Okamoto Kazuya, Uchiyama Toshio, Takemura Tadamasa, Kume Naoto, Adachi Takayuki, Kuroda Tomohiro, Uchiyama Tadasu, Yoshihara Hiroyuki
Division of Medical Information Technology & Administration Planning, Kyoto University Hospital, Japan.
Stud Health Technol Inform. 2013;192:1031.
In Japan, medical staff must select a diagnosis procedure combination (DPC) code for each inpatient upon admission. We report on the development and evaluation of a supporting system for DPC code selection. This system, based on a machine learning method developed by Okamoto et al., makes DPC code suggestions that are derived from medical practice information pertaining to inpatients. The use of the suggestions helps medical staff select an appropriate DPC code for each inpatient. We asked health information management professionals to evaluate the system and to compare the suggested DPC codes with those selected by doctors. They reported that the system was generally useful and that using this system they could find some cases of hospitalized patients whose DPC codes needed correction. However, they also determined the precision of the system needs improvement.
在日本，医护人员必须在每位住院患者入院时选择一个诊断程序组合（DPC）代码。我们报告了一个用于DPC代码选择的支持系统的开发和评估情况。该系统基于冈本等人开发的机器学习方法，根据与住院患者相关的医疗实践信息生成DPC代码建议。这些建议有助于医护人员为每位住院患者选择合适的DPC代码。我们请健康信息管理专业人员对该系统进行评估，并将建议的DPC代码与医生选择的代码进行比较。他们报告称，该系统总体上很有用，使用该系统可以发现一些住院患者的DPC代码需要修正的情况。然而，他们也确定该系统的精度有待提高。
