Combined model for diabetes lifestyle support

Area of Science:

• Biomedical Engineering
• Endocrinology
• Computational Biology

Background:

• Diabetes mellitus poses a significant public health challenge, with current insulin dosage estimation methods being experience-based and inefficient.
• Existing mixed meal models for diabetes management lack detailed nutritional analysis, limiting their practical application.

Purpose of the Study:

• To develop and evaluate a novel computational model for predicting blood glucose levels in individuals with diabetes.
• To improve the accuracy of insulin dosage recommendations by integrating detailed nutritional information and advanced modeling techniques.

Main Methods:

• A new predictive model was created by combining established nutrient absorption and glucose control models.
• The model incorporates detailed nutritional composition (protein, lipid, monosaccharide, fiber, starch) from expert dietary systems.
• Parameter training was performed using genetic algorithms to optimize patient-specific glucose level tracking.

Main Results:

• The combined model effectively tracks blood sugar levels, considering nutrition, insulin, and initial glucose levels.
• The model demonstrates improved performance compared to existing mixed meal models, especially with detailed nutritional data.
• Genetic algorithm-based parameter training significantly enhanced the model's accuracy in tracking individual patient glucose levels.

Conclusions:

• The proposed integrated model offers a more precise approach to predicting glucose levels in diabetes management.
• Detailed nutritional analysis and advanced parameter optimization, such as through genetic algorithms, are crucial for improving diabetes care.
• This approach has the potential to enhance the efficiency and effectiveness of diabetes mellitus treatment.