The digital transformation of healthcare relies heavily on achieving semantic interoperability across information systems. A key step toward this goal is the analysis and integration of reference terminologies, such as SNOMED CT. This paper introduces a structured model for manually and semi-automatically mapping concepts from psychological assessment instruments to SNOMED CT. The model incorporates three main agents and is organized into multiple phases of data discovery and validation. Large language models, including ChatGPT and Copilot, were employed to support various tasks within this process. Beyond promoting consistent terminology use in underrepresented domains, the proposed model also contributes to the enrichment of clinical coding standards and the advancement of health data interoperability.
Oliveira et al. (Thu,) studied this question.