GluPredKit: Development and User Evaluation of a Standardization Software for Blood Glucose Prediction

TL;DR

GluPredKit tackles the critical problem of non-standardized BG prediction research by delivering an open-source, modular platform that standardizes data handling, model training, evaluation, and cross-model comparisons. The authors implement a CLI-driven workflow, integrate multiple data sources and prediction models, and reproduce benchmark results to demonstrate consistency with prior work. A user study with four participants shows high usability (average SUS = 86) and qualitative feedback emphasizing documentation quality and potential improvements, including a GUI and broader data-source support. This work advances reproducibility and comparability in BG prediction, supporting collaborative development and educational use, with future work focused on expanding algorithms, user interfaces, and real-world validations.

Abstract

Blood glucose prediction is an important component of biomedical technology for managing diabetes with automated insulin delivery systems. Machine learning and deep learning algorithms hold the potential to advance this technology. However, the lack of standardized methodologies impedes direct comparisons of emerging algorithms. This study addresses this challenge by developing GluPredKit, a software platform designed to standardize the training, testing, and comparison of blood glucose prediction algorithms. GluPredKit features a modular, open-source architecture, complemented by a command-line interface, comprehensive documentation, and a video tutorial to enhance usability. To ensure the platform's effectiveness and user-friendliness, we conducted preliminary testing and a user study. In this study, four participants interacted with GluPredKit and provided feedback through the System Usability Scale (SUS) and open-ended questions. The findings indicate that GluPredKit effectively addresses the standardization challenge and offers high usability, facilitating direct comparisons between different algorithms. Additionally, it serves an educational purpose by making advanced methodologies more accessible. Future directions include continuously enhancing the software based on user feedback. We also invite community contributions to further expand GluPredKit with state-of-the-art components and foster a collaborative effort in standardizing blood glucose prediction research, leading to more comparable studies.

Figures (5)

• Figure 1: High-level visualization of the GluPredKit ecosystem. The upper image symbolize data storage and how GluPredKit acquires its data, while the bottom graph exemplifies the potential output of BG predicted trajectories in GluPredKit. The trajectories represent predictions for 120 minutes. Predictions are cut off when there are no more measurements to compare with.
• Figure 2: Methodological flowchart for the study.
• Figure 3: Overview of the GluPredKit pipeline. This diagram delineates the sequential stages involved in the BG prediction process, starting from data parsing from existing datasets, through preprocessing and model training, culminating in model evaluation and real-time predictions.
• Figure 4: Overview of the CLI commands and how they interact with the file structure in GluPredKit.
• Figure 5: Bar chart of individual SUS scores alongside a line indicating the average SUS score.