Estimates of Workload Associated With Suicide Risk Alerts After Implementation of Risk-Prediction Model

The findings of this cohort study suggest that suicide risk models can accurately stratify risk but will generate additional workload for clinicians. The degree of added workload depends on the risk threshold selected, the strategy for responding to repeated alerts per patient, and the overlap of electronic alerts with risks already identified by treating clinicians. Selecting a risk threshold also depends on the relative importance of avoiding false-negative and false-positive errors. Without estimates for these factors, attempts to design a system to respond to alerts will be hampered.