Predicting the first smoking lapse during a quit attempt: A machine learning approach.

摘要: Abstract Background Just-in-time adaptive interventions (JITAI) aim to prevent smoking lapse using tailored support delivered via mobile technology in the moments when it is most needed. Effective cessation JITAI rely on development of accurate decision rules that determine someone likely lapse. The primary goal present study was identify strongest predictors first among smokers undergoing a quit attempt. Methods Smokers attending clinic-based program (n = 74) were asked complete ecological momentary assessments five times daily study-provided smartphones for 4 weeks post-quit. A three-stage modeling process utilized Cox proportional hazards regression examine time function 31 predictors. First, univariate models evaluated relationship between each predictor and Second, elastic net machine learning algorithm used select best Third, backwards elimination further reduced set optimize parsimony. Results Univariate identified seven significantly related retained five: perceived odds today, confidence ability avoid smoking, motivation urge smoke, cigarette availability. model demonstrated inadequate approximation non-penalized baseline model. Conclusions Accurate estimation high risk remains an important JITAI. These results demonstrate utility exploratory data-driven approaches variable selection. this can inform future by highlighting targets intervention.