Using machine learning to identify predictors of imminent drinking and create tailored messages for at-risk drinkers experiencing homelessness

Abstract

Adults experiencing homelessness are more likely to have an alcohol use disorder compared to adults in the general population. Although shelter-based treatments are common, completion rates tend to be poor, suggesting a need for more effective approaches that are tailored to this understudied and underserved population. One barrier to developing more effective treatments is the limited knowledge of the triggers of alcohol use among homeless adults. This paper describes the use of ecological momentary assessment (EMA) to identify predictors of “imminent drinking” (i.e., drinking within the next 4 h), among a sample of adults experiencing homelessness and receiving health services at a homeless shelter. A total of 78 mostly male (84.6%) adults experiencing homelessness (mean age = 46.6) who reported hazardous drinking completed up to five EMAs per day over 4 weeks (a total of 4557 completed EMAs). The study used machine learning techniques to create a drinking risk algorithm that predicted 82% of imminent drinking episodes within 4 h of the first drink of the day, and correctly identified 76% of nondrinking episodes. The algorithm included the following 7 predictors of imminent drinking: urge to drink, having alcohol easily available, feeling confident that alcohol would improve mood, feeling depressed, lower commitment to being alcohol free, not interacting with someone drinking alcohol, and being indoors. The research team used the results to develop intervention content (e.g., brief tailored messages) that will be delivered when imminent drinking is detected in an upcoming intervention phase. Specifically, we created three theoretically grounded message tracks focused on urge/craving, social/availability, and negative affect/mood, which are further tailored to a participant’s current drinking goal (i.e., stay sober, drink less, no goal) to support positive change. To our knowledge, this is the first study to develop tailored intervention messages based on likelihood of imminent drinking, current drinking triggers, and drinking goals among adults experiencing homelessness.

Inferred networks, machine learning, and health data

“…One of our clusters paired homelessness with sex workers [ 61 ], while two paired homelessness with the use of methamphetamine [ 62 ], heroin, or opiates [ 63 ]. Two other clusters were particularly affected by alcoholism, with one predominately recovering from alcohol use [ 64 ], while the other had over 20 alcoholic drinks in the last month [ 65 ]. Most of the individuals in these two clusters have been to self help group meetings such as Alcoholics Anonymous, showing that some participants benefit greatly from such treatment while others may require alternatives.…”

Development and preliminary effectiveness of a smartphone‐based, just‐in‐time adaptive intervention for adults with alcohol misuse who are experiencing homelessness

“…Prior JITAIs for alcohol use have most often been designed for young adults (O'Donnell et al, 2019; Suffoletto et al, 2018; Wright et al, 2018) or domiciled adults (Attwood et al, 2017; Dulin et al, 2014; Gonzalez & Dulin, 2015), both of whom tend to be more stable populations with better access to traditional treatment services. This study is also unique in that real‐time risk assessments of imminent drinking were derived from a machine learning algorithm that incorporated empirically established risk factors for drinking, including urge to drink, negative mood, and social/availability of alcohol (Walters et al, 2021). Prior JITAIs for alcohol use have utilized geographic location (i.e., GPS) (Attwood et al, 2017; Dulin et al, 2014; Gonzalez & Dulin, 2015) or a single factor (e.g., self‐reported drinking or low self‐efficacy to avoid drinking) to determine when to deliver treatment messages (Suffoletto et al, 2018; Weitzel et al, 2007).…”

“…The Smart-T Alcohol trial had three phases. In the first phase previously reported, we collected EMA data from 78 adults experiencing homelessness (mean age [SD] = 46.6 [9.2]; 15.4% women) who reported hazardous drinking (Walters et al, 2021). Participants completed up to five EMAs per day for 4 weeks.…”

“…Like A‐CHESS, Smart‐T Alcohol provides on‐demand motivational messages to help people avoid or reduce drinking. In addition, Smart‐T Alcohol assesses momentary risk factors for imminent drinking via EMA and automatically delivers tailored messages at the end of each EMA based on individual and contextual factors, including current drinking goal (if any), time of day, and imminent risk factors for drinking (Walters et al, 2021). This approach directly targets real‐time influences on alcohol use without having to rely on treatment providers or a sober social support network, neither of which may be present for these individuals.…”

Ecological Momentary Assessment and Personalized Networks in Cognitive Bias Modification Studies on Addiction: Advances and Challenges

“…Machine learning has also been applied to EMA to predict drinking episodes. For example, a machine learning algorithm has been developed to predict drinking episodes in at-risk drinkers (Walters et al, 2021), based on which intervention messages can be delivered.…”