Algorithm-Enabled, Personalized Glucose Management for Type 1 Diabetes at the Population Scale: A Prospective Evaluation in Clinical Practice
ABSTRACT
Background:
Continuous glucose monitors (CGM) are recommended as standard of care by the American Diabetes Association for individuals with type 1 diabetes on insulin. These devices generate glucose readings every 5-15 minutes and use cloud-based platforms to share data. This remotely reviewed data can be used by members of diabetes care team to provide remote care.
Objective:
To design an automated tool that facilitates timely, personalized, population-level guidance for glucose management through asynchronous telehealth.
Methods:
Using CGM data from six clinical trials and two observational datasets, we developed manufacturer-agnostic algorithms to generate generic (e.g., mean glucose (MG) > 170mg/dL) and personalized (e.g., MG increased by >10mg/dL) flags. We developed and deployed an automated tool in a pediatric type 1 diabetes clinic, measured sensitivity for identifying who may benefit from telehealth, and measured the time saved reviewing data with the use of the tool.
Results:
The eight cohorts contained 1,365 patients with 30,017 weeks of data collected by seven types of CGMs. In the cohort with the highest MG, 81.3% (26 of 32) and 3.1% (1/32) of people had a generic and personalized flag every week, respectively. In the clinic, on average, 57.2% of patients were flagged per week, corresponding to a sensitivity of 98.6% and a 42.8% reduction in the time required to review data.
Conclusions:
The automated analysis of CGM data may help identify people requiring guidance on glucose management while reducing the workload for care providers. The rules-based approach provided fully interpretable representations of patient status relative to the latest guidelines. When deployed in a clinic, an automated tool to generate flags identified 98.6% of patients who would benefit from asynchronous telehealth contact while reducing the time required to review patient data by 42.8%. Guideline-based population health management may become more accessible through the use of automated tools.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.