NCT06676657

Closed Loop Context Aware AID

Assessment of a Fully-closed Loop AID System With a Context-aware Hyperglycemia Pattern Detection and Dosing Algorithm in People With Type 1 Diabetes

• Status: Recruiting
• Phase: N/A
• Study type: Interventional
• Enrollment: 10 (estimated)
• Sponsor: Oregon Health and Science University · Academic / Other
• Sex: All
• Age: 18 Years – 100 Years
• Healthy volunteers: Not accepted

Summary

An artificial pancreas (AP) is a control system for automatic insulin delivery. The investigators have implemented a high blood sugar detection and dosing algorithm for use within an AP control system. If a high blood sugar pattern is detected, correction insulin will be calculated and delivered. The investigators will test how well the new algorithm manages glucose compared to the AP control system without high blood sugar detection and dosing. This type of algorithm may improve glucose control for high risk patient populations.

Detailed description

Participants will be on study for approximately 4 weeks. During the study, participants will wear an Omnipod to deliver insulin. Participants will also wear a Dexcom G6 CGM. The CGM system will send sensed glucose data every 5 minutes wirelessly via Bluetooth Low Energy (BTLE) to an Android smartphone running the iPancreas app. The closed-loop system will receive activity data through an activity watch worn by the participant. Participants will complete system training on Day 1 in clinic and then spend the rest of the 4 weeks under free-living conditions. The first 3 weeks of the study will be the training period when the system will collect patterns from the glucose sensor, insulin, and fitness data that lead to high blood sugar. After the 3-week training period, participants will complete a virtual visit to train the participant on the high blood sugar detection and dosing algorithm, then continue to use for an additional week. Participants will complete meal and exercise challenges.

Conditions

• Type 1 Diabetes

Interventions

Type	Name	Description
DEVICE	iPancreas automated insulin delivery system	The Model Predictive Control (MPC) insulin infusion algorithm contains a model within the controller that takes as an input the aerobic metabolic expenditure in addition to the CGM and meal in puts. The algorithm uses heart rate and accelerometer data collected on the patient's body to calculate metabolic expenditure (METs). The METs then acts on the model for the insulin dynamics, whereby more energy expenditure and longer duration exercise can lead to a more substantial effect of insulin on the CGM. The MPC also has missed meal insulin bolus detection where the system will calculate the amount of insulin that was missed for a meal. The missed meal boluses can be delivered automatically without any input from the user. This feature can also be disabled. The MPC has a new feature called hyperglycemia pattern detection and dosing algorithm that will analyze problem patterns associated with high blood sugar and automatically calculate and deliver a correction dose.

Timeline

• Start date: 2024-11-22
• Primary completion: 2025-07-30
• Completion: 2025-08-30
• First posted: 2024-11-06
• Last updated: 2025-04-11

Locations

1 site across 1 country: United States

Regulatory

• FDA-regulated device study