Our technical team, in partnership with clinicians from the areas of mental health, emergency medicine, and psychiatry, developed an automatic and data-driven predictive model, the Behavioral Health Acuity Risk (BHAR) model, to examine information available in the medical record captured during the normal course of care and assess the likelihood of suicidal behavior. This model leverages a machine learning technique called random forests, which can be natively hosted in the Electronic Health Record and update in near-real time, with results immediately available to clinical staff.

Click here to view the slide deck.

Karen Hegarty, PhD, VP Cognitive Computing, Novant Health