Adding a therapy program to buprenorphine and voucher rewards: a Pear-funded reanalysis found more negative tests and fewer dropouts

The short version

reSET-O is a prescription phone app for opioid use disorder, developed by Pear Therapeutics for use alongside buprenorphine. This paper, paid for by Pear and written in part by its employees, reanalyzed an older trial of the same therapy content. People who got the program on top of buprenorphine and voucher rewards had more negative urine tests near the end of treatment and were less likely to drop out. What the trial tested, though, was a web version on clinic computers, not the phone app.

What they did

The original trial ran at one clinic in Arkansas and enrolled 170 adults with opioid dependence. Everyone started buprenorphine and was then randomly assigned: 79 to usual care and 91 to usual care plus a self-guided computer program called the Therapeutic Education System.

Usual care was not bare-bones. It included buprenorphine/naloxone, a 30-minute clinician visit every other week, urine tests three times a week, and contingency management for everyone: vouchers that grew in value with each negative test, up to $997.50 over 12 weeks of continuous abstinence.

The program added 67 interactive modules based on the Community Reinforcement Approach, a behavioral therapy that builds skills and rewards a life without drug use through work, relationships and recreation. Modules also covered preventing HIV, hepatitis C and other infections. Participants used it on clinic computers, without a clinician guiding them.

The trial was first reported in 2014. This 2020 paper went back to the data to look at the last four weeks of treatment, the window the authors say the National Institute on Drug Abuse recommends, and to report safety for the first time. Missing urine tests counted as positive.

What it found

In weeks 9 to 12, the estimated likelihood of an opioid-negative test was 77.3% with the program and 62.1% without it (odds ratio 2.08, 95% CI 1.10 to 3.95; p = 0.02). For opioids and cocaine together, it was 75.9% and 60.6%. For cocaine alone, 82.4% and 64.5%.

People using the program were about half as likely to leave treatment at any given point (hazard ratio 0.49, 95% CI 0.26 to 0.92). By the end, 17.6% had dropped out, compared with 31.6% in usual care.

Program users completed an average of 77.3 modules, with a range from 4 to 150. The program had 67 modules, and participants could go back to ones they had already done.

Adverse events were common in both groups, as expected for people with opioid use disorder: 62.6% with the program and 69.9% without it, not a significant difference (p = .42). No suicide-related events were reported, and none of the adverse events were judged to be caused by the program.

What it does not show

This is a reanalysis, not a new trial. The main outcome window was chosen years after the trial ended, and the paper made no adjustment for testing several outcomes.

The program ran in a web browser on clinic computers. reSET-O, the product this paper supports, is a phone app used at home. The paper says the content was the same, and that these data were the basis for the FDA's market authorization of reSET-O. It does not show how people use the app on their own phones.

Everyone got contingency management. The authors note that high abstinence in both groups may have hidden some of the program's effect. It also means the result says nothing about the program without the rewards.

The trial had one site and a sample that was 95.3% White. It was open-label, so everyone knew who got what, and no one was followed after 12 weeks.

The original trial was paid for by the National Institute on Drug Abuse and the Wilbur Mills Endowment, with no money from Pear. Pear paid for this analysis. Two authors were Pear employees, one was a Pear consultant, and one sat on Pear's scientific advisory board. Pear staff helped write the paper. Warren Bickel, a coauthor and the trial's principal investigator, disclosed being a principal of HealthSim, LLC and three other companies, advisory roles at two companies, and consulting for three drug makers.

Why it matters

This trial is the evidence behind one of the first prescription apps for opioid use disorder. Pear filed for Chapter 11 bankruptcy in April 2023, and our entry on reSET and Pear covers what happened. The result still matters for anyone weighing digital therapy for opioid use disorder: in this trial, structured therapy content plus medication plus rewards did better than medication plus rewards alone.

For people in treatment, the finding is about adding something to medication, not replacing it. Everyone in this trial was on buprenorphine.

For programs, the paper also shows why contingency management keeps turning up in these products. The authors point out that clinics often skip it for lack of funding, and that the app bundled it into the prescription. When a digital program is pitched to you, ask what the comparison group got, who paid for the analysis, and whether the version tested is the version you would get.