薬理と治療

Abstract

Background　In recent years, there has been an increasing expectation to use real‒world data in the development of drugs and medical devices. However, medical information from electronic medical records is not data collected following research plans. This may lead to problems in which the data is not collected with the quality expected for research, or the data is biased and is different from the target population. Therefore, when using medical information data, it is important to understand the characteristics and properties of the data in advance and to confirm and select the data for the research purpose. Therefore, we conducted a survey to clarify the characteristics and properties of medical information data from Osaka University Hospital（OUH）. Method　We extracted items frequently used in clinical studies from medical information data of all patients who had their first medical examination at OUH in the past 10 years. Among the data of each item, we examined the distribution and transition over time of patient background and the frequency of missing or abnormal data. Results and Conclusions　This study showed that the medical information data of OUH includes some items that were entered according to the status of missing data and the original rules. When using a database, it is necessary to check the error data and examine the cause and handling. Understanding the characteristics and properties of medical information data in advance is expected to facilitate identifying the cause of errors and the selection of databases will be more efficient.（Jpn Pharmacol Ther 2022；50 suppl 1：s37‒50）