Introducing Glycemic Index to Blood Glucose Simulations and its effect on Food Recommendation Systems

Abstract

The availability of In-Silco blood glucose simulators that accurately depict all aspects of Diabetes Mellitus patient’s daily life is crucial for the development of safe and effective new treatment technologies. Current state of the art simulators focus on modeling a meal’s effect on the glucose-insulin system solely based on the quantity of carbohydrates consumed. However not only the quantity of carbohydrates consumed effects the glycemic response, but also the quality of these consumed carbohydrates.

Regular exercise is an important factor in the treatment of diabetes as it yields many benefits for the patient’s health. However, states of hypoglycemia are common during physical activities which may cause patients to avoid exercising.

This thesis proposes an extension to the currant state of the art blood glucose simulator that introduces food’s glycemic index and its effect on the glucose insulin system. This extension is then used to develop food recommendation systems for type-1 diabetes patients, to recommend the optimal food to keep the blood glucose concentrations in the normoglycemic range during exercise sessions. And a experiment is conducted to determine how knowing food’s glycemic index affects these systems.

The simulations of the proposed extension is shown to correctly capture the effect a food’s glycemic index has on the postprandial blood glucose response as described in literature. And the result of the experiment depict a positive impact of using glycemic index knowledge in food recommendation systems that grows as the length and intensity of the exercise sessions increases.