AI-based large-scale text data analysis for optimizing outcomes of specific health guidance.

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K10620
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 58030:Hygiene and public health-related: excluding laboratory approach
• Research Institution: Shiga University (2021-2022); Kyoto University (2019-2020)
• Principal Investigator: Ikenoue Tatsuyoshi 滋賀大学, データサイエンス学系, 講師 (70761443)
• Co-Investigator(Kenkyū-buntansha): 山田 ゆかり 京都大学, 医学研究科, 特定講師 (00306846) ; 福間 真悟 京都大学, 医学研究科, 准教授 (60706703)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 特定保健指導 / 特定健診 / 機械学習 / テキスト解析 / ディープラーニング

Outline of Final Research Achievements

In this study, a longitudinal evaluation of the content and results of specific health guidance was conducted by comparing health guidance records with health check-up information. The health guidance records were classified and analysed using machine learning, and it was visualised that the content of health guidance could be classified into four or six categories. Furthermore, a model for predicting a 5 cm reduction in abdominal circumference at the health check-up one year later was constructed from the health guidance records using deep learning, and a model with an accuracy of 62% was constructed. On the other hand, existing analysis methods were used to examine the recipients' reactivity to 'specific health guidance.' It was found that attitudes and behaviours towards obesity were associated with weight loss due to health guidance.

Free Research Field

衛生学および公衆衛生学分野関連：実験系を含まない

Academic Significance and Societal Importance of the Research Achievements

本研究はこれまでブラックボックスとなっていた特定保健指導がその結果にどのように影響するかを機械学習を用いて検討した。。今後、より精緻なモデルの作成により保健指導の質的な向上に寄与することが期待される。