The goal of this clinical trial is to learn if a mobile health app called Recovery Connect can help people stay in treatment for opioid use disorder. The study is being done with adults receiving medication-assisted treatment in outpatient opioid treatment programs.

The main questions it aims to answer are:

Does Recovery Connect increase the number of people who remain in treatment after 30 days?

Does Recovery Connect improve continuity of care during the first month of treatment?

Researchers are using a stepped-wedge design, which means that all participating clinics will eventually use the app, but they will begin at different time points in a randomized order. This allows researchers to compare outcomes before and after each clinic starts using Recovery Connect.

Participants will:

Use the Recovery Connect app for daily check-ins, coping strategies, and reminders to support their recovery

Share progress and risk information through the app so counselors can provide timely, personalized support

Work with treatment teams trained to use the app to guide care and reinforce accountability

This study is a stepped-wedge cluster randomized clinical trial evaluating the implementation and effectiveness of Recovery Connect, a mobile health application designed to support individuals receiving medication-assisted treatment (MAT) for opioid use disorder (OUD). Recovery Connect is intended to enhance treatment delivery by:

Providing clinicians with real-time, self-monitored patient progress data to support measurement-based and data-informed care.

Delivering personalized care plans, reminders, and evidence-based coping strategies directly to patients.

Facilitating continuous communication and accountability between patients and counselors between sessions.

Study Design

The trial is conducted across nine outpatient opioid treatment program (OTP) clinics, grouped into eight clusters. Using a stepped-wedge design, each cluster will transition from usual care to the Recovery Connect intervention at randomized, sequential time points, such that all clinics will eventually receive the intervention. All clinicians at each cluster will receive standardized training on Recovery Connect at rollout, with follow-up reinforcement training provided two weeks later. Auxiliary staff (e.g., front desk, peer support, navigators) will also be trained to support patient onboarding.

Participants All clinicians at participating clinics will be included. Patient-level data will be analyzed for all new admissions and readmissions during the study period.

Data Collection

App Usage and Engagement: Patient and clinician usage data, including daily check-ins, logins, and secure messages, will be captured directly through the app.

Patient Treatment Outcomes: Treatment retention and continuity data will be obtained via the Methasoft medical record system, following standard clinic protocols.

Operational Definitions

Exposure: Patients are considered "exposed" if their admission occurred after clinicians at their clinic had completed Recovery Connect training.

Linkage: Patients are considered "linked" if they downloaded the app and were connected with a trained clinician during admission.

Engagement: Defined by app activity within the first week after linkage, including logins, daily check-ins, and secure messaging.

Analyses The primary analyses will use an intent-to-treat framework, including all eligible patients whose clinic underwent Recovery Connect training. Outcomes will be assessed using multilevel logistic regression and cluster-controlled discrete-time survival analyses, accounting for the stepped-wedge design.

Inclusion Criteria:

• All newly admitted patients to participating clinics

Exclusion Criteria:

• Dosing suggesting a transfer from a non-participating clinic, or if they had previously used the app prior to their most recent admission during the study period.

The datasets generated and analyzed during this study will not be publicly available due to data confidentiality agreements in place between Bright Therapeutics and Community Medical Services, but select de-identified data will be made available from the corresponding author on reasonable request.