Diabetic Retinopathy Risk Estimation Using Fuzzy Rules on Electronic Health Record Data

摘要

Diabetic retinopathy is an ocular disease that involves an important healthcare spending and is the most serious cause of secondary blindness. Precocious and precautionary detection through a yearly screening of the eye fundus is difficult to make because of the large number of diabetic patients. This paper presents a novel clinical decision support system, based on fuzzy rules, that calculates the risk of developing diabetic retinopathy. The system has been trained and validated on a dataset of patients from Sant Joan de Reus University Hospital. The system achieves levels of sensitivity and specificity above 80%, which is in practice the minimum threshold required for the validity of clinical tests.