JMIR Publications

ABSTRACT

Native Americans experience disproportionately high suicide rates. While machine learning (ML) models leveraging electronic health records have emerged as promising tools for suicide risk identification, the optimal integration of these models with existing practices, including screening approaches remains unclear. This study evaluated parallel and serial testing strategies combining an ML risk model with the Ask Suicide Questionnaire (ASQ) to identify patients at risk for suicide attempts or deaths within 90 days of Emergency Department visits at an Indian Health Service facility. Through the Native-RISE project, we examined 26,896 ED visits among 7,897 Native American patients to determine how combined approaches compare to screening alone in terms of sensitivity, specificity, and predictive values.