Suicide risk classifications do not identify those at risk: where to from here?

Dear Sir,

Wyder and colleagues contributed compelling data to the growing literature and experience base showing risk stratification to be a poor predictor of subsequent suicide death and an insufficient basis for determining appropriate care. Their findings raise a key question: what can clinicians and care systems do instead?

In their implementation of a Zero Suicide Framework, Gold Coast Mental Health and Specialist Services (GCMHSS) and other Hospital and Health Services across Queensland, Australia, replaced risk categorisation with a combination of Chronological Assessment of Suicide Events (CASE) and Prevention-Oriented Risk Formulation. The CASE approach is an interviewing strategy for uncovering withheld intent through specific validity techniques and a collaborative, non-judgemental stance. This approach could help mitigate the concern raised by Wyder et al. who observed that almost half of consumers in their study denied feeling suicidal during their recent contact with the mental health service. Prevention-Oriented Risk Formulation is a sophisticated framework that aids communication and care planning through personalised, contextually anchored judgements of consumer’s (1) risk status (risk relative to a specified subpopulation), (2) risk state (risk compared to baseline or other specified time points), (3) resources available to the consumer when in crisis, and (4) foreseeable changes that may exacerbate risk. The conceptual shift from a predictive to a preventive formulation addresses the imperative need for suicide risk assessment to identify modifiable risk factors that can be the focus of therapeutic interventions and safety planning, ensuring the availability of appropriate supports in the community.

Starting in 2016, GCMHSS pursued a coordinated, systems-wide reform that focused on training and supporting staff with skills, attitudes, and culture consistent with the paradigm shift. The transition from traditional risk stratification to a more clinically useful model required concerted effort but achieved results in a relatively short timeframe. Within a few months of implementation, strong fidelity to the new model enabled the elimination of ‘low’, ‘medium’ or ‘high’ risk in clinical documentation. Most importantly, a 35% reduction in suicide attempt re-presentations was observed after shifting to prevention-oriented risk assessment in combination with safety planning, consumer and carer education, and assertive follow-up.

We share Wyder and colleagues’ concern about the futility of risk stratification and recommend continued reform of risk assessment practices to health services worldwide.