Smart Diabetes Shared Care: Data-Driven AI Diabetic Case Management Model

The contemporary lifestyle and dietary habits have led to a yearly increase in the number of people diagnosed with diabetes, and the irreversible nature of this disease necessitates an increasing number of people to coexist with it. To address this social phenomenon, our government actively promotes and implements various supportive care mechanisms. Among them, the most popular is the co-care model extolled in the literature. This model not only allows patients to support each other but also effectively reduces the incidence of complications. However, since our country began promoting the policy of the "Diabetes Co-care Network" based on this model, the participation rate still needs to be strengthened even after several years. The broad nature of case intake leads to uneven quality of care. Therefore, taking the patients as the center to understand whether they are suitable to join the co-care network, we can accurately and effectively use resources to help diabetic patients. This research achieves the abovementioned purpose through a three-stage experiment design based on machine learning technology. The first stage is training a model of blood glucose stability in people with diabetes ten years after diagnosing people with diabetes-cond step is preparing a prediction model for whether people with diabetes need aggressive treatment. The third stage is training a prediction model for whether people with diabetes need to join the Diabetes Co-care Network. Finally, the performance of the models produced by this study is evaluated by calculating related indicators using a fuzzy matrix. The results of this study hope to ultimately provide more thoughtful and more precise care for people with diabetes.