Mobile Application Development for Gestational Blood Sugar Prediction

Abstract

Gestational Diabetes Mellitus (GDM) is a critical health situation that endangers pregnant women, requiring regular blood sugar monitoring to ensure both mother and fetus well-being. Many GDM patients are unaware and unable to determine their status in view of limited medical personnel and high cost of laboratory tests. This paper gave the design process of a Gestational Blood Sugar Tracker (GBST) smartphone application aimed at assisting pregnant women with predicting, managing and regulating their blood sugar. The GBST mobile app was implemented with Android Studio, the primary programming language being Java, the app uses SQLite for local database and firebase authentication. Integrated in the app is a fast forward neural network-based prediction model to predict the GDM status. The mobile app had 12 screens designed in Figma that allow data capture and result display. The FNN model with 77% accuracy was implemented for prediction of GDM status. It was recommended that the GBST app should be evaluated using the Mobile Applications Rating Scale (MARS).