When doctors are not present: Understanding patients trying generative AI in the post-prescription stage.

Background: When physicians are not physically present in patients' everyday illness management, individuals must continue making health-related decisions after receiving medical prescriptions. In these between-visit contexts, generative artificial intelligence (GenAI) is increasingly used as a supplementary source of health information and guidance. However, the psychological mechanisms driving user engagement and subscription intention in healthcare-oriented GenAI remain underexplored. Methods: Guided by the Theory of Trying, this study examined how digital self-efficacy and health-AI skepticism are associated with attitudestoward success, failure, and the process of trying in health-related GenAI interactions. Survey data from 634 active ChatGPT users who reported using it for health-related queries were analyzed using component-based partial least squares structural equation modeling (PLS-SEM). Results: Attitude toward the process of trying emerged as the strongest predictor of user engagement, followed by attitude toward success. Attitude toward failure did not predict engagement and, contrary to the hypothesized relationship, was positively associated with subscription intention. Health-AI skepticism was negatively associated with attitude toward success and the process of trying while being positively associated with anticipatory concerns about possible failure. Conclusions: In post-prescription and between-visit health management contexts, the findings highlight the importance of process-oriented evaluations alongside success-related evaluations in understanding engagement and subscription intention. These results clarify distinct motivational pathways associated with health-related GenAI use and offer implications for the design of supportive and ethically responsible GenAI systems.