JMIR Publications

ABSTRACT

Background:

Sexualized drug use (SDU)—the use of recreational substances before or during sex—is increasingly prevalent among gay, bisexual, and other men who have sex with men (GBMSM) and is linked to HIV, hepatitis C, sexually transmitted infections (STIs), and mental health and substance-related harms. Existing interventions to prevent SDU are limited. Interventions tailored to one’s current stage of change (SOC) are effective in reducing substance use. A hybrid chatbot was developed for this study. The rule-based component ensures that carefully co-designed SOC-tailored interventions are delivered as intended. The AI component will use natural language processing (NLP) to support real-time question-and-answer (Q&A) functions.

Objective:

This randomized controlled trial (RCT) will evaluate whether accessing the hybrid chatbot on top of web-based educational videos is more effective than watching web-based videos alone in reducing SDU in the past 3 months among GBMSM in Hong Kong.

Methods:

This is a partially blinded (outcome assessors and data analysts) and parallel-group RCT. A total of 100 Cantonese-speaking GBMSM aged 18 years or older in Hong Kong will be randomized evenly into either the intervention group or the control group. In the control group, participants will receive links via WhatsApp to access 12 weekly web-based educational videos. In addition to the web-based videos watched by the control group, those in the intervention group will have access to the hybrid chatbot during the 12-week intervention period. The rule-based component of the chatbot will first assess participants’ SOC regarding sexualized drug use and then deliver co-designed SOC-tailored interventions across 12 weekly sessions. The Q&A component, supported by a co-created Q&A database and natural language processing (NLP) functions, will allow users to raise questions and receive real-time responses. Participants will be interviewed at baseline (T0), after completion of the intervention period (T1), and 3 months after T1 (T2). The primary outcome is the presence of SDU in the past 3 months. Secondary outcomes include SOC related to SDU, condomless anal intercourse in the past 3 months, HIV and other sexually transmitted infection (STI) testing uptake in the past 3 months, use of pre-exposure prophylaxis (PrEP) and depressive and anxiety symptoms. Intention-to-treat analyses use generalized estimating equations and mixed-effects models, with sensitivity analyses for missing data.

Results:

: Recruitment commenced in February 2026. By June 2026, 102 participants had been enrolled and completed baseline assessments; of these, 51 (50%) were randomized to the intervention group and 51 (50%) to the control group. The follow-up assessment at T1 started in June 2026. Data collection is expected to be completed in January 2027.

Conclusions:

The findings will contribute to the evidence of the effectiveness of SOC-tailored and chatbot-delivered interventions. If the hybrid chatbot is proven effective, it can be integrated within existing community and non-governmental infrastructure and adapted for other Chinese-speaking GBMSM populations. Clinical Trial: ClinicalTrials.gov NCT07472582; https://clinicaltrials.gov/study/NCT07472582