Abstract
<title>Abstract</title> <p> <bold>Background</bold> Despite potential benefits of delivering digital health interventions (DHIs), engagement rates remain low. This study adopted the Digital Health Engagement Model to investigate how intention mediates the influence of behavioural determinants on caregivers’ engagement with ParentText, a parenting chatbot designed to promote positive parenting and improve adolescent health. <bold>Methods</bold> Data included caregiver-reported baseline demographics (N = 1,034 caregivers of adolescent girls ages 10-17), intentions (to use and spend mobile data on the chatbot), and behavioural determinants collected at baseline. Engagement data were digitally tracked and recorded from August to December 2023. Structural Equation models were conducted to investigate the mediator effects of pre-intervention intentions in the relationship between baseline behavioural determinants and engagement. <bold>Results</bold> Results from structural equation models indicate that more favourable behavioural determinants predict higher enrolment (OR = 1.06, 95% CI: 1.01 –1.10), higher module (β= 1.69, 95% CI: 0.47–2.91), goal (β = 1.30, 95% CI: 0.37–2.23) and more days interacting with the chatbot (β = 1.38, 95% CI: 0.30–2.46). There was no evidence of mediation by intentions. Despite limited internet access, caregivers with stronger intentions to use the chatbot were more likely to enrol. Despite analyses indicating behavioural determinants as a predictor of engagement, model fit results suggests that variation in caregivers’ engagement with the chatbot is influenced by unmeasured factors rather than behavioural determinants or intention outcomes. <bold>Conclusions</bold> This is the first study to employ a behavioural science approach in understanding user engagement in a chatbot-led parenting programme. Our results highlight the importance of developing low-bandwidth DHIs to optimally engage digitally excluded populations, as well as adopting a qualitative approach to identify the factors driving user engagement in DHIs. Our findings further underscore the relevance of longitudinal research investigating the temporal stability of intention to better understand its predictive significance for user engagement in digital parenting interventions. <bold>Trial Registration:</bold> Open Science Framework (OSF); https://doi.org/10.17605/OSF.IO/WFXNE Clinical Trial Number: not applicable. </p>