A context-adaptive smoking cessation system using videos

Cigarette smoking is the primary preventable cause of death and disease worldwide. Studies reveal that smoking is associated with psychiatric symptoms, sociodemographic characteristics, social stressors, and lack of social support. In general, smokers report poorer mental health and benefit from support to be able to quit smoking (Jorm et al., 1999). In this paper, a tailored smoking cessation system has been developed in which the counseling and support is delivered via video-messaging. The system engages users in adaptive motivating video access. Users can interact with the system and the system selects the best matching video for them by processing their messages using Natural Language Processing (NLP). We have tailored 77 videos for interactive contents that encompass important issues users might face during the process of smoking cessation. A novel application-based data driven approach has been taken for categorizing videos to push to participants. The approach is based on analyzing 750 messages of people in the cessation process. We observed that most of the messages’ contents were about smoking health effects, cravings, triggers, relapse, positive mood, low cessation self efficacy, medications, and culturally specific targeting inquiries. Considering these categories, videos are categorized to the corresponding groups by an intelligent approach. The information underlying the data driven categories allows for improving and facilitating smoking status assessment. The system has the potential for improving future smoking cessation decision-making adaptive interventions and health monitoring systems. The goal is to tailor the system to meet the needs of the users in real-time and maximize the potential impact.

Duke Scholars

Author Ming-Chun Huang DKU Faculty