Improving Medical Knowledge and Diagnostic Accuracy of Type-2 Diabetes Mellitus Through Probabilistic GRU-Oriented Problem-Based Learning

In this study, we propose a variant of Problem-based Learning (PBL) in the field of medicine that uses a probabilistic gated recurrent unit to improve the diagnostic and therapeutic outcomes of type 2 diabetes mellitus. Furthermore, when this approach is incorporated into the medical curriculum, students are provided with comprehensive case-based information consisting of patient history, symptoms, lab results, and diabetes. The probabilistic GRU model is trained on a large and general dataset, as it has to deal with the uncertainty of diabetes cases and treatment outcomes. This innovative educational model not only enhances students’ knowledge about diabetes to the level of 30% compared to the traditional approach but also enhances their abilities to diagnose and make the right decision by using the capabilities of simulation that predict real-life situations. This enhancement to the method is a significant improvement in terms of training future healthcare providers as medical students to manage chronic diseases.