Multi-Timescale Rhythmicity of Blood Glucose and Insulin Delivery Reveals Key Advantages of Hybrid Closed Loop Therapy

End-to-end offline reinforcement learning for glycemia control

The development of closed-loop systems for glycemia control in type I diabetes relies heavily on simulated patients. Improving the performances and adaptability of these close-loops raises the risk of over-fitting the simulator. This may have dire consequences, especially in unusual cases which were not faithfully - if at all - captured by the simulator. To address this, we propose to use model-free offline RL agents, trained on real patient data, to perform the glycemia control. To further improve the performances, we propose an end-to-end personalization pipeline, which leverages offline-policy evaluation methods to remove altogether the need of a simulator, while still enabling an estimation of clinically relevant metrics for diabetes.

Practical Guidance on Open Source and Commercial Automated Insulin Delivery Systems: A Guide for Healthcare Professionals Supporting People with Insulin-Requiring Diabetes

As increasing numbers of people with insulin-managed diabetes use automated insulin delivery (AID) systems or seek such technologies, healthcare providers are faced with a steep learning curve. Healthcare providers need to understand how to support these technologies to help inform shared decision making, discussing available options, implementing them in the clinical setting, and guiding users in special situations. At the same time, there is a growing diversity of commercial and open source automated insulin delivery systems that are evolving at a rapid pace. This practical guide seeks to provide a conversational framework for healthcare providers to first understand and then jointly assess AID system options with users and caregivers. Using this framework will help HCPs in learning how to evaluate potential new commercial or open source AID systems, while also providing a guide for conversations to help HCPs to assess the readiness and understanding of users for AID systems. The choice of an AID system is not as simple as whether the system is open source or commercially developed, and indeed there are multiple criteria to assess when choosing an AID system. Most importantly, the choices and preferences of the person living with diabetes should be at the center of any decision around the ideal automated insulin delivery system or any other diabetes technology. This framework highlights issues with AID use that may lead to burnout or perceived failures or may otherwise cause users to abandon the use of AID. It discusses the troubleshooting of basic AID system operation and discusses more advanced topics regarding how to maximize the time spent on AID systems, including how to optimize settings and behaviors for the best possible outcomes with AID technology for people with insulin-requiring diabetes. This practical approach article demonstrates how healthcare providers will benefit from assessing and better understanding all available AID system options to enable them to best support each individual.

How It Started, How It Is Going: The Future of Artificial Pancreas Systems (Automated Insulin Delivery Systems)

Originally, the future of automated insulin delivery (AID) systems, or artificial pancreas systems (APS), was having them at all, in any form. We've learned in the last half dozen years that the future of all artificial pancreas systems holds higher time in range, less work required to manage automated insulin delivery systems to improve quality of life, and the ability to input critical information back into the system itself. The data and user experience stories make it clear: APS works. APS are an improvement over other diabetes therapy methods when they are made available, accessible, and affordable. Understanding the unmet expectations of current users of first generation APS technology may also aid in the development of improved technology and user experiences for the future of APS.

An Open-Source Web Portal for Managing Self-reported Data and Real-world Data Donation in Diabetes Research: Feasibility Study. (Preprint)

Background

People with diabetes and their support networks have developed open-source automated insulin delivery systems to help manage their diabetes therapy, as well as to improve their quality of life and glycemic outcomes. Under the hashtag #WeAreNotWaiting, a wealth of knowledge and real-world data have been generated by users of these systems but have been left largely untapped by research; opportunities for such multimodal studies remain open.

Objective

We aimed to evaluate the feasibility of several aspects of open-source automated insulin delivery systems including challenges related to data management and security across multiple disparate web-based platforms and challenges related to implementing follow-up studies.

Methods

We developed a mixed methods study to collect questionnaire responses and anonymized diabetes data donated by participants—which included adults and children with diabetes and their partners or caregivers recruited through multiple diabetes online communities. We managed both front-end participant interactions and back-end data management with our web portal (called the Gateway). Participant questionnaire data from electronic data capture (REDCap) and personal device data aggregation (Open Humans) platforms were pseudonymously and securely linked and stored within a custom-built database that used both open-source and commercial software. Participants were later given the option to include their health care providers in the study to validate their questionnaire responses; the database architecture was designed specifically with this kind of extensibility in mind.

Results

Of 1052 visitors to the study landing page, 930 participated and completed at least one questionnaire. After the implementation of health care professional validation of self-reported clinical outcomes to the study, an additional 164 individuals visited the landing page, with 142 completing at least one questionnaire. Of the optional study elements, 7 participant–health care professional dyads participated in the survey, and 97 participants who completed the survey donated their anonymized medical device data.

Conclusions

The platform was accessible to participants while maintaining compliance with data regulations. The Gateway formalized a system of automated data matching between multiple data sets, which was a major benefit to researchers. Scalability of the platform was demonstrated with the later addition of self-reported data validation. This study demonstrated the feasibility of custom software solutions in addressing complex study designs. The Gateway portal code has been made available open-source and can be leveraged by other research groups.