Sustaining Recovery for People on Opioid Agonist Treatment With Conversational Agents

Sustaining Recovery for People on Opioid Agonist Treatment With Conversational Agents

This study aims to improve retention in buprenorphine treatment, a medication used for opioid use disorder, through the use of a smartphone-delivered recovery support intervention. The intervention involves an Embodied Conversational Agent (ECA), a virtual, animated computer agent designed to simulate natural face-to-face conversations. ECAs have been shown to help individuals manage their healthcare in other settings, and this study seeks to evaluate their potential in supporting patients on medication for opioid use disorder (MOUD).

One hundred participants will be enrolled in the study from the Office-Based Addiction Treatment (OBAT) clinic at Boston Medical Center (BMC). The participants will be randomly assigned to one of two groups. In the intervention group, participants will engage in ECA use and will receive (1) technical support, (2) text message reminders, and (3) monetary incentives. Alternatively, participants will be randomized to the treatment as usual group where they do not access the ECA application. Assessments will be conducted in person at baseline and over the phone at 6 months and 12 months. At 3 and 9 months, research staff will make brief telephone contacts to confirm or update participant contact information. The primary outcome is retention in buprenorphine treatment over 12 months.

Inclusion Criteria:

• Diagnosis of Opioid Use Disorder (OUD)
• Within 30 days of intake at recruitment site (BMC OBAT)
• Possession of a smartphone that has minimum requirements
• English-speaking and reading
• Willing to release electronic health record (EHR) data
• Able to provide at least two alternate contacts who usually know how to get in touch with them
• Currently prescribed buprenorphine from an outpatient clinic

Exclusion Criteria:

• Incarceration anticipated within 12 months of enrollment
• Inability to comprehend the study protocol, defined as failing three times to answer correctly a set of questions during the consent process
• Inability to use the ECA app (i.e., due to markedly limited visual or auditory acuity or motor function required to interact with the ECA)