Systematic Review: Semantics-driven Improvements in Electronic Health Records Data Quality

Details

This article summarizes a systematic review published in BMC Medical Informatics and Decision Making titled 'Semantics-driven improvements in electronic health records data quality: a systematic review'. The study reviews various approaches—including NLP and fuzzy methods—to enhance the Data Quality (DQ) of EHRs. The core finding is that leveraging standardized vocabularies and knowledge graphs provides deep semantic understanding, which is critical for improving EHR data. Technically, EHR data structures are defined by standards like HL7 FHIR based on ISO 13606. Semantic approaches utilize standard medical terminologies such as MeSH and SNOMED-CT (vocabularies). Furthermore, knowledge graphs can be constructed using tools like Protégé or Jena in RDF/OWL format, allowing for information extraction via SPARQL queries. The review comprehensively summarizes how these semantic technologies contribute to EHR data quality improvement. Approaches utilizing standardized vocabularies and ontologies (OWL) are shown to achieve high accuracy in information extraction.