Optimizing Nursing Data for Al and CDSS: SNOMED CT Mapping Using K-MIMIC

This study mapped K-MIMIC (Korean Medical Information Mart for Intensive Care) nursing data to SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms), an international standard terminology, to assess its potential for data standardization and utilization in AI (Artificial Intelligence)-based CDSS (Clinical Decision Support System) development. Of approximately 12,000 nursing data entries from two hospitals, 8,424 were included in the final mapping, with exact matches accounting for 90.4%, broad matches for 8.5%, and unmatched entries for 1.1%. Additionally, 76.9% of the data were classified as 1:N mappings, with the primary categories being Procedures (57.1%) and Clinical findings (38.2%). The results confirm that SNOMED CT is suitable for representing intensive care nursing data, supporting AI development and clinical decision-making. This study highlights the importance of standardizing and optimizing nursing data for CDSS development and suggests that future research should focus on improving the mapping process and validating the effectiveness of CDSS based on standardized data.