DEKODE—A cloud-based performance feedback model improved DKA care across multiple hospitals in the UK

AIM: A current gap in Diabetes-related ketoacidosis (DKA) research is understanding the factors contributing to variations in care and outcomes between people admitted with DKA. We aimed to create a system to facilitate gathering data on DKA management across multiple centres and identify trends in complications and outcomes associated with DKA.

RESEARCH DESIGN AND METHODS: Between January 2020 and December 2022, we set up a cloud-based Quality improvement project (QIP) that provided regular feedback to 11 hospitals in the United Kingdom (UK).

RESULTS: Of the 1977 episodes, we observed an increase in adherence in fluid prescription in hospitals C, D, E, F and G (C- 23% vs. 75% p = <0.001; D- 27% vs. 60%, p = <0.001; E- 17 vs. 79% p = <0.001; F- 16% vs. 57%, p = <0.001; G- 36% vs. 75% p = <0.001). Notable improvements in adherence to glucose monitoring were observed in hospitals B, D, and G (B- 11 vs. 38% p = <0.001; D- 36% vs. 56%, p = 0.05; G- 22% vs. 67% p = <0.001). Although we didn't observe significant changes in complications and outcomes among participating hospitals from the start to the end of the reported period, notable fluctuations were evident across quarters. These variations were relayed to the respective hospitals, underscoring how feedback and interventions could influence the care provided. This initiative also marks the initial move towards establishing and improving data collection practices in acute diabetes.

CONCLUSIONS: We demonstrate a sustainable QIP that improves adherence to national guidelines in some indicators for DKA care and serves as an early warning system to identify adverse trends.