AID-trial Assertive Intervention After Deliberate Self-harm

SPECIFIC AIMS/HYPOTHESES

The aim of this study is to investigate if assertive outreach incorporating hands-on guidance and motivational support of compliance with follow-up treatment after suicide attempts is able to reduce the frequency of non-fatal and fatal suicide acts.

BACKGROUND AND SIGNIFICANCE

Suicidal behavior is associated with considerable risk of subsequent self-harm, including completed suicide. Owens et al. (Owens, Horrocks, \& House 2002a) conducted a systematic review of fatal and non-fatal repetition of self-harm. Owens found that the median one-year repetition rate was 16 percent non-fatal and two percent fatal. Results from the epidemiological European Multicentre Study (WHO/EURO) also shows a repetition rate between 12 and 30 percent within the first year, making suicide attempts a very important high-risk factor (Cedereke, Monti, \& Ojehagen 2002;Schmidtke et al. 1996).

Unfortunately there are still very few randomized controlled trials with evidence based recommendations for suicide prevention. The studies conducted are often flawed due to selection bias or lack statistical power (Crawford et al. 2007;Hawton et al. 1998b;van der et al. 1997). A systematic review of randomised clinical trials of psychosocial and pharmacological interventions for persons who have attempted suicide revealed that until now, no intervention has proven effective in reducing suicide rate or repetition rate (Hawton et al. 1999). The authors identified non-significant trends toward reduced repetition of deliberate self-harm for problem-solving therapy compared with standard aftercare, for provision of an emergency contact card in addition to standard care compared with standard aftercare alone, and for intensive aftercare plus outreach compared with standard aftercare. In one relatively large study in this group, which evaluated community follow-up of patients who did not attend outpatient appointments, there was a near significant difference in repetition of deliberate self-harm of 10.7 percent compared with 17.4 percent (0.57; 0.32 to 1.02), (Van Heeringen et al. 1995).

Hawton et al. concluded that the results of the systematic review indicate that there is currently insufficient evidence on which to base firm recommendations about the most effective forms of treatment for patients who have recently deliberately harmed themselves. Given the size of the problem, they consider this situation serious (Hawton, Townsend, Arensman, Gunnell, Hazell, House, \& Van Heeringen 1999). The main problem with nearly all trials in the meta-analysis was that they included far too few subjects to have the statistical power to detect clinically meaningful differences in rates of repetition of deliberate self-harm between experimental and control treatments, if such differences existed. Even when the results from similar trials were synthesized with meta-analytical techniques, there were insufficient numbers of patients to detect such differences.

After the Cochrane review, six randomised clinical trials have been published, among them the POPMATC study, which has a large sample size of 480 patients. The study examines the effect of up to five plus two booster sessions of manual-assisted cognitive-behavioral-therapy sessions versus various control conditions, some of which bear some resemblance to cognitive therapy (problem solving). In the study, there was no significant difference in repetition of self-harm between experimental group and control group (Tyrer et al. 2003). In a randomised clinical trial, Brown et al. found that cognitive therapy (mean 9 sessions) for adults who attempted suicide (N=120) reduced repetition rate and increased the repetition-free period (Brown et al. 2005). Recently, Vaiva et al. published the results of a large French multi-centre trial in which telephone contact after one and three months were compared to standard treatment (Vaiva et al. 2006). Comparison of both experimental groups combined versus standard treatment did not show that telephone contact was superior to standard treatment, but the authors point out that there was a significant difference between the proportion that repeated suicide attempt in the one-month telephone call group compared to control treatment in the first 6 months of the trial. The odds ratios in all trials point in the same direction, namely toward a protective effect of the experimental conditions.

As pointed out by Hawton, it might be necessary for such conditions to be combined with assertive outreach.

RESEARCH METHODS

METHOD, DESIGN AND PROCEDURE

The AID-study is a randomized, clinical trial. The patients included will be centrally randomized to either standard treatment (n = 120) or intervention treatment (N = 120), representing 6 - 8 assertive outreach contacts after suicide attempts or deliberate self-harm. The randomization is computerized and will, in prevention of allocation bias, be conducted by an independent research assistant in the research unit. The intervention team therefore will not know the block size of the study groups. The randomization is stratified for three items:

1. First-evers or previous suicide attempts,
2. Previous psychiatric contact or no contact of the kind,
3. Alcohol or no alcohol involved.

Stratification variables will reported when calling the research assistant in charge of the randomization. The research assistant in turn will provide the intervention team with the assigned treatment and participant number. The intervention team is not blinded. The patients will be recruited from the emergency room, intensive care, psychiatric emergency room as well as those admitted to psychiatric wards. After discharge they will be contacted by mail or phone at home. The experimental intervention treatment will go on for the first six months followed by six months of no contact. One year from the inclusion date both patients in standard treatment and intervention treatment will receive a final call focusing on on-going compliance with follow-up treatment. Sociodemographic data, medical status, information concerning abuse, previous psychiatric treatment and previous suicide attempts will be obtained at baseline.

INTERVENTION - ASSERTIVE OUTREACH

Colleagues in Norway have developed a preventive intervention focusing on the 'gap' between admission/emergency room and outpatient treatment (Dieserud, Loeb, \& Ekeberg 2000). The Bærum-model is named after the suburb north of Oslo where the intervention first took place. The concept is to support and guide the person shortly after the in-hospital treatment for self-injury through a recommended follow-up or after treatment based on assertive principles. Poor compliance with treatment is addressed using active, caring individual support to assist clients with feelings of being abandoned, depressed, shame, and ambivalent (Dieserud, Loeb, \& Ekeberg 2000). The Bærum-model has never previously been as researched in a randomized controlled study (Dieserud, Loeb, \& Ekeberg 2001). Knowing that the precipitating factors for suicide are diverse, the prevention strategies have to be likewise. The Bærum-model is a psychosocial approach to suicide prevention that underlines the need for intervention tailored according to individual client needs. The intervention is an indicated prevention strategy targeting people with suicide attempts and deliberate self-harm as a high-risk group. They will be offered 6-8 assertive outreach contacts. The outreach contacts will be home visits focusing on providing support and motivating patients to comply with follow-up treatment. However, depending on individual needs, the contacts may include or be substituted with admittance to alcohol units, somatic/oncological treatment, counselling with a family doctor, family meetings and so on. The main target of the intervention is to encourage and ensure that the patient will be referred and accompanied to the necessary treatment and actually receive the recommended treatment.

STANDARD TREATMENT

Standard treatment consists of referral to a range of different treatment modalities depending on the diagnosis and clinical and social condition of the patient. In standard treatment there is no procedure for ensuring that the patient will actually receive the recommended treatment. Patients are often referred to available treatment modalities such as general practitioner, psychological treatment, treatment for alcohol abuse, and most often, the patients are themselves responsible for getting into contact with the treatment to which they are referred.

OUTCOME

The primary outcome measure is the number of suicides assessed by the Danish death-register and the rate of repeated suicide attempts/deliberate self-harm registered by the collaborating wards and units in their routine procedure of treating people applying for help in relation to suicide behaviour. (Standardized hospital registration procedure). Where it is not clear if the act was a suicide attempt medical an external evaluation committee will review records.

STATISTICAL POWER

Alfa is assessed to 0.05 and beta to 0.2, which gives a power of 80 percent (power = 100- beta). According to international literature(Hawton et al. 1998a;Owens, Horrocks, \& House 2002b) the expected repetition rate (primary outcome) will be 30 percent in the standard group and 15 percent in the intervention group at one year. Calculated that requires 120 people in each randomization treatment group(Dupont \& Plummer, Jr. 1990), which makes 240 people included all together. Primary outcome based on register data and medical records dropout does not have to be accounted for. The data collection period will be limited to 2 years. Patients will be drawn from two University Hospitals in Copenhagen and based on routine registration; it is likely that we will be able to recruit 240 patients in a two-year period.

STATISTICS AND DATA ANALYSIS

Statistical survival analysis will be used for investigating the differences in repeated fatal and non-fatal acts of self harm between the standard group and the intervention group. Cox-regression analysis will be used to investigate differences between the two groups in the frequency of suicidal behaviour (suicide, suicide attempts and deliberate self-harm). Analysis will be adjusted for confounding factors such as gender, age, and treatment with antidepressants. The survival rates for the two randomization groups will be illustrated by a Kaplan-Meyer Plot. Chi-square-test, Mann-Whitney and T-test will be used in the analysis of compliance intensity in both standard and intervention groups.