Understanding Usage Intention and Behavior for a Health Chatbot: An Integrated Model Applied to Non-Latino, English-preferring Latino, and Spanish-preferring Latino Populations

This study examined factors influencing intentions to use and actual usage of a genetics education chatbot among non-Latino, English-preferring Latino, and Spanish-preferring Latino populations using an integrated Theory of Planned Behavior (TPB) and Health Belief Model (HBM) framework. This integration captures both health-specific perceptions and social influences, providing a more comprehensive account of adoption than either model alone. A cross-sectional survey was conducted with 963 participants stratified by language and ethnicity group, sex, and family history of cancer (FHC). Participants were exposed to a message about a Genetics Education Virtual Assistant (EVA) and completed measures assessing TPB and HBM constructs, usage intentions, and actual behavior, operationalized as choosing to interact with the chatbot immediately. Structural equation modeling revealed the integrated model explained 42.3% of variance in usage intention but only 4.0% in initial engagement behavior, indicating that the framework was more successful in accounting for intention than for actual use. Predictors from both TPB (subjective norm) and HBM (perceived benefits, barriers, susceptibility) significantly predicted intention, which in turn predicted behavior. Spanish-preferring Latinos showed higher intentions and more positive attitudes than non-Latinos but perceived significantly greater barriers. Men reported higher intentions than women across most constructs. FHC was associated with stronger motivational factors through indirect pathways. Given the cross-sectional design, these findings may inform the development and testing of culturally tailored approaches that leverage social influence and perceived benefits while addressing perceived barriers, particularly for Spanish-preferring Latino populations.