Efficacy of a Smartphone-Based JITAI to Reduce Drinking Among Adults Experiencing Homelessness

Overview

Approximately 1 out of every 3 adults experiencing homelessness has a current alcohol use disorder (AUD). Alcohol use plays a central role in increased morbidity and mortality in this population, with some studies suggesting it plays a direct role in as many as 17% of deaths. Alcohol is also a leading cause of homelessness and contributes to its chronicity by inhibiting progress toward obtaining employment and a stable living environment. Shelter-based treatments are common, but compliance is often poor. There are many documented challenges to effectively implementing traditional evidence-based interventions in shelter settings. In the current project, we will conduct a randomized controlled trial. Adults currently experiencing homelessness who report hazardous alcohol use (N = 600) will be recruited from five shelters across the Dallas, TX and Oklahoma City, OK metropolitan areas. Individuals will attend screening and training visits, then complete one week of self-monitoring (via smartphone app) before being randomized to receive either standard shelter-based treatment (Usual Care; UC), or the Smart-T Alcohol intervention + UC. The Smart-T Alcohol intervention offers on-demand content and automated contextually tailored messages to reduce alcohol consumption. The intervention period will span eight-weeks (i.e., weeks 2-9), with follow-up assessments occurring at weeks 9, 13, and 26. All participants will complete brief daily smartphone-based surveys for 13 weeks. Aims of the project will be to assess the effects of the Smart-T Alcohol intervention on alcohol use and alcohol-related problems (Aim 1), intervention effects on key drinking risk indicators and their mediation of use outcomes (Aim 2), and identify specific subpopulations (e.g., women, racial/ethnic minorities, younger adults; Exploratory Aim 3) for which the intervention is most effective. Findings will provide a rigorous evaluation of the efficacy of the Smart-T Alcohol intervention and guide future smartphone-based interventions for this population.

The study will test the efficacy of a novel smartphone-based alcohol use reduction intervention (i.e., Smart-T Alcohol) relative to Usual Care in an RCT of heavy-drinking adults experiencing homelessness (N = 600). Participants will complete a screening visit on-site at one of our partnering shelters. Eligible participants will receive training in our smartphone-based assessment platform and may be provided with a compatible device if they do not own one. All participants will use the app to complete four brief (1-2 minute) daily EMAs for 1 week to obtain baseline drinking levels and psychosocial drinking risk factors (e.g., drinking urge, mood, alcohol availability, abstinence motivation). Individuals that complete at least 50% of the daily surveys will then be told if they were randomized to either standard shelter-based treatment (Usual Care; UC) or UC + Smart-T Alcohol intervention (1:1 allocation). The Smart-T Alcohol Intervention will last for eight weeks. An additional 4 weeks of EMA following treatment completion will be used to determine if intervention effects are maintained in the 4-weeks following the end of treatment. Follow-up assessments will occur via Insight(TM) app, at shelters, or via phone at end-of-treatment (i.e., 9 weeks post-enrollment), as well as 13- and 26-weeks post-enrollment. Participants in both groups will complete daily EMAs for a total of 13 weeks (see participant flow in Figure 2). Considering both established recommendations for treatment trials of AUDs and our emphasis on a harm reduction approach for this population, our primary efficacy outcome (Aim 1) will be total alcohol consumption, with # of drinking days and # of heavy drinking days serving as secondary outcomes. Given our preliminary study revealed EMA reports had substantially better agreement with objective markers of alcohol use compared to TLFB, EMA will be used to calculate primary and secondary outcomes. However, analyses will be repeated using TLFB-based outcomes (i.e., 9, 13, and 26-weeks) to facilitate direct comparisons to prior research and examine intervention effects. Additional analyses will examine intervention effects on key drinking risk factors per our conceptual framework (i.e., drinking urge, mood, alcohol availability, abstinence motivation) and whether these effects mediate treatment outcome (Aim 2), as well as subpopulations for whom the intervention may be more or less effective (e.g., women, younger adults, those with comorbid mental health or substance use problems, individuals engaging more heavily with shelter services; Exploratory Aim 3). Interventions. Over the past 9 years, Dr. Businelle (MPI) has led the development of the InsightTM mHealth Platform. This platform enables researchers to rapidly create innovative smartphone-based studies that include EMA and just-in-time adaptive interventions (JITAI) in a matter of weeks. The platform has supported 113 research studies to date (60 of these studies received funding from the NIH). The InsightTM platform was used to create the Smart Treatment (Smart-T) suite of apps, including the Smart-T Alcohol app that we tested in our recently completed NIAAA-funded R34 study. All participants will receive a version of the InsightTM app (UC or Smart-T Alcohol) on their personal phone or a study phone that is loaned to them. Usual Care (UC). We will carefully document engagement with AUD treatments (shelter and non-shelter based) via EMA each day for examination in exploratory moderator analyses (Aim 3). All participants at the recruitment sites have access to 12-step support groups (e.g., Alcoholics Anonymous), intensive or supportive outpatient services, and medical and/or psychiatric care. Though initially conceived as part of the Smart-T Alcohol intervention, for ethical reasons individuals in either group who report heavy drinking episode(s) will receive feedback about the negative health effects of heavy drinking and those who report high-intensity drinking (i.e., Women: 8 drinks/day; Men: 10 drinks/day) will be immediately referred for additional treatment services via a designated shelter referral/contact (i.e., shelter contacts will receive encrypted emails that state when study participants report high-intensity drinking). All five shelters conduct a needs assessment as part of their intake process. Access to supportive services (e.g., computer access, laundry facilities, physical and mental healthcare) are also provided as part of standard shelter services either in-house or through partnerships with local organizations. Usual Care + Smart-T Alcohol. Participants in this condition will receive UC and gain access to the Smart-T Alcohol intervention, which includes two core components: 1) an on-demand "Tips" function/button; and 2) an algorithm that uses EMA responses to assess current risk for alcohol use and automatically delivers tailored messages to help participants achieve their daily drinking goal (i.e., stay sober, cut down on drinking, no drinking goal). On days when there is an abstinence or reduction goal, the app will provide tailored treatment messages that aim to help participants achieve their chosen goal. On days when there is no drinking goal, the app's gain-framed messages will focus on increasing motivation to set a reduction or abstinence goal. The app has been customized using iterative, user-centered design procedures to increase clarity and helpfulness of messages and ensure app usability in this population.