The goal of this clinical trial is to evaluate the effectiveness of a Multi-Theory Model (MTM)-based AI agent intervention for smoking cessation in early-stage cancer patients (clinical stage cTNM 0\~II) who currently smoke. The main questions it aims to answer are: Does the AI agent intervention improve the biochemically verified 7-day point prevalence abstinence rate at the 6-month follow-up compared to control groups? Is the AI agent intervention feasible and acceptable for early-stage cancer patients? Researchers will compare the AI agent intervention group to an professional counseling group and a routine health education groupto see if the AI agent yields higher smoking cessation rates and better maintenance of abstinence. Participants will: Be randomly assigned to one of three groups to receive either AI agent support via WeChat, professional counseling via Phone, or routine health education. Interact with the AI agent (if in the intervention group) which provides personalized guidance, emotional support, and resource matching based on the Multi-Theory Model constructs (e.g., participatory dialogue, emotional transformation). Complete questionnaires regarding smoking behavior, nicotine dependence, self-efficacy, and quality of life at baseline and follow-ups (1 week, 1 month, 3 months, and 6 months). Provide exhaled carbon monoxide (CO) and saliva cotinine samples for biochemical verification if they report successful smoking cessation.
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
TREATMENT
Masking
SINGLE
Enrollment
156
An AI agent powered by a Large Language Model with Retrieval-Augmented Generation (RAG). It provides 24/7 personalized smoking cessation support based on the Multi-Theory Model (MTM). Key Functions: Initiation Phase: Participatory dialogue to weigh pros/cons and goal setting to build behavioral confidence. Maintenance Phase: Emotional transformation support, habit tracking (practice for change), and social/physical environment resource matching (e.g., peer support). Dynamic Adaptation: Adjusts content and push frequency based on user interaction and quitting stage.
WeChat-based counseling provided by smoking cessation specialists twice a month, following WHO guidelines.
Biochemically validated 7-Day Point Prevalence Abstinence Rate
Participants are considered abstinent if they self-report having smoked 0 cigarettes (not even a puff) in the past 7 days, confirmed by a biochemical validation of exhaled Carbon Monoxide (CO) concentration \< 4 ppm and saliva cotinine concentration \< 115 ng/ml
Time frame: 6 month follow-up after randomization
Self-Reported 7-Day Point Prevalence Abstinence Rate
The proportion of participants who self-report having smoked no cigarettes in the past 7 days, without biochemical verification at interim time points.
Time frame: 1 week, 1 month, 3 months, and 6-month follow-up
Smoking Reduction Rate
Defined as a reduction in daily cigarette consumption by ≥50% compared to baseline levels.
Time frame: 1 week, 1 month, 3 months, and 6-month follow-up
Change in Smoking Self-Efficacy
Measured using the Smoking Self-Efficacy Questionnaire (SEQ-12). The scale contains 12 items rated on a 5-point Likert scale. Total scores range from 12 to 60, with higher scores indicating greater confidence in the ability to refrain from smoking in various situations.
Time frame: Baseline, 1 week, 1 month, 3 months, and 6-months follow-up
Change in Quality of Life
Health-related quality of life will be assessed using the EuroQol 5-Dimension 5-Level (EQ-5D-5L) questionnaire. The EQ-5D-5L assesses five dimensions of health: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Responses will be converted to an EQ-5D-5L index score using the prespecified value set, with higher scores indicating better health-related quality of life. The outcome will be reported as the change in EQ-5D-5L index score from baseline.
Time frame: Baseline,1 week, 1 month, 3 months, and 6-months follow-up
Total Duration of AI Agent Use
AI agent use will be assessed using automatically recorded backend system logs. The cumulative duration of AI agent use for each participant during the follow-up period will be calculated. The unit of measure is minutes.
Time frame: From randomization through 6 months
Mean AI Agent Response Time
Mean AI agent response time will be calculated using backend system timestamps as the average time between submission of a participant message and generation of the corresponding AI agent response. The unit of measure is seconds.
Time frame: From randomization through 6 months
Mean AI Agent Session Duration
Mean session duration will be calculated from backend system logs as the total duration of AI agent use divided by the number of usage sessions for each participant. The unit of measure is minutes per session.
Time frame: From randomization through 6 months
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.