Phenotypes vs Processes: Understanding the Progression of Complications in Type 2 Diabetes. A Case Study

Patients with type 2 diabetes mellitus are at high risk of developing several types of complications. Previous studies show a clear association of tight glycemic control, both in terms of glycated hemoglobin and the time in range, with different metabolic pathways that contribute to nervous, micro, and macro-vascular complications. As a matter of fact, a few works attempted to obtain phenotyping to describe the evolution of type 2 diabetes patients; however, due to the complexity and variability of the disease progression, together with factors related to the process of care focused on medical specialties, made it difficult to obtain clinically meaningful processes to understand the progression of the disease. In this work, we apply process mining techniques to automatically detect temporally restricted patterns based on routinely collected clinical data to define possible pathways and disease progression process models to support clinicians in understanding the process of care in type 2 diabetes mellitus. Our work is focused on the most prevalent complications, cardiovascular and nephropathy, based on a retrospective cohort of 16,187 patients during eight years of follow-up and demonstrates the ability of process mining approaches to identify disease progression patterns.