Significance: As recent national controversy over Joint Commission mandates proves, universal suicide risk
screening in emergency departments (ED) will not achieve widespread adoption because confusion remains
around which specific risk indicators to assess, and clinicians fear that such screening will lead to massive
surges in psychiatric evaluations. To address these two implementation barriers, the proposed study will derive
a clinical decision rule to support universal risk detection and optimize patient care workflow in adult patients.
Investigators: The Project Team has extensive expertise in ED-based suicide risk screening and assessment
(Boudreaux, Larkin), clinical decision rule design (Boudreaux, Stiell), predictive analytics (Wang, Liu, Simon),
machine learning and informatics (Liu, Simon), industrial engineering (Johnson), and healthcare economics
(Clements). A Clinical Advisory Panel ensures that the proposal is grounded in the practical realities of the ED.
Innovation: The proposed study will be the first to apply industry standards for deriving decision rules to
suicide risk and will directly inform the controversy regarding the relative strengths and weaknesses of
universal versus targeted screening. We will pioneer new statistical innovations for rule derivation and will
integrate simulation of potential workflow impact using industrial engineering modeling and economic analyses.
Approach: We have already developed a pool of empirically supported, clinician-acceptable candidate suicide
risk indicators. Data on these candidate indicators will be collected by trained research staff on 500 adult
medical patients and 500 adult psychiatric patients from a large ED. Participants will undergo a comprehensive
suicide risk assessment by a research clinical psychologist blinded to the indicators who will assign the
participant to a criterion reference risk group: Negligible, Mild-Moderate, or High risk. Participants will be
followed for 24 weeks after the visit to assess suicidal behavior, our secondary outcome. In Aim 1, we will
derive a universal screening decision rule for “all comers,” as well as a variant to be used with patients
presenting with a psychiatric chief complaint (targeted). In Aim 2, we will test whether a previously validated
risk stratification algorithm using data from the electronic health record improves the performance of the
decision rules. In Aim 3, we will model the potential operational impact of the rules through dynamic modeling
of clinical workflow and economic costs and assessing clinician and patient acceptability in a new sample of
100 ED clinician-patient dyads.
Environment: UMass has demonstrated its capability to support this study through several key preliminary
studies, including the ED-SAFE studies, System of Safety, and other suicide-related studies set in the ED.
Impact: By providing clear, evidence-based recommendations on universal screening and optimized workflow
using standards accepted by emergency clinicians, this study will address two pivotal barriers to universal
suicide risk screening, transforming the “right thing” into the “easy thing” so it becomes the “usual thing.”