Diabetic foot ulcers risk prediction in patients with type 2 diabetes using classifier based on associations rule mining

Abstract

AbstractIdentifying diabetic patients at risk of developing foot ulcers, as one of the most significant complications of diabetes, is a crucial healthcare concern. This study aimed to develop an associative classification model (CBA) using the Apriori algorithm to predict diabetic foot ulcers (DFU). This retrospective cohort study included 666 patients with type 2 diabetes referred to Shahid Beheshti Hospital in Iran between April 2020 and August 2022, of which 279 (42%) had DFU. Data on 29 specific baseline features were collected, which were preprocessed by discretizing numerical variables based on medical cutoffs. The target variable was the occurrence of DFU, and the minimum support, confidence, and lift thresholds were set to 0.01, 0.7, and 1, respectively. After data preparation and cleaning, a CBA model was created using the Apriori algorithm, with 80% of the data used as a training set and 20% as a testing set. The accuracy and AUC (area under the roc curve) measure were used to evaluate the performance of the model. The CBA model discovered a total of 146 rules for two patient groups. Several factors, such as longer duration of diabetes over 10 years, insulin therapy, male sex, older age, smoking, addiction to other drugs, family history of diabetes, higher body mass index, physical inactivity, and diabetes complications such as proliferative and non-proliferative retinopathy and nephropathy, were identified as major risk factors contributing to the development of DFU. The CBA model achieved an overall accuracy of 96%. Also, the AUC value was 0.962 (95%CI 0.924, 1.000). The developed model has a high accuracy in predicting the risk of DFU in patients with type 2 diabetes. The creation of accurate predictive models for DFU has the potential to significantly reduce the burden of managing recurring ulcers and the need for amputation, which are significant health concerns associated with diabetes.