Stanford Seminar - How Design Decisions Impact the Effectiveness of Digital Health Interventions

January 24, 2025 Pedja Klasnja, University of Michigan When and How Design Matters: Investigating How Design Decisions Impact the Effectiveness of Digital Health Interventions Digital health interventions that are based on health-behavior theory are, on average, more effective than interventions that are developed without a robust use of theory. But differences in effectiveness abound even among theory-based interventions. There is growing evidence, both from qualitative research and clinical trials, that how interventions are designed plays a major role in their effectiveness. Of course, that 'design matters' is a commonplace in HCI, but exactly how and when design decisions impact a system's effectiveness is often left unspecified. In this talk, I will review some of the recent research that my colleagues and I have been doing to investigate how differences in design impact the effectiveness of digital health interventions. By drawing on causal modeling and optimization study designs like micro-randomized trials and factorial experiments, we can demonstrate differences in the effectiveness of different designs of an intervention component that implements the same behavior-change technique, such as goal-setting or planning, and articulate specific, testable causal hypotheses for why these differences occur. This type of research, I suggest, can provide guidance for when design is likely to have a significant impact on how well an intervention can support a health behavior like physical activity, and it can build evidence for how best to design digital interventions for different populations and different contexts. About the speaker: Predrag 'Pedja' Klasnja is an Associate Professor in the School of Information at the University of Michigan. He focuses on the design and optimization of novel mHealth technologies for health behavior change. He is particularly interested in the design and evaluation of just-in-time adaptive interventions (JITAIs), interventions th