Leveraging Mobile Phone Sensors, Machine Learning, and Explainable Artificial Intelligence to Predict Imminent Same-Day Binge-drinking Events to Support Just-in-time Adaptive Interventions: Algorithm Development and Validation Study

Abstract

Background Digital just-in-time adaptive interventions can reduce binge-drinking events (BDEs; consuming ≥4 drinks for women and ≥5 drinks for men per occasion) in young adults but need to be optimized for timing and content. Delivering just-in-time support messages in the hours prior to BDEs could improve intervention impact. Objective We aimed to determine the feasibility of developing a machine learning (ML) model to accurately predict future, that is, same-day BDEs 1 to 6 hours prior BDEs, using smartphone sensor data and to identify the most informative phone sensor features associated with BDEs on weekends and weekdays to determine the key features that explain prediction model performance. Methods We collected phone sensor data from 75 young adults (aged 21 to 25 years; mean 22.4, SD 1.9 years) with risky drinking behavior who reported their drinking behavior over 14 weeks. The participants in this secondary analysis were enrolled in a clinical trial. We developed ML models testing different algorithms (eg, extreme gradient boosting [XGBoost] and decision tree) to predict same-day BDEs (vs low-risk drinking events and non-drinking periods) using smartphone sensor data (eg, accelerometer and GPS). We tested various “prediction distance” time windows (more proximal: 1 hour; distant: 6 hours) from drinking onset. We also tested various analysis time windows (ie, the amount of data to be analyzed), ranging from 1 to 12 hours prior to drinking onset, because this determines the amount of data that needs to be stored on the phone to compute the model. Explainable artificial intelligence was used to explore interactions among the most informative phone sensor features contributing to the prediction of BDEs. Results The XGBoost model performed the best in predicting imminent same-day BDEs, with 95% accuracy on weekends and 94.3% accuracy on weekdays (F1-score=0.95 and 0.94, respectively). This XGBoost model needed 12 and 9 hours of phone sensor data at 3- and 6-hour prediction distance from the onset of drinking on weekends and weekdays, respectively, prior to predicting same-day BDEs. The most informative phone sensor features for BDE prediction were time (eg, time of day) and GPS-derived features, such as the radius of gyration (an indicator of travel). Interactions among key features (eg, time of day and GPS-derived features) contributed to the prediction of same-day BDEs. Conclusions We demonstrated the feasibility and potential use of smartphone sensor data and ML for accurately predicting imminent (same-day) BDEs in young adults. The prediction model provides “windows of opportunity,” and with the adoption of explainable artificial intelligence, we identified “key contributing features” to trigger just-in-time adaptive intervention prior to the onset of BDEs, which has the potential to reduce the likelihood of BDEs in young adults. Trial Registration ClinicalTrials.gov NCT02918565; https://clinicaltrials.gov/ct2/show/NCT02918565

Alcohol and the Brain–Gut Axis: The Involvement of Microglia and Enteric Glia in the Process of Neuro-Enteric Inflammation

“…At the 46th Research Society of Alcoholism-which was held from 24-28 June 2023 in Bellevue, USA-Dr. George F. Koob, Director of the NIAAA, mentioned during his opening lecture that acute and chronic alcohol consumption not only affects the brain, by causing cognitive and neurological disorders [10,11], but also affects every organ of the body, leading to arrhythmia, cardiovascular disease, hepatitis, and pancreatitis, and especially gut dysbiosis [12][13][14][15][16][17]. Binge alcohol consumption is characterized by 3-5 drinks in a short time of 2-3 h, whereas chronic alcohol consumption is considered as the daily intake of alcohol for several weeks [18]. Binge 2 of 18 drinking is more common in adolescents, especially among college students.…”

Acceptability of Personal Sensing Among People With Alcohol Use Disorder: Observational Study

“…More recently, research using personal sensing of raw data streams other than self-reporting is emerging for mental health, including alcohol and other substance use disorders. This includes methods to sense geolocation [ 13 - 16 ], cellular communication [ 14 - 16 ], sleep [ 17 ], and physiology [ 15 , 16 , 18 ], for example. These alternative personal sensing methods provide benefits and opportunities that are not possible with EMA.…”

“…These methods can also be used to monitor psychiatric symptoms or even predict the future risk of symptom recurrence or other harmful behaviors (eg, suicide attempts and risky or otherwise harmful drinking episodes) [ 24 - 27 ]. For alcohol and other substance use disorders, there is emerging research on using sensed data to predict craving [ 13 , 18 ]; alcohol [ 15 , 27 - 29 ], cannabis [ 16 ], or opioid use [ 14 ]; and lapses or relapse [ 14 , 30 , 31 ]. Personal sensing measures or risk indicators may be shared, with patient consent, to health care providers to allow for cost-effective, targeted allocation of limited mental health resources to patients with the greatest or most urgent need [ 32 ].…”

“…This research has predominantly involved proof-of-concept studies that typically include only healthy controls or other convenience samples rather than people with psychiatric disorders [ 16 , 17 ]. It has also often used very small sample sizes or short monitoring periods [ 15 , 16 , 18 ]. Recent reviews of this emerging literature have highlighted gaps in reporting on participant exclusions, attrition, and adherence that are necessary to assess selection biases and the feasibility of these novel personal sensing methods [ 37 - 39 ].…”

Behavioral health and generative AI: a perspective on future of therapies and patient care

“…These technological progress have led to a number of applications that address an extensive range of challenges encountered in healthcare. Such applications have introduced improvements in diagnostic accuracy and treatment planning [2][3][4] , have personalized medical interventions [5][6][7] , expedited drug discovery and development 8 , and broadened the reach of medical services through remote monitoring and telemedicine 9,10 .…”