A Smartphone App Cut Illicit Drug Use by Half Among Homeless People in Recovery

More than 70% of the people who enrolled in a Missouri-based smartphone app for substance use disorder recovery had insecure housing - living in shelters, on the street, in treatment facilities, or staying temporarily in others' homes. These are among the hardest individuals to reach with conventional treatment: mobile, disconnected from stable care environments, and facing overlapping crises that make keeping regular clinic appointments difficult.

A study from Washington University School of Medicine in St. Louis, published in Drug and Alcohol Dependence, examines whether a digital health tool designed for this population actually moves the needle. The answer, within the constraints of a one-month observational study, is measurable and specific.

What the App Does

The app, called uMAT-R, was designed to support recovery from any substance use disorder - opioids, stimulants, hallucinogens - with the specific goal of reducing overdose deaths. It is available on iOS and Android and provides a suite of tools: information about local treatment options and social services, a sobriety tracker, educational content on managing withdrawal and maintaining recovery, and mental health support materials.

The differentiating feature is human contact. Users can message through a chat function with trained coaches - called e-coaches - available Monday through Friday. These are not automated chatbots; they are people trained to provide advice, encouragement, and referrals.

The app emerged from an observation by principal investigator Patricia Cavazos-Rehg, a professor at WashU School of Medicine and School of Public Health: people in recovery communities were already using social media to support each other, asking questions about clinical care and seeking mental health resources in spaces that felt less stigmatized than formal treatment settings.

"I observed how folks use social media venues to support each other, even during times when they were feeling most alone or stigmatized," Cavazos-Rehg said. "They would ask questions on social media about clinical care and seek out resources for mental health help. This inspired me to develop a tool for people who are going through mental health problems or are in recovery support."

Study Design and Key Numbers

The analysis included data from 972 people enrolled in uMAT-R between 2020 and 2024. Of those, more than 70% self-identified as housing-insecure. The study sample consisted of 378 participants who completed both an intake survey and a follow-up survey one month later.

At intake, participants reported substance use patterns, drug cravings, and baseline health literacy - the ability to find and use health information. The surveys also measured two psychosocial factors associated with depression: perceived burdensomeness (feeling like a burden to others) and a sense of not belonging.

After one month:

• App users were approximately 50% less likely to have used non-opioid illicit drugs compared to those who had not logged on.
• Participants who used the e-coach messaging function were approximately 50% less likely to report opioid use than those who did not use it.
• Higher overall app engagement was associated with greater reductions in illicit drug use.
• App use was associated with reduced drug cravings and lower perceived burdensomeness.
• E-coach chat use was associated with improved health literacy and reduced feelings of social isolation and alienation.

Reading the Limitations Carefully

The study design cannot establish causation. People who logged into the app or messaged with coaches may differ systematically from those who did not - perhaps they were more motivated to reduce use from the start, or had different levels of social support, or faced different intensities of housing crisis. The month-long follow-up window is short for a condition characterized by relapse patterns that play out over years.

The sample also represents people who enrolled in the app and completed the follow-up survey - likely a self-selected group more engaged with recovery than those who enrolled and dropped out. The 378-person sample from an initial 972 enrollees reflects significant attrition that the study does not fully account for.

What the study can say is that within a population with very high rates of housing insecurity and co-occurring mental health challenges, those who used the app showed measurably better outcomes on substance use measures at one month than those who did not.

"These results speak to the promise of what mobile health and digital health technologies can bring to the treatment of hard-to-reach populations," said co-author Vijay Eswaran, who is planning a similar study targeting recently incarcerated people with substance use disorders.