Development of a Hypothesis-generating Support System Using Medical Records for Clinical Knowledge Acquisition

Abstract

The authors developed a support system for finding hypotheses about clinical knowledge from medical records. The support system has a retrieval function using medical records and uses coverage-oriented views displaying search results. The coverage-oriented views are snippets, graphs, and tag clouds. In addition, users can narrow the search results utilizing keywords, ranges of values, and labels of codes, corresponding to snippets, graphs, and tag clouds, 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 search results using coverage-oriented views based on interactive retrieval. We evaluated the developed system and received positive responses from medical staff. I. INTRODUCTION Clinical knowledge is expected to be extracted from accumulated medical records in order that higher quality medical care can be provided based on the clinical knowledge. In this study, clinical knowledge is defined as the universal relationship between subjective data and objective data of patients, diagnoses and medical practices for patients. Namely, we treat universal relationships in medical records as clinical knowledge. Generally, in order to obtain knowledge from large and variant information, we limit the amount of data to information having specific characters and conditions. When the information is narrowed, we can easily find trends of narrowed information so that we would generate hypotheses based on the trends. Narrowing information is realized by using a retrieval system. Moreover, we can easily grasp the trends by using coverage-oriented views.