(P608) Using Predictive Analytics to Improve CGM Use in the High-Risk Medicare Population

This poster discusses how an AI-powered predictive and interventional platform effectively identified approximately 28 percent of a group of 55,000 Medicare FFS patients at risk of lapsing on CGM therapy. This group was then segmented into defined risk personas. The following two tailored interventions illustrate the striking impact of this platform on Medicare patients’ ability to continue CGM: 1)CGM use improved 46% when patients received messages aligned with provider trust; and 2) Rural patients showed a 30% boost to CGM use when reminded of at-home support options. As a result of applying this AI-powered platform to a Medicare population, more than $10 million in taxpayer dollars have already been saved in unused devices and unnecessary care. An AI-powered predictive and interventional platform proved effective in identifying barriers to CGM use in a large Medicare population. Personalized digital interventions applied to this high-risk population significantly improved patients’ ability to continue CGM therapy. These findings support the use of data-driven, tailored communication strategies to enhance CGM use and diabetes self-management in the high-risk, high-cost Medicare population.