Evaluating Effectiveness and Implementation of a Risk Model for Suicide Prevention Across Health Systems

Kaiser Foundation Research Institute, Oakland CA

Investigators

Linked publications, trials & patents

Abstract

PROJECT SUMMARY/ABSTRACT: Despite decades of prevention research, suicide remains a persistent public health problem in the United States. In the last decade, rates of suicide death increased among both males and females, and across age groups between 15 and 44. However, arguments for universal suicide risk screening are weakened by the pragmatic realities of suicide prevention in health systems, the rarity of suicide mortality, and the lack of 1) association between suicide ideation and near-term risk of suicide mortality, 2) evidence for risk stratification based on suicide ideation, and 3) screening instruments validated for suicide mortality. Risk identification models using electronic health records (EHR) data have demonstrated superior accuracy compared to traditional clinical assessment or self-reported screening measures. Yet, use of these models in clinical settings outside of the Veterans Administration has not been evaluated. Our ongoing clinical trial, “Evaluating Effectiveness and Implementation of a Risk Model for Suicide Prevention Across Health Systems” (R01 MH130548), is designed to test the ability of a suicide risk identification model to reduce suicide attempts in real-world settings that represent typical care. The clinical trial is limited to adults receiving behavioral health services; this is the population our clinical partners (three large health care delivery systems) were interested in testing first. However, all of these systems acknowledged interest in spreading the model to youth and to primary care or other settings. In this competitive revision application, we propose to collect additional data to evaluate the effect of the model on an adolescent population that is currently receiving the suicide risk identification model intervention but not being evaluated as part of the original study, and to describe characteristics of other at-risk groups who do not receive any intervention in the original study (e.g., patients insured by but receiving behavioral health outside one of these health systems, those presenting in primary care, those not presenting for care, and patients at varying risk thresholds lower than those set within the study). Studying these other groups will help to prepare for broader scale implementation within and beyond the health systems participating in the trial. Based on previous experiences, if the clinical trial shows that the risk model reduces suicide attempts, we expect our delivery system partners will eagerly proceed with broader implementation ahead of evidence justifying its use in these other populations. Therefore, it is urgent that we leverage the opportunities provided in the original study to learn what we can to inform next steps and to help new adopters considering these implementation decisions. The proposed competitive revision is innovative in that it will be the first study to evaluate the impact of a suicide risk identification model in an adolescent sample. It also offers the opportunity to study potential primary-care-based and outreach-based approaches in a non-Veteran sample.