A new method for predicting the timing of chronic disease severity based on electronic medical records

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K10348
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 58010:Medical management and medical sociology-related
• Research Institution: Kochi University
• Principal Investigator: Hatakeyama Yutaka 高知大学, 教育研究部医療学系連携医学部門, 教授 (00376956)
• Co-Investigator(Kenkyū-buntansha): 奥原 義保 高知大学, 教育研究部医療学系連携医学部門, 教授 (40233473) ; 兵頭 勇己 高知大学, 教育研究部医療学系連携医学部門, 助教 (50821964)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 医療データ解析 / 予測モデル

Outline of Final Research Achievements

The prediction algorithms for time series laboratory data were developed based on medical text data such as progress records and medical interview items, which were combed with laboratory test values and these data. The prediction results confirmed that the combination of these information improved the accuracy. These algorithms showed that patient status can be estimated even from unstructured data, and that more detailed patient status can be obtained based on the integration of patient information. These results suggest that the construction algorithms will be useful for hospital information systems, which will be able to acquire various types of data in the future.

Free Research Field

医療情報学

Academic Significance and Societal Importance of the Research Achievements

手法の新規性としては、経過記録などの非構造化データから患者状態の定量的な指標に変換して検査データなどの構造化データと統合して処理を行った点が挙げられる。電子カルテ情報以外のテキスト情報が電子情報として蓄積され始め膨大なデータとなり、これらの非構造化データを解析する需要が増大すると考えられる。そのため、膨大な医療データに対して統合処理を行う本手法は今後さらに必要とされる手法である。