Evaluation of Factors Affecting Neuropathy in Patients With Type 2 Diabetes Using Artificial Neural Networks

Abstract

Introduction: Neuropathy is a common and debilitating complication in type 2 diabetes, affecting quality of life and increasing healthcare costs. Identifying risk factors is essential for early intervention and management. This study aims to evaluate the factors influencing the occurrence of neuropathy in patients with type 2 diabetes using artificial neural networks. Methods: In this cohort study, data from 371 patients with type 2 diabetes from Fereydunshahr, Iran, were analyzed over a 12 -year follow-up period. Participants were selected based on diabetes screenings conducted in 2008 and 2009. Artificial neural networks with varying architectures were trained and validated, and their performance was compared to logistic regression models using receiver operating characteristic (ROC) curve analysis. Results: The prevalence of neuropathy in this cohort study was 31.2. The best-fitted artificial neural network and logistic regression model had area under the curve (AUC) values of 0.903 and 0.803, respectively. Significant risk factors identified included gender, race, family history of diabetes, type of diabetes treatment, cholesterol levels, triglyceride levels, high -density lipoprotein (HDL) levels, and duration of diabetes. Notably, women, patients with a family history of diabetes, and those using injectable or combined injectable and oral medications were at higher risk of developing neuropathy. Conclusion: These findings highlight the importance of vigilant monitoring and proactive management of neuropathy risk factors, especially in women, patients with a family history of diabetes, and those using injectable or combined diabetic medications.