Clinical implementation of AI-based screening for risk for opioid use disorder in hospitalized adults

The article reports on an NIH-supported clinical trial showing that an artificial intelligence (AI)–based screening tool for opioid use disorder (OUD) can improve hospital outcomes without reducing quality of care. Embedded in the electronic health record, the AI system continuously scanned clinical notes to identify hospitalized adults at risk for OUD and prompted referrals to inpatient addiction specialists. In the trial of over 50,000 hospitalizations, the AI-assisted approach generated addiction medicine consultations at rates comparable to usual provider-initiated screening (about 1.5% vs. 1.3% of patients) while requiring less ad hoc effort from clinicians. 

Critically, patients identified through AI screening had 47% lower odds of being readmitted within 30 days (approximately 8% vs. 14% readmission rates), leading to an estimated $109,000 in healthcare savings over the eight-month implementation period. The findings, published in Nature Medicine, suggest that AI tools can serve as a scalable, cost-effective complement to clinicians in detecting OUD early, supporting addiction care, and reducing avoidable hospital use, while still requiring careful evaluation and thoughtful integration into clinical workflows. 

Click here to access the National Institute of Health news release