A Methodology to Compare Insulin Dosing Recommendations in Real-Life Settings

Area of Science:

• Endocrinology
• Biomedical Engineering
• Data Science

Background:

• Complex real-life glucose data from insulin pump users presents analysis challenges.
• Type 1 diabetes management requires accurate insulin dosing and lifestyle adjustments.

Purpose of the Study:

• To propose and evaluate a methodology for analyzing complex glucose data from insulin pump users.
• To assess the performance of a proposed decision aid (PDA) for insulin bolus recommendations, alcohol consumption, and exercise in type 1 diabetes.

Main Methods:

• Recruited 15 patients with type 1 diabetes on insulin pumps.
• Collected 30 days of glucose data, insulin bolus amounts, and self-reported alcohol/exercise events.
• Developed rules to compare insulin pump bolus calculator (IPBC) recommendations with the PDA and assess PDA advice for lifestyle factors.

Main Results:

• The PDA showed varied recommendations compared to IPBC for different glucose levels (below, at, or above target).
• PDA provided appropriate advice for 64% of alcohol consumption events and 75% of exercise events.
• Preliminary analysis indicates the PDA's potential for guiding insulin dosing and lifestyle advice.

Conclusions:

• The study presents a systematic methodology for analyzing real-life insulin pump data.
• The proposed decision aid demonstrates preliminary efficacy in managing insulin dosing and lifestyle factors for type 1 diabetes.
• Further research with a broader population and prospective trials are necessary to validate the methodology and PDA performance.