The LBI-method for automated indexing of diagnoses by using SNOMED. Part 2. Evaluation

• Published in: International journal of bio-medical computing Vol. 38; no. 2; pp. 101 - 108
• Main Authors: Brigl, Birgit, Mieth, Markus, Haux, Reinhold, Glück, Ewald
• Format: Journal Article
• Language: English
• Published: Ireland Elsevier B.V 01.02.1995
• Subjects: Abstracting and Indexing as Topic; Automated indexing of diagnoses; Automatic Data Processing; Diagnosis; Evaluation Studies as Topic; Humans; Information retrieval; Information Storage and Retrieval; Medical Records; Natural Language Processing; Nomenclature; Patient Discharge; SNOMED; Software Validation; Subject Headings; Terminology as Topic; Information retrieval; SNOMED; Natural language processing; Automated indexing of diagnoses; Nomenclature
• ISSN: 0020-7101
• DOI: 10.1016/0020-7101(94)01062-6

We present a simple, formal, lexicon-based method for automated indexing of diagnoses based on the Systematized Nomenclature of Medicine (SNOMED), called LBI-method. Part 1 gave an introduction to the LBI-method and presented its realisation as application system SALBIDH. Part 2 presents the design and the results of an evaluation study to judge the quality of the LBI-method. In this evaluation study the quality of automated indexing as well as the quality of the retrieval of patient data by using automated indexed diagnoses was examined. The results show that the retrieval based on SNOMED indices is at least as good as the retrieval based on ICD classes despite a lot of indexing errors. From this we gather that our system is not yet good enough for immediate routine use but that an appropriate indexing quality and, as a result, a higher retrieval quality can be achieved after few improvements of the LBI-method, especially after revision of the lexicons.