Detecting spatiotemporal clusters of tobacco and cannabis use and co-use to inform smartphone-based interventions

Co-use of tobacco and cannabis poses adverse health consequences. Few studies have incorporated spatial analysis into smartphone-based assessments of co-use. This study aims to detect patterns of tobacco and cannabis co-use to inform the delivery of tailored interventions. We analyzed data from 30 young adults aged 18-30 in Northern California who used both tobacco and cannabis. Substance use/nonuse events were collected on participants' smartphones via geographic ecological momentary assessment over 30 days. Substance use events were categorized as tobacco use, cannabis use, or co-use and analyzed using the Space-Time Permutation Scan Statistic to identify distinct spatiotemporal clusters. Cluster characteristics were summarized to further explore use patterns. Substance use events showed clear clustering patterns in space and time, with 76% of events captured within clusters (interquartile range = 69%-87%). The number of tobacco use clusters peaked in the afternoon and dropped in the evening, while cannabis and co-use clusters peaked in the evening. Home was consistently a prevalent location. Additionally, tobacco use was clustered at others' homes (42.9%) in the early morning and then shifted to work (34.5%). Cannabis use clustered at work (17.6%) in the morning. Co-use clustered at others' homes (18.2%) in the early morning and then shifted to work (18.2%). This study identified individual patterns of tobacco and cannabis use on an event-level basis. Findings can inform the development of smartphone-based interventions that use contextual data to align with established co-use routines and intervene at the most likely times and locations of use. (PsycInfo Database Record (c) 2025 APA, all rights reserved).