Structuralization of Digestive Endoscopic Report Based on NLP

It is an inevitable transition from paper-based medical records to electronic medical records (EMR). EMR's structuralization and standardization are important for efficient information retrieve and sharing. Presently, natural language processing (NLP) and structured data entry (SDE) are two different methods realizing EMR's structuralizaion. Because SDE limits doctors' free expression, this paper designs a system based on NLP and minimal standard terminology (MST) that automatically maps digestive endoscopic narrative records to structured MST records and its accuracy is 92.3%.