Evidence supports the obvious: suicides need not happen

Although suicide is a devastating and tragic event, it is a relatively rare one. Every year in England and Wales, roughly one suicide occurs per 1000 patients in treatment with mental health providers, the group at greatest risk.¹ Yet one suicide death is too many, leaving in its wake a legacy of pain and traumatic grief. Reduction, and even elimination, of suicide deaths is a high public health and clinical priority.

One consequence of the low base rate of completed suicide is that proving the effectiveness of interventions designed to reduce suicide deaths is very difficult. Without evidence to support such practices, mustering of social, institutional, and political will to sustain them might be impossible. David While and colleagues¹ report such evidence in The Lancet for a range of nine widely used suicide prevention practices in public sector mental health service settings. They used data for suicides occurring from Jan 1, 1997, to Dec 31, 2006, in 12 881 patients who were in contact with any of 91 mental health services in England and Wales in the 12 months before death, describing changes in rates relative to regional uptake of key mental health service recommendations over time. Unsurprisingly, the investigators showed that as more service recommendations were implemented, suicide rates in the target populations of service users with mental illness declined. From 2004 onwards, there were fewer suicides in services in which seven to nine recommendations had been implemented (10·80 suicides per 10 000 in 2004, 95% CI 9·99–11·65) than in those where zero to six recommendations were implemented (12·63, 11·70–13·61). Removal of ligature points on inpatient wards was associated with reduction in overall psychiatric inpatient suicide rates (11·98 suicides per 10 000 per year [95% CI 10·92–13·11] vs 15·66 [14·65–16·73]), and implementation of an assertive outreach policy was associated with decreased rates in the subgroup of the patient population that was non-compliant with care (1·26 suicides per 10 000 per year [95% CI 1·16–1·37] vs 1·77 [1·65–1·89]). These and other reported results might seem obvious, but even the obvious can go unseen and, therefore, unfunded. While and colleagues¹ do a great service to mental health systems, their providers, and patients by lending strong support to the contention that existing recommendations are effective.

The investigators acknowledge the limitations of observational data, and the undocumented gap between policy and practice that their study cannot fill. Putting the lessons learned into practice will require increased attention to the study of factors that affect the uptake, effective implementation, and dissemination of these recommendations as best practices.² Our discipline has learned sometimes painful lessons about the failure of efficacious and even effective treatments to have positive results when taken to scale.³ Such failures underscore the need to understand far better how promising interventions should be translated to practice in a manner that is acceptable to the patients, providers, payers, and policy makers who must implement, consistently apply, and sustain them over time.⁴

While and colleagues¹ focused on mental health service rather than general population suicide rates. The recommendations they examined targeted those with an identified need for mental health care. While necessary, an approach that targets only high-risk individuals (indicated preventive intervention⁵) will have a marginal effect on overall suicide rates, because most people who take their own lives have had no contact with mental health services.⁶ Such an approach must be complemented by a range of interventions that target individuals and groups with characteristics that make them vulnerable to development of suicidal states but who may never have been at imminent risk (selective prevention⁵), and also interventions that target the population as a whole (universal prevention⁵). Coffey,⁶ for example, describes the probable effect of a system-wide quality improvement programme for depression care that resulted in substantial decreases in suicide in behavioural health services patients of the Henry Ford Health System in Detroit, USA.⁷ The multi-layered suicide prevention programme of the US Air Force offers another example.⁸ A complex intervention that targeted stigma reduction to enhance help-seeking behaviour, education to support acceptance of and adherence with care, and changing policies and social norms among all service members was associated with a sustained decline in suicide rates. This intervention was also associated with reductions in other adverse outcomes including accidental death, homicide, and moderate to severe family violence. In addition to indicated interventions, selective and universal preventive interventions are important elements of any comprehensive suicide prevention campaign aimed at reducing suicide-related morbidity and mortality.⁹

We cannot take for granted that even the most seemingly obvious intervention designed to reduce suicide is necessarily effective without evidence, or even that the reduction of suicide is a shared public concern warranting allocation of substantial resources for its prevention. While and colleagues’ contribution is important at each of these levels. To have an even greater effect, we need much more evidence.