The objective of this proposal is to optimize the design and evaluate a robust artificial pancreas (R-AP) system for use in patients with uncontrolled type 1 diabetes (T1D) with HbA1C greater than 8% and compare HbA1C outcomes in these patients relative to a decision support system that utilizes continuous glucose monitoring (CGM) and multiple daily injection (MDI) therapy. Although high risk patients have possibly the most to gain from usage of AP technology, they are oftentimes under-represented or excluded from clinical trials. This has been because of the increased risk of failure of these AP systems that were not designed to handle inconsistent reporting of meals, variable activity level, and infusion set failures. An AP system for high risk patients needs to be designed to achieve maximal benefit, including reducing the risk of acute and chronic complications. A major obstacle for enabling the AP for usage by high-risk patients is that these patients may be less compliant with use guidelines for the system including missed meal announcements, infrequent sensor calibrations, and prolonged infusion set wear leading to infusion set failures. In this grant, we will integrate new risk-mitigation features into the OHSU single-hormone AP to enable usage by high-risk patients that fall into the categories described above. We present new algorithms for automating the detection of missed meal announcements, missed calibrations, and robust handling of hybrid usage mode. While AP systems may be an optimal choice for improving glycemic control, many people with T1D prefer MDI therapy. Decision support systems such as the DailyDose decision support system developed at OHSU can be used to improve glycemic control for patients who prefer MDI therapy. The DailyDose decision support system is designed for CGM augmented MDI therapy. It enables on-demand calculation of insulin doses, automates insulin dose adjustments based on pattern recognition, and uses machine learning approaches to alert the patients to events such as predicted hypoglycemia and missed meal doses.The benefit of the DailyDose system is that it is a simple system and does not require use of an insulin pump, which may be a challenge for some patients with uncontrolled type 1 diabetes as pump therapy is more intensive and requires infusion set changes. It is unknown in this high risk group of people whether patient needs, quality of life, and glycemic control are best addressed with an AP system or decision support tool or if both treatments are appropriate. We have designed a 3-month clinical study to compare glycemic outcomes during AP vs. decision support interventions in a high- risk T1D cohort (HbA1C 8-10.5%), with the aim of demonstrating a significant clinically relevant reduction in HbA1C. Our hypothesis is that both AP and decision support therapies will decrease HbA1C relative to baseline but that the AP will provide further benefit over DailyDose.

Public Health Relevance

Diabetes is the leading cause of blindness, kidney failure, and non-traumatic amputations. We propose to integrate features into the OHSU artificial pancreas to optimize it for high risk patients with type 1 diabetes (HbA1c>8-10.5%) and compare it to DailyDose, a decision support system that uses multiple daily injections.

Agency
National Institute of Health (NIH)
Institute
National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)
Type
Research Project (R01)
Project #
1R01DK120367-01
Application #
9691740
Study Section
Special Emphasis Panel (ZDK1)
Program Officer
Arreaza-Rubin, Guillermo
Project Start
2018-09-30
Project End
2022-06-30
Budget Start
2018-09-30
Budget End
2019-06-30
Support Year
1
Fiscal Year
2018
Total Cost
Indirect Cost
Name
Oregon Health and Science University
Department
Internal Medicine/Medicine
Type
Schools of Medicine
DUNS #
096997515
City
Portland
State
OR
Country
United States
Zip Code
97239