Detecting Receptivity for mHealth Interventions in the Natural Environment

Author ORCID Identifier

Document Type


Publication Date


Publication Title



Department of Computer Science


Just-In-Time Adaptive Interventions (JITAI) have the potential to provide effective support for health behavior by delivering the right type and amount of intervention at the right time. The timing of interventions is crucial to ensure that users are receptive and able to use the support provided. Previous research has explored the association of context and user-specific traits on receptivity and built machine-learning models to detect receptivity after the study was completed. However, for effective intervention delivery, JITAI systems need to make in-the-moment decisions about a user's receptivity. In this study, we deployed machinelearning models in a chatbot-based digital coach to predict receptivity for physical-activity interventions. We included a static model that was built before the study and an adaptive model that continuously updated itself during the study. Compared to a control model that sent intervention messages randomly, the machine-learning models improved receptivity by up to 36%. Receptivity to messages from the adaptive model increased over time.




Short version of the full paper, published earlier in the Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT/UbiComp), volume 5, number 2, article 74, 24 pages. ACM, June 2021. doi:10.1145/3463492.

Original Citation

Varun Mishra, Florian Künzler, Jan-Niklas Kramer, Elgar Fleisch, Tobias Kowatsch, and David Kotz. Detecting Receptivity for mHealth Interventions in the Natural Environment. GetMobile, volume 27, number 2, pages 23–28. ACM, June 2023. doi:10.1145/3614214.3614221.