Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GlucoBench: Curated List of Continuous Glucose Monitoring Datasets with Prediction Benchmarks

Renat Sergazinov, Elizabeth Chun, Valeriya Rogovchenko, Nathaniel Fernandes, Nicholas Kasman, Irina Gaynanova

TL;DR

A comprehensive resource that provides a consolidated repository of curated publicly available CGM datasets to foster reproducibility and accessibility, a standardized task list to unify research objectives and facilitate coordinated efforts, and a set of benchmark models with established baseline performance to objectively gauge new methods' efficacy are presented.

Abstract

The rising rates of diabetes necessitate innovative methods for its management. Continuous glucose monitors (CGM) are small medical devices that measure blood glucose levels at regular intervals providing insights into daily patterns of glucose variation. Forecasting of glucose trajectories based on CGM data holds the potential to substantially improve diabetes management, by both refining artificial pancreas systems and enabling individuals to make adjustments based on predictions to maintain optimal glycemic range.Despite numerous methods proposed for CGM-based glucose trajectory prediction, these methods are typically evaluated on small, private datasets, impeding reproducibility, further research, and practical adoption. The absence of standardized prediction tasks and systematic comparisons between methods has led to uncoordinated research efforts, obstructing the identification of optimal tools for tackling specific challenges. As a result, only a limited number of prediction methods have been implemented in clinical practice. To address these challenges, we present a comprehensive resource that provides (1) a consolidated repository of curated publicly available CGM datasets to foster reproducibility and accessibility; (2) a standardized task list to unify research objectives and facilitate coordinated efforts; (3) a set of benchmark models with established baseline performance, enabling the research community to objectively gauge new methods' efficacy; and (4) a detailed analysis of performance-influencing factors for model development. We anticipate these resources to propel collaborative research endeavors in the critical domain of CGM-based glucose predictions. {Our code is available online at github.com/IrinaStatsLab/GlucoBench.

GlucoBench: Curated List of Continuous Glucose Monitoring Datasets with Prediction Benchmarks

TL;DR

A comprehensive resource that provides a consolidated repository of curated publicly available CGM datasets to foster reproducibility and accessibility, a standardized task list to unify research objectives and facilitate coordinated efforts, and a set of benchmark models with established baseline performance to objectively gauge new methods' efficacy are presented.

Abstract

The rising rates of diabetes necessitate innovative methods for its management. Continuous glucose monitors (CGM) are small medical devices that measure blood glucose levels at regular intervals providing insights into daily patterns of glucose variation. Forecasting of glucose trajectories based on CGM data holds the potential to substantially improve diabetes management, by both refining artificial pancreas systems and enabling individuals to make adjustments based on predictions to maintain optimal glycemic range.Despite numerous methods proposed for CGM-based glucose trajectory prediction, these methods are typically evaluated on small, private datasets, impeding reproducibility, further research, and practical adoption. The absence of standardized prediction tasks and systematic comparisons between methods has led to uncoordinated research efforts, obstructing the identification of optimal tools for tackling specific challenges. As a result, only a limited number of prediction methods have been implemented in clinical practice. To address these challenges, we present a comprehensive resource that provides (1) a consolidated repository of curated publicly available CGM datasets to foster reproducibility and accessibility; (2) a standardized task list to unify research objectives and facilitate coordinated efforts; (3) a set of benchmark models with established baseline performance, enabling the research community to objectively gauge new methods' efficacy; and (4) a detailed analysis of performance-influencing factors for model development. We anticipate these resources to propel collaborative research endeavors in the critical domain of CGM-based glucose predictions. {Our code is available online at github.com/IrinaStatsLab/GlucoBench.
Paper Structure (27 sections, 4 equations, 6 figures, 16 tables)

This paper contains 27 sections, 4 equations, 6 figures, 16 tables.

Table of Contents

  1. Introduction
  2. Related works
  3. Data
  4. Description
  5. Pre-processing
  6. Benchmarks
  7. Tasks and metrics
  8. Models
  9. Testing protocols
  10. Results
  11. Analysis
  12. Why does the performance of the models differ between the datasets?
  13. Are the models generalizable to patients beyond the training dataset?
  14. How does adding the covariates affect the modeling quality?
  15. Discussion
  16. ...and 12 more sections

Figures (6)

  • Figure 1: Sample of glucose curves captured by the Dexcom G4 Continuous Glucose Monitoring (CGM) system, with dates de-identified for privacyweinstock2016risk.
  • Figure 2: An illustration of static (Age), dynamic known (Date), and dynamic unknown (Heart Rate) covariate categories based on data from hall2018glucotypes and dubosson2018open.
  • Figure 3: Example data processing on weinstock2016risk. The red arrows denote segmentation, green blocks interpolate values, and red brackets indicate dropped segments due to length.
  • Figure 4: Analysis of errors by: (a) OD versus ID, (b) population diabetic type (healthy $\rightarrow$ Type 2 $\rightarrow$ Type 1), (c) daytime (9:00AM to 9:00PM) versus nighttime (9:00PM to 9:00AM).
  • Figure 5: Model forecasts on weinstock2016risk dataset.
  • ...and 1 more figures