The authors developed a support system for finding hypotheses about clinical knowledge from medical records accumulated electronically in hospital information systems. In order to support the development of various hypotheses, we matched a variety of medical and clinical information with each case and utilized the information comprehensively. This information includes clinical documents, examination information, medical practice information, and other minor factors. The system uses coverage-oriented views showing targeted medical records, in whole or in part. We selected appropriate methods to provide coverage-oriented views of natural language data, numerical data, and coded data, making use of snippets, graphs, and tag clouds, respectively. In addition, users can narrow the search results utilizing keywords, ranges of values, and labels of codes, respectively. The coverage-oriented views and the narrowing methods utilizing the views make the retrieval system interactive. We predicted that users could find hypotheses by grasping features of targeted cases using coverage-oriented views based on interactive retrieval. We evaluated the developed prototype in terms of its potential for helping users find hypotheses about clinical knowledge and we received positive responses from doctors.
[1]
R O Morgan,et al.
Diabetes increases the risk of hepatocellular carcinoma in the United States: a population based case control study
,
2005,
Gut.
[2]
David R. Karger,et al.
Enabling web browsers to augment web sites' filtering and sorting functionalities
,
2006,
UIST.
[3]
Satoshi Nakamura,et al.
Rerank-by-Example: Efficient Browsing of Web Search Results
,
2007,
DEXA.
[4]
T Suzuki,et al.
Automatic DPC code selection from electronic medical records: text mining trial of discharge summary.
,
2008,
Methods of information in medicine.