A Model-Based Insulin Dose Optimization Algorithm for People With Type 1 Diabetes on Multiple Daily Injections...

Area of Science:

• Biomedical Engineering
• Computational Biology
• Endocrinology

Background:

• Type 1 diabetes (T1D) management commonly uses multiple daily injections (MDI) for insulin therapy.
• Achieving optimal glycemic control with MDI is challenging due to significant inter- and intra-individual variability in insulin response.
• Accurate insulin dosing is critical for preventing hyperglycemia and hypoglycemia in T1D patients.

Purpose of the Study:

• To develop and evaluate a novel model-based iterative algorithm for optimizing insulin doses in T1D patients on MDI therapy.
• To leverage previous-day glucose, insulin, and meal data for personalized insulin dose recommendations.
• To enhance the precision of insulin dosing to improve overall glycemic control.

Main Methods:

• A maximum-a-posteriori method was used to estimate parameters of a mathematical model for basal-bolus insulin effects.
• The algorithm integrates parameter estimates, confidence intervals, and goodness-of-fit metrics to generate updated insulin dose recommendations.
• The algorithm's performance was assessed using a 150-day clinical dataset, 60-day simulations, and a 6-day clinical experiment.

Main Results:

• The developed model demonstrated a good fit to clinical data, with a root-mean-square error of 1.75 mmol/L.
• Simulations showed a significant improvement in time spent within the target glucose range (64% to 77%).
• Hypoglycemia (time < 3.9 mmol/L) was reduced from 8.1% to 3.8%, and the algorithm proved feasible in a clinical setting.

Conclusions:

• The novel algorithm shows significant potential for improving glycemic control in individuals with T1D using MDI.
• This iterative, model-based approach offers a promising step towards advanced decision support systems for diabetes management.
• The findings suggest a pathway to enhance the quality of life for T1D patients through more precise insulin therapy.