Millions of U.S. adults living with prediabetes, a high risk state for future type 2 diabetes, do not receive appropriate lifestyle counseling to lower their risk of type 2 diabetes. Mobile health (mHealth) technologies represent a potential scalable solution to address this far-reaching problem. The objective of this project is to compare the real-world effectiveness of a digital diabetes prevention program (dDPP) to standard of care in- person diabetes prevention programs (ipDPPs). This study will test a novel, fully-automated digital health platform (Sweetch Health, Ltd.) that uses artificial intelligence technology to provide just-in-time and adaptive lifestyle change coaching for prediabetic adults. Preliminary evidence from feasibility or observational studies suggests that JITAIs, which are often delivered via smartphone apps by virtue of their ability to provide continuous self-monitoring and feedback, can be effective. However, it is currently not known whether dDPPs that deliver a JITAI are as effective as ipDPPs in improving health outcomes in patients with prediabetes, a susceptible patient population that is positioned to benefit from such an intervention. The overarching goal of this project, therefore, is to compare the effectiveness of the Sweetch digital diabetes prevention program (dDPP) to real-world in-person diabetes prevention programs (ipDPPs) for promoting weight loss, increasing physical activity, and reducing hemoglobin A1C in prediabetic adults. The proposed study addresses an evidence gap in the science of chronic disease prevention and health behavior change and is supported by promising short-term results from a previous pilot trial conducted by our team. Building on our previous study and leveraging the collective expertise of our multidisciplinary study team, we will conduct a randomized controlled trial of 382 overweight/obese, prediabetic adults ages 18-75 with 6 and 12 month follow-up visits: Arm 1 (N=191) will receive the fully automated Sweetch digital health kit (?dDPP? arm) and Arm 2 (N=191) will be referred to a local CDC-recognized ipDPP. Both arms will have physical activity measured serially during the trial using actimetry at baseline and 2 month intervals. We hypothesize that the dDPP will be more effective than the ipDPP for the outcomes of weight loss, physical activity, and lowering of hemoglobin A1C at 6 months, with sustained effects at 12 months. We further hypothesize that the overall engagement and acceptability will be greater in the dDPP, and that the superiority of the dDPP on clinical outcomes will be mediated by higher engagement in this arm. This project will advance chronic disease prevention and behavioral science research by elucidating the extent to which fully-automated digital interventions using artificial intelligence technology can deliver effective, scalable, sustainable, and cost-effective health-promoting behavioral change interventions in high-risk populations. The implications of this fully-automated approach for scalability in diabetes prevention are profound.
Prediabetes, a high risk state for future type 2 diabetes, currently affects nearly 1 in 3 Americans, the vast majority of whom do not receive appropriate lifestyle counseling to lower their risk for type 2 diabetes. In this project, we propose to evaluate the real-world effectiveness of a fully-automated digital diabetes prevention program for weight reduction, promotion of physical activity, and lowering of hemoglobin A1C (a measure of average blood glucose control) compared to standard of care in-person diabetes prevention programs. Since this digital diabetes prevention program uses artificial technology and requires no human component, it would potentially offer a breakthrough in improving population health in a scalable and cost- effective way, helping the tens of millions of affected U.S. patients and hundreds of millions of patients globally living with prediabetes.
