Adolescents who self-harm are at increased risk of health and social problems as young adults

ABSTRACT FROM: Mars B, Heron J, Crane C, et al Clinical and social outcomes of adolescent self-harm: population based birth cohort study. BMJ 2014;349:g5954.

What is already known on the topic

Approximately 1 in 10 young people report having engaged in self-harm. Self-harm is the strongest clinical predictor of death by suicide and the behaviour causes great concern among family members, friends, teachers and clinicians.¹ Recently, in Diagnostic and Statistical Manual Fifth Edition (DSM 5), a distinction has been made between ‘non-suicidal self-injury’ (NSSI) and attempted suicide. Yet little is known about whether this distinction is clinically meaningful or informative.² In particular, it is unclear whether adolescent NSSI and self-harm with suicidal intent, are associated with distinctive patterns of health outcomes.

Methods of the study

The authors used data from the Avon Longitudinal Study of Parents and Children (ALSPAC) to investigate the early adult outcomes of adolescent self-harm with and without suicidal intent. ALSPAC is a large population-based birth cohort of over 14 000 children born in 1991–1992. The children have been intensively followed-up over two decades, with detailed assessments of a wide range of domains, including their mental health, use of substances and educational and occupational outcomes. When the children were aged 16 years, 4799 respondents (51% of potential participants) completed a self-report questionnaire about the lifetime occurrence of self-harm. Additional standard questions were used to identify those who had self-harmed with suicidal intent. When the participants were aged 18 years, they completed self-administered computerised assessments for depression, anxiety and substance use. At 21, they completed a further questionnaire enquiring about the occurrence of self-harm in the past year. Educational outcomes and data on employment were obtained from linked national data or self report. The authors dealt with the problem of missing data, using multiple imputation, before conducting regression analyses to examine the independence of associations between adolescent self-harm with and without suicidal intent and each of the young adult outcomes.

What does this paper adds

• This is the first population-based longitudinal study to investigate the outcomes of adolescent self-harm with and without suicidal intent.

• Compared with those who had never self-harmed, adolescents who engaged in ‘non-suicidal self-injury’ (NSSI) had a greater than twofold increase in the odds of depression and anxiety at age 18 years. Suicidal self-harm was associated with a fivefold increase in the odds of both depression and anxiety at age 18 years.

• Adolescents who self-harmed were also more likely to experience substance use problems at age 18 years. Compared with those who had never self-harmed, those engaging in NSSI experienced a greater than twofold increase in the odds of problematic cannabis use at age 18 years. Suicidal self-harm was associated with a greater than sixfold increase in problematic cannabis use at age 18 years. The odds of harmful alcohol use at age 18 years was more than doubled for both non-suicidal and suicidal self-harm.

• Self-harm with suicidal intent was associated with a doubling in the odds of not being in education, employment or training at age 19 years

Limitations

• Fifty-one per cent of those who were sent the self-harm questionnaire at age 16 years completed it. More importantly, those who returned the questionnaire differed from non-respondents on a range of characteristics. While the response rate is fairly typical for population-based studies, some of the associations in the complete case analysis may have been biased by non-random response. Notwithstanding, the researchers used a careful statistical approach (multiple imputation) which should have corrected for this possibility.

• Data on self-harm and health outcomes were all based on self-report. While this approach is commonplace in the field of self-harm research, it remains possible that some of the associations reflect response bias. The assessment of self-harm took place around the same time that the General Certificate of Secondary Education qualifications were obtained (an examination set especially for secondary-school pupils of about age 16 in England, Wales and Northern Ireland), thus reverse causality for these educational associations cannot be ruled out.

What next in research?

This study sheds important light on the health and social outcomes of adolescent self-harmers as they enter young adulthood. As the cohort participants grow older, further data will emerge about the long-term health and social outcomes of these young people. Other longitudinal research indicates that the majority will cease self-harming³ yet the behaviour may still signal an underlying vulnerability to serious difficulties in later adult life.⁴ The mechanisms underpinning these adverse outcomes need to be investigated. Both suicidal and non-suicidal self-harm were associated with adverse outcomes in adulthood and it remains unclear whether these behaviour are better conceptualised as part of a continuum, or as separate behavioural disorders. Replication of the ALSPAC findings in other international cohorts, will be helpful. Perhaps most urgently, we need to identify effective ways of preventing and treating self-harm in adolescents, irrespective of the intent associated with such behaviour.

Do these results change your practices and why?

Yes—because it is now evident that adolescent self-harm is an important indicator of future mental health status in young adulthood. Adolescents who report self-harming behaviour (regardless of whether or not they report suicidal intent) should be carefully followed-up to assess their need for support and treatment. Interventions should not only focus on reducing self-harm, but should also treat the anxiety, depression and substance use problems that may accompany self-harming behaviour.