6  What interventions work?: Selected case studies

The available evidence on effectiveness and cost-effectiveness is mixed. It is strong in some areas, where a number of methodologically sound systematic reviews or meta-analyses have been undertaken. These are often based on evidence from the USA, where the number and variety of interventions permits comparisons. In other areas the analysis is not robust and does not provide clear evidence of intervention effectiveness.

A number of general interventions focus on early childhood and have beneficial effects much later in life. There is evidence that some intensive early interventions, involving both the child and family, prevent later adverse outcomes from teenage pregnancy to youth offending. However, targeted interventions, at this early stage in particular, run a risk of Type I and Type II errors.

Health interventions cover the spectrum from before birth to young adulthood and include universal and targeted components. Child health services provide an opportunity for early identification of conditions and risk factors that put children at risk of adverse outcomes and for interventions to minimise their adverse consequences. More specific interventions are designed to reduce childhood accidents and youth suicide. The available evidence on youth suicide, a major concern in New Zealand, suggests that there is no single intervention that has been identified as effective, let alone cost-effective.

Education interventions are designed primarily to improve educational achievement. Because cognitive ability and education are factors that protect against other adverse outcomes, such as criminal behaviour, they have beneficial indirect effects on well-being.

Behavioural interventions are those that are specifically designed to reduce risk-taking behaviour among adolescents and young people in order to prevent adverse outcomes. The evidence suggests that the graduated drivers’ licences are effective in reducing road fatalities among young drivers, and this is reflected in the New Zealand statistics. The evidence on the prevention of teenage pregnancy is much less clear, although it suggests that both sexual and non-sexual factors should be targeted, for example by improving access to, and knowledge of, contraception and improving economic opportunities for young women.

The remainder of this section summarises the available evidence on specific interventions to address selected adverse outcomes in New Zealand: youth suicide, teenage pregnancy, educational underachievement and youth inactivity.

6.1  Youth suicide

The causes of youth suicidal behaviour, like many other youth behaviours of concern, are complicated. Such behaviour arises through the dynamic interplay of individual, social and environmental factors. There seem to be two main models of suicide risk that imply different intervention strategies (Hider 1998). One model considers that risk is largely confined to young people with recognisable mental disorders, based on the strong association between mental disorder and suicidal behaviour. This model underpins the design of interventions based on the strengthening of health services and the targeting of individuals with psychiatric morbidity.

An alternative view is that suicide is a response to overwhelming and untenable life stresses that could happen to any adolescent and that mental illness is not the most important variable. This model implies that population-based interventions are needed to help young people to cope with stress and that, in particular, there is a need for more employment opportunities for young people. However, increasing scientific understanding of suicide, mental illness and substance abuse disorders indicates that suicide does not result from stress alone (US Surgeon General 1999b).

A review of studies on risk factors for suicides suggests that suicide is the endpoint of adverse life experiences in which multiple risk factors combine (Beautrais 2000). Young people at increased risk of attempting suicide often come from educationally and socially disadvantaged backgrounds; often have disturbed or unhappy family and childhood backgrounds; typically display mental health problems and immediately before the suicide attempt they may face some severe stress or life crisis (Beautrais 1998). The mental health problems most commonly associated with suicide are depression, substance abuse and behavioural difficulties.

The CHDS identified that mental health disorders, especially depression, were the most important predictors of suicide attempts. Adverse life events, and being raised in a family with low socio-economic status, were also associated with increased rates of suicide attempts, even when other confounding and intervening factors were taken into account (Fergusson et al 2000). Such research provides some support for both the main models of youth suicide discussed above.

Protective factors against suicide include individual characteristics, such as a person’s neurobiological makeup, attitudes or behaviours and skills in areas such as problem solving and conflict resolution. They also include family factors such as warm and supportive relationships, high but realistic expectations on the part of parents, clear and reasonable limits and religious and cultural beliefs that discourage suicide. School and community influences can also confer resilience, including the development of warm and supportive relationships, conventional peers and involvement in school activities.

The risk factors that lead to suicide (especially mental and substance abuse disorders) and the protective factors that confer resilience and safeguard against it can provide the basis for the design of appropriate and effective interventions (White 1998). Since one of the strongest indicators of suicidal behaviour is a mental health disorder (including substance abuse and antisocial behaviour) effective policies focus on the identification, prevention, treatment and management of these disorders (Beautrais 1998). Policies that provide support to high-risk families in which children might be at risk of developing a range of adjustment problems may reduce the number of children exposed to adverse family environments and inter alia reduce the risk suicide.

The meta-analysis undertaken in New Zealand identifies a number of limitations in the published literature that examined the effectiveness of interventions (Hider 1998). In particular, many studies do not exclude chance, bias, or the possibility of alternative explanations for their findings or rely on expert opinion.

The study considers the prevention of suicidal behaviour in terms of universal interventions applied to young people in general and designed to prevent the development of suicidal behaviour (ie, before the fact prevention); and targeted interventions applied to individuals established at being at high risk of suicide and aimed at preventing suicidal behaviour (ie, after the fact intervention) (see Table 9). The study uses a structured approach to critically appraising the literature, grading the level of evidence from I (randomised control trials and meta-analyses) to V (opinions of respected authorities or reports of expert committees).

No single intervention was been found to prevent suicide in well-conducted randomised control trials. Programmes that targeted at-risk groups seem most promising in reducing suicidal behaviour among young people. There is some uncertainty about the effectiveness of school-based programmes that are popular in the USA.

Restricting the means of suicide is also a promising intervention. Barriers at jumping sites can be effective (Beautrais 2001). Reducing the lethal emissions in car exhausts appears to have some potential (Gunnell and Frankel 1994). The potential for reductions in the availability of means of committing suicide is confirmed by an early report of an in-depth review (Fonagy 2000). Marked reductions in suicide attempts were found, not specifically in young people, from restricting handguns in the USA and domestic gas in the UK.

The main focus of the meta-analysis was on suicide prevention by primary health practitioners. The study reveals little evidence on the effectiveness of psychological or psychosocial treatments or pharmacological therapies. However, group support and cognitive behaviour therapy are probably effective in reducing suicidal behaviour. The evidence on family therapy, crisis intervention and psychoanalysis is less promising.

There is a dearth of research on the effectiveness of pharmacological treatments for reducing suicidal thoughts and behaviours in young people. Most of the research has been conducted on adults (US Surgeon General 1999a). While medication (using selective serotonin re-uptake inhibitors) appears to be effective in treating mental illness, it is less effective in preventing suicidal behaviour. Office-based interventions by GPs, and education programmes to help them recognise and treat mental illness, appeared to be effective.

Key: (Note: the effectiveness and the level of evidence need to be read together)

Probable = Reasonable evidence from at least one good quality study or established consensus from several studies

Possible = Evidence from study/s that may have methodological limitations

Unlikely = Results from several currently available studies that generally suggest that the intervention is ineffective, although further research is needed to enable a definitive conclusion

I = evidence obtained from at least one meta-analysis

II-1 = evidence obtained from at least one randomised controlled trial

II-2 = evidence obtained from at least one controlled trial without randomisation

III-1 = evidence obtained from at least one cohort analytical study, preferably based in more than one centre or research group

III-2 = evidence obtained from at least one case control analytical study, preferably based in more than one centre or research group

IV = evidence obtained from at least one study that has used a primarily descriptive study design such as a cross sectional, ecological or time series methodology

V = opinions of respected authorities based on clinical experience, or reports of expert committees.

Source: Hider (1998) Table 38

Postvention refers to interventions after the suicide of family members or close associates and is in part designed to prevent imitative suicide. Very little research is available to assess these strategies, but experts seem to agree that it is a potentially valuable means of preventing further suicides.

It is possible that using a range of interventions is the most effective way of addressing the challenge of reducing suicidal behaviours and suicides. Prevention and treatment are not competing strategies. Rather, they complement one another, offering both broad-based strategies designed to reduce the risk of suicidal behaviours and individually focused treatments for individuals identified as being at risk of suicide.