Technical Name |
SmartEHR Coderr: LLM-based Automatic EHR Clinical Coding Support System |
Project Operator |
工業技術研究院 |
Project Host |
張慈映 |
Summary |
We present SmartEHR Coder, a clinical coding support system powered by large language models (LLMs), developed to transform unstructured clinical notes into standardized, computable data. Designed to address the fragmentation of Taiwan’s healthcare IT environment, the system maps electronic health records (EHRs) into international coding standards—LOINC, SNOMED CT, and RxNorm—while supporting ethical, human-in-the-loop workflows. |
Scientific Breakthrough |
SmartEHR Coder leverages large language models (LLMs) and retrieval-augmented generation (RAG) to transform raw clinical text into structured, interoperable codes across ICD-10, SNOMED CT, LOINC, and RxNorm. SmartEHR Coder offers contextual understanding, semantic preservation, and better adaptability to noisy, multilingual clinical text. SmartEHR Coder uniquely supports human-in-the-loop workflows—providing traceable, explainable suggestions while preserving coder oversight. |
Industrial Applicability |
This technology extracts semantic data from medical records and converts it into standardized codes, solving cross-hospital integration challenges and improving EMR structuring. Applicable across all healthcare levels, it supports clinical use, AI, and research. Deployable via cloud, on-premises, or within HIS, it offers scalability and value—forming a foundation for smart healthcare in Taiwan with global potential. |
Keyword |
Large Language Model Electronic Health Record Clinical Coding Support FHIR ETL ICD-10 SNOMED CT LOINC RxNorm Interoeprability |