JMIR Publications

ABSTRACT

Intravenous (IV) fluid therapy is a cornerstone of haemodynamic management, yet its understanding is often limited by inconsistent clinical terminology. As the medical field moves toward artificial intelligence (AI)-empowered Clinical Decision Support Systems (CDSS-AI) with fully automated infusion systems, there is a critical need for high-quality, verified training data. The reliability of these systems is threatened by biased or misleading data resulting from ambiguous definitions and the interchangeable use of non-synonymous terms, such as "fluid overload" and "hypervolaemia". This expert opinion article identifies some drawbacks in current terminology and proposes ashift toward adequacy-based phenotypes, categorizing fluid status as "insufficient," "optimal," or "excessive" rather than relying on imprecise quantitative measurements. We suggest a phenotype matrix designed to provide a structured, systematic description of fluid requirements that can be integrated into clinical algorithms and machine learning databases. By establishing a consistent medical language, we can reduce the risks of erroneous AI recommendations and initiate expert consensus to harmonize terminology for the era of automated therapy.