Leveraging Temporal Networks to Develop and Personalize a Digital Intervention for Comorbid Cannabis Use and Binge Eating

Cannabis use (CU) and binge eating (BE) disorders are increasingly prevalent and often co-occur, contributing to poor treatment outcomes when addressed in isolation. CU has more than doubled over the past decade, with up to 20% of individuals reporting annual use. Among those with substance use disorders, up to 39% report disordered eating behaviors, and 16% have a formal eating disorder diagnosis—most commonly binge-spectrum disorders. BE and CU share several dynamic, momentary risk factors, including emotion dysregulation, executive dysfunction, high cognitive restraint, and low distress tolerance. These overlapping risk factors can reinforce each other over time; for example, CU may increase food cravings that lead to BE, while post-binge distress may trigger CU as a coping strategy.

Current treatments typically target CU and BE sequentially, failing to address their shared underlying mechanisms. This fragmented approach reduces overall intervention effectiveness. Just-in-time adaptive interventions (JITAIs), which deliver personalized treatment strategies in real-time based on individual risk profiles, offer a promising solution. While effective JITAIs exist for CU and BE separately, no such intervention currently addresses their co-occurrence, leaving a significant portion of individuals underserved.

This study aims to inform the development of a JITAI tailored to comorbid CU and BE using network analysis; a method that conceptualizes psychopathology as arising from dynamic interactions between symptoms. Thirty adults with comorbid cannabis use disorder and binge-spectrum eating disorders will complete ecological momentary assessment (EMA) surveys four times daily for 28 days. Phase 1 will construct both idiographic and nomothetic network models to identify shared and distinct risk pathways for CU and BE. Phase 2 will evaluate participants’ perceptions of these models’ accuracy and the acceptability of prototype intervention messages. Findings will support development of personalized, real-time digital interventions for individuals with comorbid CU and BE.