What question did this study set out to answer?

The review aims to evaluate automated mapping methods for clinical terminologies and identify research gaps.

March 5, 2026Open Access

A systematic review of automatic mapping of clinical terminologies

MAMohammad Mekayel AnikBangladesh University of Engineering and Technology SASyed AhmmedBangladesh University of Engineering and Technology MMM.Rubaiyat Hossain MondalBangladesh University of Engineering and Technology

Key Points

The review aims to evaluate automated mapping methods for clinical terminologies and identify research gaps.
Conducted a systematic review following PRISMA guidelines.
Analyzed published articles from January 2018 to July 2025.
Reviewed applications of AI, NLP, ML, and DL in clinical data mapping.
Examined studies on mapping to LOINC, SNOMED-CT, and ICD.
ML methods achieved up to 0.99 accuracy for LOINC and 0.954 for SNOMED-CT.
LightGBM achieved 0.952 accuracy for ICD-10.
NLP models reached 0.91 accuracy for LOINC.
Identified significant research gaps and opportunities in automated mapping.

Abstract

Automated mapping of clinical terminology is important for standardizing clinical data, ensuring interoperability across electronic health records. Artificial intelligence (AI), natural language processing (NLP), machine learning (ML), and deep learning (DL), have the potential for automating clinical data mapping to different terminologies. There is a lack of review papers on automated clinical terminology mapping, particularly regarding mapping to multiple terminologies. In this paper, an evaluation of research articles published from 1 January 2018 to 31 July 2025 is presented, following the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. The novelty of this review is that it presents various clinical datasets originating from different countries and stored in different languages that are used in the current automated mapping of clinical terminology. Moreover, existing applications of different technologies used for automated mapping are briefly described, and the results of those applications are listed. Additionally, the existing studies on automated mapping to Logical Observation Identifiers Names and Codes (LOINC), Systematized Nomenclature of Medicine - Clinical Terms (SNOMED-CT), and International Classification of Diseases (ICD) are objectively reviewed. The review of the existing studies indicates that ML methods achieve up to 0.99 accuracy for LOINC, 0.954 for SNOMED-CT, with LightGBM obtaining 0.952 for ICD-10, and NLP models achieving 0.91 for LOINC. Finally, the paper concludes by highlighting current research gaps and future research opportunities in automated mapping.

Ask AI

Helpful

Bookmark

View Full Paper