Evaluating the Dietary Factors Most Closely Associated with Diabetes Mellitus Using a Decision-Making Tree Algorithm

Introduction: The development of type ۲ diabetes mellitus (T۲DM) is associated with lifestyle factors, including dietary patterns. A diet rich in macro- and micronutrients has been reported to reduce the risk of T۲DM. Therefore, this study aimed to identify the dietary factors most closely associated with T۲DM in subjects within the MASHAD cohort using a decision tree algorithm. Methods: This cross-sectional study was conducted on ۹۷۰۴ individuals from the Mashhad Stroke and Heart Atherosclerotic Disorders (MASHAD), of whom ۵۹۳۶ participants completed a ۲۴h dietary recall questionnaire. Macronutrients and micronutrients were estimated using Diet Plan ۶ software. A decision tree algorithm was utilized to evaluate the most crucial dietary nutrient intakes concerning T۲DM. Results: The algorithm showed a high specificity (۸۱.۳۴%) but low sensitivity (۳۴.۲۱%), which could predict T۲DM with a low-to-moderate diagnostic ability (AUC=۰.۵۸). Based on the decision tree, eight features, including dietary potassium, total sugar, sucrose, riboflavin, thiamin, sodium, total nitrogen, and magnesium, were T۲DM’s most critical dietary components. Conclusion: Based on the results, consuming sugar, salt, and vitamin B was the most critical related dietary intake to T۲DM. Dietary interventions may be a cost-effective strategy for preventing T۲DM.

Keywords:

Diabetes Mellitus , Nutrients , Diet , Cohort Studies