase og Ejnar Dainielsens Foundation, Denmark.
Self-harm in young people generates more anxiety in those working with them than any other psychological problem.1 Several promising interventions following self-harm have been tried, ranging from postcard interventions to individualised outpatient therapy.2 Morthorst and collegues describe the latter appoach as assertive intervention for deliberate self-harm (AID) and it is similar to the treatment offered by a few specialised self-harm services operating in the UK.3 This study corroborates previous findings showing little benefit of targeted interventions over treatment as usual (TAU).2 Yet this null finding may be explained by a combination of modest statistical power and patient heterogeneity.
While cognisant of the dangers of ‘data dredging’, the subgroup analysis hints at some positive findings; adults (aged 20–39 years old) were the least, and adolescents the most, responsive to AID. This implies that AID might be more effective among younger patients, or alternately that older patients are more resistant to treatment. While subgroup analysis can provide clues to better target treatment, the nature of adolescent and adult self-harm is likely to be substantively different. Another complication is the eclectic nature of AID, which makes isolating the ‘active’ therapeutic component (problem solving, crisis-management, etc) difficult.
Routine clinical care alone may be effective, and may be considerably better than no care. NICE guidelines acknowledge the inadequate research base for interventions, but suggest good engagement with, and offering support for a young person with various problems in their lives is the best approach. What is now needed are more focused research studies on those young people not responding to ‘TAU’, in order to identify which young people at first presentation to medical services are likely to have poor outcomes, and evidence for what sorts of targeted interventions will work for this vulnerable group.