P Nadkarni1, R Chen, C Brandt. 1. Yale University School of Medicine, New Haven, CT 06520-8009, USA. Prakash.Nadkarni@yale.edu
Abstract
OBJECTIVES: To explore the feasibility of using the National Library of Medicine's Unified Medical Language System (UMLS) Metathesaurus as the basis for a computational strategy to identify concepts in medical narrative text preparatory to indexing. To quantitatively evaluate this strategy in terms of true positives, false positives (spuriously identified concepts) and false negatives (concepts missed by the identification process). METHODS: Using the 1999 UMLS Metathesaurus, the authors processed a training set of 100 documents (50 discharge summaries, 50 surgical notes) with a concept-identification program, whose output was manually analyzed. They flagged concepts that were erroneously identified and added new concepts that were not identified by the program, recording the reason for failure in such cases. After several refinements to both their algorithm and the UMLS subset on which it operated, they deployed the program on a test set of 24 documents (12 of each kind). RESULTS: Of 8,745 matches in the training set, 7,227 (82.6 percent ) were true positives, whereas of 1,701 matches in the test set, 1, 298 (76.3 percent) were true positives. Matches other than true positive indicated potential problems in production-mode concept indexing. Examples of causes of problems were redundant concepts in the UMLS, homonyms, acronyms, abbreviations and elisions, concepts that were missing from the UMLS, proper names, and spelling errors. CONCLUSIONS: The error rate was too high for concept indexing to be the only production-mode means of preprocessing medical narrative. Considerable curation needs to be performed to define a UMLS subset that is suitable for concept matching.
OBJECTIVES: To explore the feasibility of using the National Library of Medicine's Unified Medical Language System (UMLS) Metathesaurus as the basis for a computational strategy to identify concepts in medical narrative text preparatory to indexing. To quantitatively evaluate this strategy in terms of true positives, false positives (spuriously identified concepts) and false negatives (concepts missed by the identification process). METHODS: Using the 1999 UMLS Metathesaurus, the authors processed a training set of 100 documents (50 discharge summaries, 50 surgical notes) with a concept-identification program, whose output was manually analyzed. They flagged concepts that were erroneously identified and added new concepts that were not identified by the program, recording the reason for failure in such cases. After several refinements to both their algorithm and the UMLS subset on which it operated, they deployed the program on a test set of 24 documents (12 of each kind). RESULTS: Of 8,745 matches in the training set, 7,227 (82.6 percent ) were true positives, whereas of 1,701 matches in the test set, 1, 298 (76.3 percent) were true positives. Matches other than true positive indicated potential problems in production-mode concept indexing. Examples of causes of problems were redundant concepts in the UMLS, homonyms, acronyms, abbreviations and elisions, concepts that were missing from the UMLS, proper names, and spelling errors. CONCLUSIONS: The error rate was too high for concept indexing to be the only production-mode means of preprocessing medical narrative. Considerable curation needs to be performed to define a UMLS subset that is suitable for concept matching.
Authors: S M Huff; R A Rocha; C J McDonald; G J De Moor; T Fiers; W D Bidgood; A W Forrey; W G Francis; W R Tracy; D Leavelle; F Stalling; B Griffin; P Maloney; D Leland; L Charles; K Hutchins; J Baenziger Journal: J Am Med Inform Assoc Date: 1998 May-Jun Impact factor: 4.497
Authors: John M Fisk; Pradeep Mutalik; Forrest W Levin; Joseph Erdos; Caroline Taylor; Prakash Nadkarni Journal: J Am Med Inform Assoc Date: 2003 Jan-Feb Impact factor: 4.497