Pilot Testing Implementation of Suicide Risk Prediction Algorithms to Support Suicide Prevention in Primary Care

点击翻译按钮获取中文摘要

PROJECT SUMMARY In 2020, suicide was among the top three causes of death for adolescents and young adults (age 10-34) and among the top nine for adults (age 35-64). Recently, researchers have successfully developed suicide risk prediction algorithms that have potential to vastly improve identification of individuals at high risk of suicide and support primary care-based suicide prevention practices. However, there is very little evidence to guide routine use of suicide risk prediction algorithms during healthcare encounters. The most well-known implementation work to date has focused on outreach, like the ReachVet program, which researchers recently reported is associated with greater treatment engagement and safety plan documentation and fewer psychiatric hospitalizations, emergency department visits, and suicide attempts. Visit-based implementation efforts have been less common. Kaiser Permanente Washington leaders piloted this approach using a using a visit-based “flag” among providers in one mental health specialty clinic. The leader of this proposal partnered with healthcare system leaders to conduct a mixed-method implementation evaluation, which found the visit-based risk “flag” did not consistently prompt additional suicide risk assessment as intended by mental health providers and provided a roadmap for quality improvement and future implementation efforts. Therefore, the purpose of this project is to address the RFA-MH-22-120 objective by conducting a practice- based pilot implementation evaluation designed to guide the use of suicide risk predictive analytics in primary care. Specifically, a multi-disciplinary team of researchers, including developers of suicide risk predictive analytics and primary care providers, will work in partnership to build and support implementation of clinical decision support tools designed to identify and engage primary care patients (adults and adolescents) at high risk of suicide in risk mitigation and follow-up care pathways. The research team will use the Discover, Design and Build, and Test Human-Centered Design framework to support three specific aims: 1 (DISCOVER): Conduct qualitative and statistical analyses to identify opportunities to use predictive analytics to guide clinical decision making to support suicide prevention in primary care. 2 (DESIGN & BUILD): Develop and iteratively refine clinical decision support using suicide risk predictive analytics that will augment workflows for both identifying and engagement primary care patients (age 13+) at high-risk of suicide. 3 (TEST): Pilot test implementation of clinical decision support prototypes in 1-3 primary care clinics and evaluate the implementation via 1) provider surveys and 2) statistical analysis of clinical process and suicide attempt outcomes. This work will support use of suicide risk predictive analytics by healthcare systems nationwide and lay a strong foundation for future evaluations of the effectiveness of this intervention for preventing suicide.