A Medical Decision Support System Using Text Mining to Compare Electronic Medical Records

摘要

The electronic medical records (EMRs) contain information about the patient such as their date of birth and blood type as well as other medical information such as prescription history and previous syndromes. Physicians usually have limited time to identify critical information on medical records and to provide a summary before they make a decision. However, the content of EMRs usually be complicated, repeated, and contain many consistency problems; these issues are not only cost a lot of time for physicians to filter information out from the medical records but also increase the probability of wrong medical decisions. Therefore, this study proposed a new EMR interface to identify the new medical information such as new syndromes or the turning point in the medical records. The Metathesaurus database which contains medical information such as medical terms or classification codes in the Unified Medical Language System will be used. This study uses MetaMap tools to compare medical terms within EMRs using MetaMap and also compares the vocabulary using the bigram technique to highlight the similarities in the EMR.