New platform standardizes analysis of continuous glucose monitoring data

Continuous glucose monitors (CGMs) produce an overwhelming amount of data on how the body regulates glucose. As researchers harness the potential of CGMs in metabolic studies, they face the challenge of transforming this massive data into coherent and comparable insights. To tackle this issue, a team of scientists from institutions across the United States and United Kingdom, including Pennington Biomedical Research Center and King's College London, has developed GVC-Calc, a new platform aimed at standardizing the analysis of CGM data.

Published in Nature Metabolism, the paper "GVC-Calc enables large-scale analysis of continuous glucose monitoring data" describes how GVC-Calc overcomes inconsistencies in data preprocessing and metric definitions, which often hinder the comparison of findings across studies and integration into large-scale analyses. Unlike existing platforms that rely on specific manufacturers or proprietary software, GVC-Calc seamlessly processes raw CGM files from various devices, enabling batch processing for hundreds of participants while linking individual-level data to cohort-level results.

The platform incorporates consensus clinical metrics and exploratory measures of glucose dynamics, such as geometric and probabilistic approaches, which reveal temporal patterns that conventional summary statistics might miss. By providing multiple transparent implementations of widely used metrics, GVC-Calc allows researchers to align their analyses with prior studies and compare results across different analytical methods.

Researchers from Pennington Biomedical, including postdoctoral researcher Dr. Kaja Falkenhain and adjunct professor Dr. Leanne Redman, played a crucial role in evaluating GVC-Calc from a clinical perspective. They tested the tool's visualizations and workflow, offering recommendations to enhance its usability and interpretation in clinical settings.

To illustrate the platform's capabilities, the researchers applied GVC-Calc to four independent CGM studies involving diverse populations, devices, and study designs. The same analytical framework ensured consistent preprocessing, metric extraction, and downstream statistical analysis across these varied datasets. The study demonstrates that a standardized and scalable analytical framework like GVC-Calc can facilitate researchers' ability to work with large and heterogeneous CGM datasets more effectively.

The complete GVC-Calc platform is available as a web application, and its source code is publicly accessible, fostering transparency, reproducibility, and ongoing use within the research community.