Evaluating Insulin Delivery Systems Using Dynamic Glucose Region Plots and Risk Space Analysis

Area of Science:

• Endocrinology
• Biomedical Engineering
• Diabetes Technology

Background:

• Continuous glucose monitoring (CGM) provides valuable data for managing diabetes.
• Automated insulin delivery (AID) systems aim to improve glycemic control.
• Assessing the performance of different AID systems is crucial for patient care.

Purpose of the Study:

• To compare the performance of two AID systems (A and B) using dynamic glucose region plots.
• To evaluate glucose homeostasis by analyzing glucose rate of change (RoC) and glucose levels.
• To identify clinically relevant differences in glycemic control between AID systems.

Main Methods:

• Utilized 60-day CGM data from individuals with type 1 diabetes using AID systems A (n=65) and B (n=85).
• Employed dynamic glucose region plots to visualize simultaneous glucose and RoC values.
• Defined risk spaces to predict glucose trajectory (staying in, leaving, or returning to target range) within 30 minutes.

Main Results:

• System A showed a significantly higher fraction of (RoC, glucose) values predicted to remain in the target range (62.5% vs. 56.8%, p=0.002).
• System A exhibited a slightly lower fraction of values in range with a risk of progressing to hyperglycemia (glucose > 10.0 mmol/L).
• Dynamic glucose region plots effectively demonstrated differences in glucose homeostasis between the two AID systems.

Conclusions:

• Dynamic glucose region plots and risk spaces offer novel insights into glucose homeostasis.
• These methods can reveal clinically significant performance differences between AID systems.
• System A demonstrated superior performance in maintaining glucose within the target range compared to System B.