Evaluating the Comparative Validity and Reliability of SERAS: A Decision Support Tool for Assessing Near Term Risk of Suicide in Emergency Departments

Period of Performance: 05/01/2017 - 04/30/2018


Phase 1 SBIR

Recipient Firm

Wiser Systems, LLC
Principal Investigator


PROJECT SUMMARY The goal of the NIMH Zero Suicide Initiative is prevention of suicide among individuals being served by health care systems. The ?Products to Support Applied Research Towards Zero Suicide Healthcare Systems? FOA specifically calls for development of health IT products designed to meet this goal. Our Systematic Expert Risk Assessment of Suicide (SERAS?) is a neural-network-based decision support tool that replicates the critical thinking process of experienced clinicians in weighing risk factors to assess an individual's near-term risk of suicide (Desjardins et al., 2016). SERAS? triages patients by level of near-term risk. It takes less than 1 minute to administer and 91% of patients rate the interaction favorably. We aim to further establish the scientific merit of SERAS? to improve immediate, near-term suicide prevention in hospitals by comparing the performance of SERAS? to the most commonly used suicide risk detection method: the Columbia-Suicide Severity Rating Scale (C- SSRS). Every year in the United States, there are more deaths from suicide than from motor vehicle accidents (National Center for Health Statistics 2014). Up to 80% of those who die from suicide had received health care services for reasons unrelated to suicide or mental health in the year preceding their death, and up to 64% had visited a doctor within the prior month (Ahmedani, 2015). This latter finding helped prompt the Joint Commission (JC) to mandate screening of suicide potential for all patients in contact with healthcare settings nationwide (most recent JC Sentinel Event Alert, Feb 24, 2016). At present there is a lack of efficient and validated measures of suicide-related behaviors necessary for real-time clinical decision making (Bongiovi-Garcia et al., 2009; Ronquillo et al., 2012; Randall et al., 2011; Horowitz et al., 2012; Doran et al., 2016). This is especially true for the assessment of near-term risk for suicide, which is variably defined and poorly assessed by existing screening tools. The Columbia Suicide Severity Rating Scale (C-SSRS) (Posner et al., 2011; Youngstrom et al., 2015) is widely used for the detection of lifetime suicidal ideation, behaviors, and related non-suicidal self- injury (Posner et al, 2011; Youngstrom et al., 2015; Brown et al., 2015). However, it does not address near-term risk. There remains a critical need for a new tool that can be rapidly administered in the ED, allows for prediction of near-term risk, and provides decision support. The current prospective clinical study proposes to collect a sample of 480 patients in an emergency department with aims to 1) compare the predictive ability, sensitivity, specificity, and test-retest reliability of SERASTM compared to the C-SSRS and to 2) confirm level of patient satisfaction with SERAS?. Completion of these aims will allow us to establish the technical reliability and scientific merit of SERAS? and prepare us for a Phase II proposal in which we will conduct cost-benefit analyses, obtain a broader sample of patients, finalize the technical feasibility optimization (data model and interface) of SERAS?, and prepare a detailed commercialization plan.