(50) Real-World Healthcare Utilization With SROM Versus Methadone: A Target Trial Emulation

Background & Introduction: Over the past 25 years, the opioid-related overdoses and deaths in North America have escalated through multiple waves, reaching an unprecedented scale [1]. In the United States, the total cost of excess health care utilization associated with opioid use disorder (OUD) is currently estimated at approximately $90 billion per year, primarily attributed to costs associated with hospital admissions and emergency department (ED) encounters [2]. Initiating and adhering to effective medications for OUD (MOUDs) are consistently associated with reduced overdose risk, improved health outcomes [3], and reduced healthcare use over time [4]. However, direct comparative evidence on healthcare utilization across MOUDs remains limited, with no available evidence for slow-release oral morphine (SROM), a new alternative MOUD. The objective of this study is to examine the comparative effect of SROM vs methadone on healthcare utilization—a policy- and public-health-relevant marker of morbidity, care engagement, and system burden.

Methods: This is an observational emulation of a real randomized controlled trial (pRESTO: NCT03948464) [5] utilizing electronic medical records from Vancouver Coastal Health Authority linked with population-based administrative health data from British Columbia, Canada. Eligible patients aged between 18 and 65 years with a new prescription of SROM or methadone for OUD between July 1, 2017 and June 30, 2023 were identified. Healthcare utilization outcomes over 1 year following treatment prescription or initiation were examined, including (1) all-cause ED visits, (2) all-cause hospitalizations, (3) all-cause physician visits, (4) OUD-related physician visits, and (5) non-OUD-related primary care physician visits. We estimated the observational analogues of the intent-to-treat (ITT) (i.e., receipt of a new prescription of SROM or methadone) and the per-protocol (PP) (i.e., initiating and adhering to the assigned treatment) effect. The daily rate of healthcare utilization over 12 months was modelled using weighted modified Poisson regression with generalized estimating equations on pooled person-day data. Cumulative healthcare utilization at 12 months and corresponding rate ratios were estimated. Inverse probability of treatment weights and instrumental variable approach were used to account for baseline confounding; marginal structural model with inverse probability of censoring weights was used to adjust for time-varying censoring at treatment deviation.

Results: We identified a total of 3,254 unique individuals (median [Q1–Q3] age 37 [30–46] years, 64.2% male) with a new prescription of SROM or methadone contributing to 4,059 person-trials (32.6% SROM) for the ITT analysis. The PP analysis included 1,992 unique individuals contributing to 2,276 person-trials (27.4% SROM), restricted to trials in which the assigned treatment was initiated within a pre-specified 4-day grace period. ITT analyses showed similar rates per person-year over 12 months under SROM vs methadone for ED visits (5.36 vs 4.89 ; adjusted rate ratio [aRR] 1.10, 95% CI 0.997–1.20), days hospitalized (7.12 vs 7.22; aRR 0.99, 95% CI 0.81–1.18), all-cause physician visits (25.34 vs 25.35; aRR 1.00, 95% CI 0.94–1.06), OUD-related physician visits (14.35 vs 15.31; aRR 0.94, 95% CI 0.87–1.01), and non-OUD-related primary care visits (3.01 vs 2.72; aRR 1.11, 95% CI 0.98–1.24). PP analysis yielded consistent results suggesting similar healthcare utilization among patients initiating and adhering to the prescribed SROM vs methadone.

Conclusion & Discussion: This emulated trial on OUD patients under a universal healthcare setting in Canada found similar risk of healthcare utilization outcomes between the SROM and methadone groups. This supports SROM as a viable alternative, expanding treatment options without increasing health system burden. This study has several limitations. Causal inference from observational data relies on key assumptions—some cannot be verified empirically with data. Given study population was predominantly urban (>90%), findings are less generalizable to rural settings. Estimates for less common outcomes, such as hospitalization, had wider confidence intervals. Future studies with larger sample sizes or longer follow-up are warranted to improve precision to detect small but clinically meaningful effects.

References: 1. Humphreys K, Shover CL, Andrews CM, et al. Responding to the Opioid Crisis in North America and Beyond: Recommendations of the Stanford-Lancet Commission. Lancet. 2022;399(10324):555-604. doi:10.1016/S0140-6736(21)02252-2
2. Velez FF, Colman S, Kauffman L, Ruetsch C, Anastassopoulos K, Maricich YA. Comparison of Healthcare Resource Utilization Between Patients Who Engaged or Did Not Engage With a Prescription Digital Therapeutic for Opioid Use Disorder. Clinicoecon Outcomes Res. 2021;13:909-916. doi:10.2147/CEOR.S334274
3. National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Health Sciences Policy; Committee on Medication-Assisted Treatment for Opioid Use Disorder. Medications for Opioid Use Disorder Save Lives. (Mancher M, Leshner AI, eds.). National Academies Press (US); 2019. Accessed December 29, 2025. https://www.ncbi.nlm.nih.gov/books/NBK538936/
4. Onuoha EN, Leff JA, Schackman BR, McCollister KE, Polsky D, Murphy SM. Economic Evaluations of Pharmacologic Treatment for Opioid Use Disorder: A Systematic Literature Review. Value Health. 2021;24(7):1068-1083. doi:10.1016/j.jval.2020.12.023
5. Anand R, Penta S, Cui Z, Fairbairn N, Socias ME. Redesigning a Stopped Clinical Trial as an Emulated Trial Using Real-World Data to Explore the Effectiveness of Slow-Release Oral Morphine as a Treatment for Opioid Use Disorder. Subst Use Addctn J. 2025;46(2):384-390. doi:10.1177/29767342241279167