Researchers at Ann & Robert H. Lurie Children’s Hospital of Chicago have introduced four artificial intelligence (AI) models designed to address significant challenges in mental health care. The tools aim to reduce the workload for clinicians, improve patient access and safety, and support more personalized treatment.
The demand for mental health services has grown, especially since the COVID-19 pandemic. AI systems are being used to help with early detection of mental illnesses, enhance safety and treatment planning, provide ongoing patient support, and ease clinician fatigue.
One tool, called Neuropsychology Outcome Reporting Assistant (NORA), automates the process of creating clinician reports. Traditionally, this process takes between three to six hours but can now be completed in about three minutes using NORA. The model uses data from multiple sources and is trained on each clinician’s style. It flags uncertain areas for review by a clinician, allowing manual edits that help improve future AI output. This approach could decrease administrative tasks for providers and speed up delivery of results to patients.
Another tool, Medication Information for Neuropsychological Disorders (MIND), provides information about medications, side effects, and dosing to patients and their families. Clinicians currently spend an estimated 30% of their time answering repetitive questions from patients; MIND aims to reduce this burden. Dr. Arjun Nagendran stated: “Training the model on continuously updated, doctor-verified content enables it to provide confidence measures for its responses, deliver real-time, context-specific information, and appropriately acknowledge when information is insufficient and a provider should be contacted.” However, after just 12 months of testing and study, further validation is needed regarding its long-term effectiveness.
The Suicide Assessment Fidelity Evaluator (SAFE) evaluates the quality of suicide risk safety plans in real time—a process that can reduce review times by up to 70%. SAFE offers immediate feedback during plan creation with patients and may also predict emergency department return risk based on available data. Preliminary evidence supports these capabilities. According to researchers, the ultimate goal is for SAFE to assist clinicians by identifying cues that might otherwise be missed.
The fourth tool—the Suicide Prevention AI Role-play Kit (SPARK)—uses simulated child patients powered by AI for suicide prevention training among clinicians. Traditional role-play training is hard to scale due to cost and limited demographic representation; SPARK addresses this through dual-AI simulation: one model acts as the patient while another provides feedback on clinical performance afterward. Project work has begun on SPARK as teams seek additional grant funding.
Dr. Nagendran said that “MIND, NORA and SAFE have been identified as highly implementable,” but added that SPARK will require more time before widespread use: “We are now collaborating with innovation teams for efficacy studies and moving towards commercialization.” Dr. Scott Compton noted all platforms were developed within clinical settings following evidence-based practices; institutional rollout requires approval from an AI governance body.
