Effect of the LifeSpan suicide prevention model on self-harm and suicide in four communities in New South Wales, Australia: a stepped-wedge, cluster randomised controlled trial

Background There have been few rigorous evaluations of population, multi-strategy, suicide prevention programmes, despite increasing global recognition that such approaches are needed to reduce suicide. Objective To examine the effects of a multi-strategy suicide prevention model on age-standardised rates of hospital presenting self-harm and suicide after 24 months of implementation. Methods A stepped-wedge cluster randomised trial was conducted in four sites across New South Wales (NSW), Australia, from 2016 to 2020. Sites were randomised to a starting order and implemented the same set of interventions over a 24-month period. Changes in rates of hospital presenting self-harm and suicide deaths were measured using linked administrative health data sets of persons aged 10 or older. Results Negative binomial regression models adjusted for linear trends and seasonality showed that LifeSpan was associated with a 13 8% (incident response rate 0.86; 95% CI 0.79 to 0.94) reduction in hospital-presenting self-harm rates over the intervention period, compared with preintervention. These effects were not observed in the rest of NSW. There were statistically non-significant changes in suicide death rates during the intervention across all sites. Conclusions Locally implementing a multi-strategy suicide prevention model can reduce rates of hospital presentations for self-harm, but longer implementation and evaluation periods may be required to realise the full impacts of interventions for suicide, as a more intractable outcome. Clinical implications Our findings can inform policy at all levels of government to invest in actions that may build cross-sectoral capacity in local communities to detect and respond to suicide risk.