Suicide prevention in schools: all plain SEYLE-ing?

Mar 12 2015

Suicide is one of the main causes of death in young people. According to the WHO it is the second leading cause of death in 15-29 year olds after traffic accidents, accounting for 8.5% of all deaths (WHO, 2014).

Preventing suicide is therefore the aim of several intervention programmes that range from national strategies or action plans to individual support groups.

Interventions delivered in schools have significant potential, given the suicide rates in this age group and the fact that most adolescents attend school, which makes them easy to reach. However, there are only a few trials of school-based interventions to reduce suicide attempts, and all are from the USA.

In this context, a large-scale, European programme was developed called the Saving and Empowering Young Lives in Europe (SEYLE) study (Wasserman et al, 2015). This aimed to test the effectiveness of various preventive interventions to reduce new cases of suicide attempt in a randomised controlled fashion.

Suicide is the second leading cause of death in 15-29 year olds worldwide (WHO, 2014).

Methods

This was a multi-centre, cluster-randomised controlled trial at 168 schools in ten European countries (Austria, Estonia, France, Germany, Hungary, Ireland, Italy, Romania, Slovenia and Spain). All schools were mixed-sex, and had at least 40 pupils aged 15 years. Pupils completed a questionnaire at baseline, 3 months and 12 months that assessed risk behaviours, symptoms of psychopathology, suicidal thought, plans and suicide attempts.

The primary outcome was incident suicide attempt(s) at either the 3 month or 12 month follow-up. This was defined as a pupil responding “yes” to the question: ‘Have you ever made an attempt to take your own life?’ on either follow-up questionnaire.

Schools were randomised to a control intervention, or one of three active intervention programmes:

Question, Persuade, and Refer (QPR)
Youth Aware of Mental Health (YAM)
Screening by Professionals (ProfScreen)

Only YAM involved the students themselves. QPR relied on educating teachers and other school staff to recognize suicide risk in students and encourage them to seek help. ProfScreen consisted of health professionals reviewing the baseline responses of every student and inviting those scoring above a specific cutoff to attend a mental health clinical assessment.

In contrast, YAM consisted of role-play sessions, interactive workshops, interactive lectures at the beginning and end of the intervention, and a booklet that the students could take home. The material covered risk and protective factors for suicide, coping skills, and knowledge about depression and anxiety.

Researchers used generalised linear mixed models to test for intervention group differences between the control group and each intervention individually.

The YAM program involved role-playing and interactive workshops and lectures.

Results

Of the 27,099 students approached, 11,110 were assessed at baseline, with a median age of 15 years (59% girls). Only 622 (5.6%) were unavailable at either follow-up timepoint. Roughly equal numbers of students were assigned to each condition (roughly 2,700), with those referred at baseline for psychiatric treatment excluded from the analysis.

No completed suicides were reported for any study participants
Comparisons of the individual interventions against the control showed:
- No effect on suicide attempts at 3 months follow-up,
- But a significant effect of YAM at 12 months (0.7% cases in YAM vs 1.51% cases in control; OR for YAM = 0.45, 95% CI 0.24 to 0.85)
A similar pattern of results was shown for suicidal ideation
The number-needed-to-treat for YAM was 167. This means that 167 students needed to receive YAM to prevent one suicide attempt per year

The YAM program NNT was 167. A high number but perhaps worth it for such a serious outcome as attempted suicide.

Strengths

This represents a massive effort, recruiting a very large sample across many sites. The excellent follow-up participation rate is particularly noteworthy. Key confounding characteristics were controlled for in the analysis, including past history of suicide attempt and recent history of suicidal ideation.

Weaknesses

Although this represents a very large study, it was only sufficient to detect a 50% reduction in suicide attempts with 80% power. Smaller effects in the QPR and ProfScreen may have been present but not detectable.

The statistical analysis is not ideal. Firstly, the primary outcome was assessed at two timepoints (3 and 12 months), and then each active intervention was compared to the control, with no corrections for multiple comparisons. A full statistical correction would probably have resulted in a non-significant finding. Furthermore, there was no comparison between the active interventions, so it is unclear whether there is a significant advantage for YAM. Although neither QPR or ProfScreen differed significantly from the control, their odds ratios at 12 months were below one and not that dissimilar to YAM (0.7 and 0.65 respectively). In any case, such a difference represents about five cases of suicide attempt per group, which is obviously vulnerable to chance factors at such low incidence rates.

For ethical reasons, the control group was exposed to the same educational posters that were used in YAM, which may have reduced the effect size for all three active interventions compared to controls.

Only 41% of eligible children took part in the trial, and there was a significant over-representation of girls (although sex did not significantly modify the intervention effect). It’s unclear why the participation rate was this low, and this does limit the generalizability of the study.

Conclusions

This represents a landmark study of suicide prevention interventions in school-age populations. It represents good evidence that a universal intervention can have significant impact, but it remains to be seen if the findings can be replicated and if it represents the most cost-effective approach.