Evaluating Feasibility of Personal Diabetes Device Data Collection for Research

BACKGROUND Diabetes devices, like insulin pumps and continuous glucose monitors (CGM), capture and store patient adherence and utilization data that can be retrieved or downloaded providing objective information on self-management behaviors; yet, diabetes device data remain underutilized in research. OBJECTIVE To examine the usability and feasibility of personal diabetes device data collected using a clinical download platform retooled for research purposes. METHODS Investigators evaluated the feasibility of raw diabetes device data collection. One hundred eight preteens and adolescents with type 1 diabetes and their parents provided consent/assent. RESULTS Data were successfully collected from the diabetes devices (insulin pumps and CGM) of 97 youth using a clinical download software adapted for research including data from all three commercially available CGM systems and insulin pumps brands, which contained all current and previous models of each insulin pump brand. The time required to download, mode of connection, and process varied significantly between brands. Despite the use of an agnostic download software, some outdated device brands and cloud-based CGM data were unsupported during data collection. Within the download software, dummy clinical accounts were created for each study participant, which were then linked back to a master study account for data retrieval. Raw device data were extracted into 7-8 Excel files per participant which were then used to develop aggregate daily measures. DISCUSSION Our analysis is the first of its kind to examine the feasibility of raw diabetes device data using a clinical download software. The investigators highlight issues encountered throughout the research process along with mitigating strategies to inform future inquiry. CONCLUSION This study demonstrates the feasibility of raw data collection, from a wide variety of insulin pump and CGM brands, through the retooling of a clinical download software. Data from these personal devices provide a unique opportunity to study self-management behavior and the glycemic response of individuals in their everyday environments.