Highlight
This pragmatic cluster randomized trial evaluated the effect of an electronic health record (EHR)-integrated clinical decision support system (CDSS) aimed at improving opioid use disorder (OUD) diagnosis and treatment in primary care clinics across multiple US health systems. While the CDSS did not increase OUD diagnosis rates within 30 days, it significantly improved naloxone prescribing and initiation of medications for OUD (MOUD) or specialty referrals. However, MOUD adherence over 90 days and overdose or death rates were unchanged, highlighting the potential and limitations of CDSS in addressing the opioid crisis in primary care.
Study Background and Disease Burden
The United States faces an ongoing opioid overdose epidemic, with more than 727,000 deaths between 1999 and 2022 attributed to opioid overdoses. This staggering mortality underscores an urgent need for expansive access to effective treatment for opioid use disorder. The shortage of addiction medicine specialists and the underutilization of primary care clinicians (PCCs) to diagnose and manage OUD contribute to suboptimal treatment coverage. Many PCCs report insufficient support and training to confidently treat OUD, hindering expansion of care. Integrating evidence-based clinical decision support into EHRs may facilitate PCC engagement in OUD management by offering real-time, personalized treatment recommendations, potentially increasing diagnosis and treatment rates in primary care settings where most patients receive care.
Study Design
This pragmatic cluster randomized clinical trial enrolled primary care clinics from three health systems across four US states, randomizing clinics to either intervention (EHR-integrated OUD CDSS) or usual care. The study period spanned April 2021 to December 2023, with data analyzed from September 2023 to October 2024.
Eligible patients were aged 18 to 75 years, had at least one visit to a randomized clinic during the study period, and met criteria including an OUD diagnosis within the previous two years, a documented opioid overdose in the last six months, or a predictive risk score indicating high risk for OUD or opioid overdose.
The intervention involved an OUD CDSS integrated into the EHR, providing personalized treatment recommendations for patients and PCCs. The CDSS aimed to support decisions on diagnosis, naloxone prescribing, initiation of MOUD, or referral to specialty care.
Primary outcomes assessed were:
- The likelihood of receiving a new OUD diagnosis among high-risk patients without a baseline OUD diagnosis, within 30 days of the index visit.
- The likelihood of receiving a naloxone prescription within 30 days.
- The likelihood of receiving a MOUD prescription or specialty referral within 30 days.
- The number of days covered by a MOUD prescription during the 90 days after the index visit.
Key Findings
The study included 10,891 eligible patients with mean age of 48 years; 54.3% were female.
OUD Diagnoses: There was no significant difference in the rate of new OUD diagnoses within 30 days between the intervention and usual care groups.
Naloxone Prescriptions: Intervention clinics had significantly higher naloxone orders (1.4% vs. 0.7%; odds ratio [OR] 1.76; 95% confidence interval [CI], 1.14–2.72), indicating increased secondary prevention efforts.
MOUD Prescriptions or Specialty Referrals: Patients in the intervention groups were more likely to receive MOUD prescriptions or treatment referrals within 30 days (14.0% vs. 9.4%; OR 1.48; 95% CI, 1.05–2.08).
MOUD Adherence: Median days covered by MOUD over 90 days postindex did not differ significantly (84 days intervention vs. 83 days usual care; rate ratio 1.00; 95% CI, 0.93–1.08).
Clinical Outcomes: No significant differences were observed in opioid overdose or death rates during the intervention period between groups.
Adverse safety signals related to treatments were not reported, consistent with the known safety profile of MOUD and naloxone.
Expert Commentary
The findings elucidate how a well-designed EHR-integrated CDSS can enhance certain treatment behaviors in primary care, notably increasing naloxone distribution and treatment initiation. However, the system did not influence timely OUD diagnosis rates or sustained engagement with MOUD, suggesting additional barriers beyond decision support, such as stigma, limited resources, patient adherence challenges, and structural healthcare system constraints.
Moreover, the unchanged overdose and mortality rates may reflect the multifactorial nature of these outcomes, requiring multifaceted interventions that go beyond improved prescribing practices alone.
Limitations include the pragmatic design which may allow real-world variability in implementation fidelity, possibly attenuating effect sizes. Additionally, patients’ social determinants of health, comorbidities, and competing clinical priorities were not deeply explored here, contributing to the complexity of improving long-term outcomes.
Current clinical guidelines endorse integration of MOUD and naloxone distribution in primary care. This trial’s supportive evidence for enhanced prescribing via CDSS aligns with calls for greater primary care involvement but suggests that supportive tools should be complemented by systemic interventions to address diagnosis and retention challenges.
Conclusion
This cluster randomized trial demonstrates that an EHR-integrated clinical decision support system can improve key treatment processes for OUD in primary care, increasing naloxone orders and MOUD initiation or referrals. Nevertheless, the intervention did not significantly advance diagnosis rates or medication adherence over 90 days, nor reduce overdoses or deaths. These findings underline the potential of CDSS to extend impactful OUD treatment access via primary care but also highlight persisting challenges that require multifaceted strategies integrating clinical tools, provider training, patient engagement, and broader healthcare system support for meaningful reductions in opioid-related morbidity and mortality.
References
- Rossom RC, Crain AL, Wright EA, et al. Clinical Decision Support System for Primary Care of Opioid Use Disorder: A Randomized Clinical Trial. JAMA Intern Med. 2025;185(9):1079-1089. doi:10.1001/jamainternmed.2025.2535
- Kampman K, Jarvis M. American Society of Addiction Medicine (ASAM) National Practice Guideline for the Use of Medications in the Treatment of Addiction Involving Opioid Use. J Addict Med. 2015;9(5):358-367. doi:10.1097/ADM.0000000000000166
- Volkow ND, Jones EB, Einstein EB, et al. Prevention and Treatment of Opioid Misuse and Addiction: A Review. JAMA Psychiatry. 2019;76(2):208–216. doi:10.1001/jamapsychiatry.2018.3126
- Williams AR, Nunes EV, Bisaga A, et al. Developing an opioid use disorder treatment continuum: a review of barriers and facilitators of treatment entry, retention, and outcomes. Curr Psychiatry Rep. 2019;21(2):3. doi:10.1007/s11920-019-0994-3