Learning challenges of healthcare professionals supporting open-source automated insulin delivery

Background: Open-source automated insulin delivery (AID) is a user-driven treatment modality used by thousands globally. Healthcare professionals' (HCPs) ability to support users of this technology is limited by a lack of knowledge of these systems. Aims: To describe the challenges experienced by HCPs supporting participants' use of open-source automated insulin delivery in the Community deRivEd AuTomatEd insulin delivery (CREATE) study. Methods: Data were collected prospectively from the study team's fortnightly meetings and Slack Workspace (Slack Technologies, Ltd. 2018) during the first 4 months of the trial. Key topics were identified from minutes of meetings. Slack conversations were categorised by topic, with the number of posts per conversation, number of sites per conversation and involvement of experts in open-source AID being recorded. Results: In the first 4 months of the trial, there were 254 conversations in Slack with a mean of 5.2 (+/- 4.25) posts per conversation. The most frequent learning challenge was insulin pump and cannula problems relating to the DANA-i (TM) insulin pump, which totalled 24.0% of all conversations. Experts on open-source AID use were involved in 83.3% of conversations. Conclusions: A significant proportion of challenges related to specific devices, rather than AID. Challenges relating to the functioning of open-source AID were more likely to involve input from experts in open-source AID. This is the first report of challenges experienced by a multidisciplinary team in a supported open-source environment that may inform expectations in routine clinical care.