The STRIDE ChartIndex Project

A large proportion of clinical information is contained within clinical documents, such as radiology reports, surgical pathology reports, operative reports, discharge summaries and clinic notes. Much of this document-based data is represented as unstructured narrative text with little, if any, standardization of the language used to represent important information such as diagnosis, therapy or test results. The data contained within these documents is therefore difficult to integrate into clinical or research databases that need to support efficient standards-based retrieval. To address this problem we are developing a computer system, called ChartIndex (1,2,3,4), that can structure clinical documents, use natural language processing to identify important terms within narrative text and automatically map these terms to standard biomedical coding terminologies for research use within the STRIDE system.

ChartIndex receives clinical documents via STRIDE HL7 feeds from Stanford University Medical Center. It converts these documents into a format that is compliant with the HL7 Clinical Document Architecture (CDA) standard. A natural language parser then identifies all noun phrases within the document, while retaining the context of that phrase down to the document section and sentence level. Each noun phrase is then mapped to a standard biomedical terminology concept of the National Library of Medicine's (NLM) Unified Medical Language System (UMLS). ChartIndex enhances indexing precision using a method called "Contextual Indexing" to restrict the UMLS source terminologies selected to index each document section. CharIndex also automatically detects and tags negated terms (e.g. "no evidence of X"). The link between each indexed concept and the canonical term selected to represent it is stored within a relational database to facilitate rapid searching and retrieval of document sets.

1. Huang Y, Lowe HJ "A Novel Hybrid Approach to Automated Negation Detection in Clinical Radiology Reports." J Am Med Inform Assoc 2007 May-June; 3: 14: 304-311
2. Huang Y, Lowe HJ, Klein D, Cucina RJ "Improved identification of noun phrases in clinical radiology reports using a high-performance statistical natural language parser augmented with the UMLS specialist lexicon." J Am Med Inform Assoc 2005 May-Jun; 12: 3: 275-85
3. Huang Y, Lowe HJ "A Grammar-based Classification of Negations in Clinical Radiology Reports." AMIA Annu Symp Proc 2005; 988
4. Huang Y, Lowe HJ, Hersh WR "A pilot study of contextual UMLS indexing to improve the precision of concept-based representation in XML-structured clinical radiology reports." J Am Med Inform Assoc 2003 Nov-Dec; 10: 6: 580-7