Development of a personalized blood glucose management smartphone application software based on artificial neural network

• Main Authors: Chun-Su Lin, 林純夙
• Other Authors: 胡仲祺
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/63514972570680010991

博士 === 國立中興大學 === 生物科技學研究所 === 105 === Diabetes mellitus (DM) is a chronic disease that requires constant self-management of health condition by the patients, including monitoring of blood glucose levels, dietary control. However, blood glucose response is highly variable among individuals, and the monitoring is painful, which hinder the adoption of the self-management Apps by the patients. It is therefore highly desirable to develop an App with artificial neural network to predict the blood glucose level based on the personalized diets and blood glucose responses of the patients, which is expected to increase the willingness for adoption and the efficacy of the self-management programs. Thus, the objective of this study is to establish a personalized self-management health support system for diabetic or pre-diabetic patients. Accordingly, a user-centered App has been developed in this study as a result to address the deficiencies in current Apps. The cellphone-based application, Guava Princess’s Blood Glucose Predictor, is developed with user-centered design, including functions such as automatic reminder, database management, personalized data analyses and blood glucose level prediction, as well as suggestions for therapeutic treatments to improve the balance and stability of blood glucose values of the patients. The achievements of this study include the following: (i) identification of the personalized relationship between diets and blood glucose level; (ii) real-time recording of data for reviewing and analyses; (iii) integration of diet images, positioning, and the “google similar images” functions to establish and share personalized dietary records; (iv) prediction of blood glucose level based on dietary data; (v) self-manage dietary contents based on previous meals; (vi) assistance for goal achieving by providing positive feedbacks following the comparison of the numbers of expected and actual measurements; and (vii) amendment of nutritional and self-management education, with the ability to deduce the results to generate suggestions and corrections for future improvements. The App is now freely available at the Google Store, and is expected to bring a healthier life to those with concerns of metabolic syndromes or DM.