Medical-AI trial: predicting overdose risk with machine learning and nudging safer prescribing in the EHR — curbing opioid overdose (cluster-randomized) (NCT06806163)
A cluster-randomized trial of a clinician-targeted behavioral nudge in the electronic health record (EHR) for patients flagged by a machine-learning model as at elevated risk of opioid overdose. It compares a risk flag alone, a flag plus nudges, and usual care. The primary outcome is a prescribing-practices composite score. 1,350 patients; recruiting.
Trial overview (primary data)
- StatusRecruiting
- ConditionsOpioid Overdose, Opioid Use, Opioid Use Disorder, Opioids
- InterventionsBEHAVIORAL: EHR-Embedded Elevated-Risk Flag, BEHAVIORAL: EHR-Embedded Elevated-Risk Flag with Behavioral Nudges, BEHAVIORAL: Usual Care
- SponsorUniversity of Pittsburgh
- Target enrollment1,350 participants
- Period2025-03-10 〜 2027-02-01
Key points
- A cluster-randomized trial of a clinician-targeted EHR behavioral nudge for patients flagged by a machine-learning model as at elevated risk of opioid overdose.
- It compares a risk flag alone, a flag plus nudges (best practice alerts), and usual care.
- The main goals are improving opioid prescribing safety and reducing overdose risk. The primary outcome is a prescribing-practices composite score.
- The design links AI prediction to physician action, the heart of AI implementation: from prediction to action.
- 1,350 patients; recruiting. An evaluation by prescribing practices, not an establishment of overdose-death reduction or model effectiveness.
- Predicting risk means nothing unless prescribing changes, so a behavioural nudge is attached to the flag in the record.
1Opioid overdose as a public health crisis
Opioid (strong painkiller) overdose is a serious public-health crisis in the United States. Who carries risk depends on many factors, such as history, co-prescriptions, and the course of prescribing, and it is easily missed in a busy clinic. If machine learning (ML) can predict high-risk patients from past data and deliver that to the physician, prescribing might be made safer. This study tests such a mechanism.
2Turning prediction into clinical action
Per the registry summary, this cluster-randomized controlled trial tries a clinician-targeted behavioral nudge in the EHR for patients the ML model flags as at elevated risk of opioid overdose. Specifically, it provides a flag in the EHR for high-risk individuals, with and without behavioral nudges (best practice alerts, BPAs), compared with usual care by primary care clinicians.
The main goals are to improve opioid prescribing safety and reduce overdose risk, and the primary outcome is a prescribing-practices composite score. The size is 1,350, and it is recruiting.
3Connecting prediction to action
What distinguishes this study is that it designs in not only the AI's prediction but how that prediction connects to what physicians do.
- 1Predict the riskMachine learning identifies patients at high risk of overdose from past data
- 2Deliver it to the physicianA flag marks the high-risk individual in the electronic health record
- 3Nudge the behaviourA best practice alert is attached as a behavioural nudge
- 4Prescribing changesThe primary endpoint is a composite score of prescribing practice
Predicting risk means nothing if it does not translate into changed prescribing. Comparing a flag alone against a flag plus nudge, both against usual care, goes to the heart of implementing AI: moving from prediction to action. It is a cluster-randomised controlled trial across 1,350 cases, recruiting. A prediction indicates whether risk is high or low; the individual judgement rests with the physician.
Why it matters
The core is covering not just AI prediction but how it connects to physician action. A prediction that does not change prescribing is meaningless; combining a behavioral nudge and comparing flag-alone with flag-plus-nudge gets at from prediction to action (this study evaluates prescribing practices, not an establishment of overdose-death reduction).
FAQ
What is a nudge?
Does the AI decide prescribing?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT06806163