CADApp: uma solução tecnológica para padronização e apoio no tratamento inicial da cetoacidose diabética

Abstract

Diabetic ketoacidosis (DKA) is an acute and potentially life-threatening complication of diabetes mellitus (DM), associated with high morbidity and mortality rates when treatment is not initiated promptly and appropriately. Clinical management involves multiple steps from fluid replacement and correction of electrolyte imbalances to insulin therapy, requiring precision and strict adherence to therapeutic guidelines, especially in emergency settings. However, studies indicate that adherence to clinical protocols can be compromised in real-world practice, particularly in high-demand environments where treatment complexity increases the risk of therapeutic errors, such as inadequate potassium replacement, premature initiation of insulin therapy, and inappropriate selection of intravenous solutions. In this context, the use of digital clinical decision support tools, such as CADApp, represents a promising strategy to reduce inconsistencies in DKA management. By providing algorithms based on updated guidelines, automated calculations, and interactive flowcharts, the application assists healthcare professionals in making rapid and safe decisions, potentially contributing to standardized clinical practices and minimizing errors during treatment. This study aimed to design, implement, and validate a digital tool for use on smartphones and web browsers, developed for physicians and focused on systematizing the initial management of DKA in accordance with current scientific guidelines. The application, named CADApp, was developed based on the recommendations of the Brazilian Diabetes Society, structuring treatment from the onset of hydration to the insulinization stage. The development process included a literature review, modeling, and prototype construction, followed by usability validation using the System Usability Scale (SUS) applied to 17 physicians working in emergency services. SUS scores ranged from 85.0 to 100.0, with a mean of 95, indicating excellent usability. Seven of the ten items on the scale showed a Content Validity Index (CVI) = 1.00, while the remaining three had a CVI = 0.94, demonstrating a high level of expert agreement. Internal consistency was considered acceptable (Cronbach’s alpha = 0.686), and McDonald’s omega coefficient indicated moderate consistency (ω = 0.502). The results demonstrate that CADApp has high potential as a clinical decision-support tool for managing diabetic ketoacidosis and may contribute to improved standardization, safety and effectiveness of care in emergency settings.