A community-based, pragmatic, cluster-randomized controlled trial of chat-based instant messaging support integrated with proactive brief interventions for smoking cessation

Abstract

Background: Current models of smoking cessation treatment are largely reactive and fail to reach many smokers who are not motivated to quit. Widespread use of mobile instant messaging apps (e.g., WhatsApp) offers an avenue for delivering personalized behavioral support through “chat” (synchronous, interactive messaging). This study aimed to evaluate the effect of chat-based instant messaging support combined with proactive brief interventions on smoking cessation in smokers in the community. Methods: This thesis had two phases. In phase one, a purposive sample of current smokers in Hong Kong were interviewed in a formative qualitative focus group study to explore their perspectives on chat-based instant messaging support for smoking cessation. Data were analyzed by Braun and Clarke’s method of thematic analyses. In phase two, daily cigarettes smokers from 68 community sites in Hong Kong were proactively recruited into a two-arm, pragmatic, cluster-randomized controlled trial. Community sites (n = 68) were randomized into either intervention or control groups (1:1). Participants in both groups received proactive brief advice. Those in the intervention group additionally received chat-based instant messaging support for 3 months and offers of referral to external smoking cessation services. Masking of interventionist and participants were not possible, but statistical analysts were blinded to treatment allocation. The primary outcome was carbon monoxide- and cotinine-validated 7-day point-prevalence abstinence (PPA) at 6 months post-treatment initiation. The key secondary outcome was validated 7-day PPA at 3 months (end of treatment). The main analyses were by intention-to-treat. The trial was registered with ClinicalTrials.gov (NCT03182790). Results: In phase one, 21 current smokers were interviewed in five focus groups. Mobile instant messaging was deemed an acceptable and feasible platform for chat-based support. Perceived utility of the intervention included provision of personalized and tailored cessation advice, enhanced psychosocial support, and identification of “weakness points” for improving motivations to quit. Several suggestions were made to inform the design and content of the intervention. In phase two, 1,185 smokers were recruited and cluster-randomized into the intervention (n = 591; 34 clusters) or control groups (n = 594; 34 clusters). 65.6% were not ready to quit within 30 days at baseline. The intervention group had significantly higher biochemically validated 7-day PPA at 6 months (8.1% vs 5.1%; OR = 1.68; 95% CI 1.03–2.74; P = .040), as was validated abstinence at 3 months (7.6% vs 4.0%; OR = 1.95; 95% CI 1.95–3.22; P = 0.009). Engagement in the chat-based intervention was low (16.8%) but strongly predicted higher validated abstinence as a stand-alone treatment (OR = 5.60; 95% CI 2.13–14.8; P < .001) and when combined with use of smoking cessation services (OR = 9.23; 95% CI 3.42–24.9; P < .001), comparing with use of neither interventions. Conclusions: This trial has found chat-based instant messaging support combined with proactive brief interventions effective in increasing abstinence in a cohort of largely unmotivated smokers––an understudied population. Further studies are warranted to ascertain the utility of chat-based support in other populations and to optimize its design for improving the users’ engagement.