Research Project 1: Navigation and Artificial Intelligence Technology for Indigenous Virtual Education on smoking cessation (NAITIVE)

PROJECT SUMMARY/ABSTRACT American Indian/Alaska Native (AIAN) people have the highest rates of commercial cigarette smoking of any racial/ethnic group in the US. Compared to other racial/ethnic groups, they have 6 times higher rates of developing smoking-related cancers and are only half as likely to quit smoking. The result is that commercial cigarette smoking now accounts for half of all deaths among AIAN communities nationwide. These dramatic inequities are mainly due to two causes: (1) AIAN peoples’ lack of access to smoking cessation interventions, and (2) the lack of efficacious smoking cessation interventions for AIAN populations. Addressing the dual challenges of access and efficacy requires high-impact population-level interventions for smoking cessation. Recent advances in large language model (LLM)-based smartphone app-delivered conversational chatbots offer a promising solution, leveraging machine learning and natural language processing to emotionally engage users and to provide personalized support and skills training. Chatbots can deliver supportive, empathic, and personalized responses. To date, nothing is known about the efficacy and underlying mechanisms of chatbots for helping AIAN people quit cigarette smoking. Regarding accessibility, chatbots require no in-person delivery and no provider training, do not require integration into complex medical systems, can be freely accessed through an app store, and are available at any time and any place. These smartphone-delivered chatbot apps have potentially high population-level reach given that 68-78% of AIAN people own smartphones. We developed “QuitBot” using an agile, user-centered design framework. QuitBot, the first-known chatbot with LLM-supported conversation features, offers a comprehensive 42-day smoking cessation program. We initially assessed QuitBot through a diary study, followed by a pilot RCT (N = 404), comparing it with the National Cancer Institute (NCI)’s SmokefreeTXT program. The pilot RCT proved highly feasible with a 96% retention rate. QuitBot participants interacted more frequently with their intervention and used it longer than SmokefreeTXT. Building on our preliminary data, we propose to conduct a fully powered two-arm RCT comparing QuitBot (n = 386) to SmokefreeTXT (n = 386) among nationally recruited AIAN participants in order to determine: (1) the efficacy of QuitBot relative to SmokefreeTXT for biochemically verified 30-day point prevalence commercial cigarette abstinence at 12 months post-randomization; (2) if QuitBot’s effect on 12-month smoking cessation is mediated by therapeutic alliance and engagement. We will also (1) explore if key baseline factors moderate the primary cessation outcome and (2) conduct qualitative interviews with QuitBot participants and Tribal partners, Tribal organizations, urban Indian partner organizations, and other relevant organizations and policymakers in tobacco control to identify barriers and facilitators to dissemination. This study will advance the fields of research on the use of chatbots for smoking cessation and cancer prevention among the AIAN population nationwide.