Dr. Casey Sarapas, a research scientist at Chestnut Health Systems’ Lighthouse Institute, unpacks a study he led, titled Predictive Validity of the e-Connect Suicide Risk Classification Algorithm in Youth on Probation (selected as among 2025’s best studies by the senior editors at JAACAP Open). 

In this episode, Casey discusses the science behind eConnect’s risk classifier, which combines youth self-report data, substance use patterns, and recent mental health symptoms to assign risk levels more precisely than traditional methods. Casey shares how this tool outperforms many existing assessment measures, with a prediction accuracy (AUC) of around 0.7, capturing more youth at risk than ever before. He breaks down the key components—unsupervised self-report questionnaires, risk classification algorithms, and referral pathways—and how each element addresses critical gaps in behavioral health care for underserved populations.

Casey explains how this technology is designed to work seamlessly for non-clinicians, removing the burden of interpretation from probation officers and fitting into their often-overwhelmed workflow. We discuss scalable models for other settings like schools and child welfare agencies, and explore the potential for periodic reassessment to monitor progress and adapt interventions over time. If you're involved in juvenile justice, mental health, or youth advocacy, this episode is essential, explaining how frontline workers can make faster, better-informed decisions in preventing the alarming rise in suicidal ideation among justice-involved youth.