Background: Standard suicide risk assessments rely on self-reports of suicidal ideation (SI). However, up to 75% of suicide decedents explicitly deny SI during their final communications. Thus, implementation of imminent risk assessment tools not relying on SI is critical. The Suicide Crisis Syndrome (SCS) is an acute suicidal mental state under review for inclusion in the DSM as a suicide-specific diagnosis. SCS does not require SI, and the SCS Assessment (SCS-A) is uniquely predictive of suicidal behavior near-term post-hospital discharge. In emergency rooms, the abbreviated SCS-A was a preferred tool for clinical admission/discharge decision-making and associated with decreases in post-discharge suicidal outcomes.
Aims: We aim to implement and test the effectiveness of the SCS-A for reducing post-discharge suicidal behavior in a large urban hospital system. We will compare two implementation approaches: the Diagnosis (D) arm will use a categorical diagnosis of SCS to guide clinical decision making, whereas the Symptoms (S) arm will imbed SCS symptoms in the standard of care Mental Status Examination/Suicide Risk Assessment (MSE/SRA) without the SCS diagnosis, as recommended by some.
Hypotheses: We hypothesize that SCS-A implementation will result in decreased post-discharge suicidal outcomes in both arms compared to pre-implementation. This decrease will be larger in the D vs. S arm and will be mediated by the concordance of clinicians keep/discharge and intensify/not intensify treatment choices with the SCS diagnosis (Epic electronic health records) and with the self-reported use of SCS-A in choosing treatment (Epic). Both will be superior in the D vs. S arm.
Sample: We will analyze suicidal outcomes from Epic data for all inpatient admissions over the three-year study duration (expected N=7800). Additionally, we will recruit 930 ethnically and racially diverse high-risk adult Patient Participants (PPs), as well as 420 staff participants (SPs).
Measures: Patients primary suicidal outcomes will include one-month post-discharge suicide deaths, suicide attempts, emergency room visits and hospitalizations (Epic). Patients secondary suicidal outcomes will be SCS, SI, preparatory acts, and treatment utilization (in PPs). We will assess treatment decisions through Epic by the concordance (concordant=1; reverse=0) of
keep/discharge and intensify/not intensify clinicians choices with the objective SCS diagnosis (explicit yes/no in the D arm, or implicit through the SCS criteria met in MSE/SRA in the S arm), as well as with the subjective self-reported use of SCS-A in choosing treatment (D and S arms).
Procedures: We will compare outcomes at four Mount Sinai Health System sites across four study phases: pre-implementation, implementation, post-implementation, and sustainability, using the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework. Randomization will be enacted through Epic and each site will only get Epic templates specific to that study arm.
Impact: If our hypotheses are supported, we will establish that successful implementation of the diagnostic SCS-A in major urban health system improves suicide-related outcomes, establishing a blueprint for future multi-center SCS-A use using implementation science. In line with the AFSP goal of 20% by 2025, subsequent widespread implementation of diagnostic SCS-A, alongside traditional risk assessments, will have the potential to measurably reduce US suicide rates
