Implementing a digital audit-and-feedback surveillance system for diabetic ketoacidosis: study protocol for the DEKODE programme.

INTRODUCTION: Diabetic ketoacidosis is one of the leading causes of unplanned hospital admission among people with diabetes, yet adherence to treatment guidelines remains inconsistent, contributing to variation in care and suboptimal outcomes. Conventional audit approaches are limited by non-standardised data collection, restricting benchmarking and shared learning. The Digital Evaluation of Ketosis and Other Diabetic Emergencies (DEKODE) platform was developed as a cloud-based audit-and-feedback system to support near real-time performance monitoring and continuous quality improvement, but its implementation requires theory-informed evaluation. METHODS: This multi-site, mixed-methods study involves six National Health Service hospitals, including early adopters and sites new to DEKODE. Guided by an implementation research logic model, the study links contextual determinants to tailored strategies. Quantitative outcomes, including episode duration, guideline adherence, complications, and clinical outcomes, will be analysed using interrupted time series methods. Qualitative stakeholder interviews will be analysed thematically, and findings integrated using triangulation. RESULTS: The study will assess the impact of DEKODE on clinical and process outcomes and identify determinants influencing adoption, fidelity, and sustainability. DISCUSSION: Findings will inform scalable implementation of digital audit-and-feedback systems to reduce variation in acute diabetes care and improve patient outcomes. CLINICAL TRIAL REGISTRATION: https://www.isrctn.com/ISRCTN17096714, identifier ISRCTN17096714.