Improvement of predictive accuracies of functional outcomes of stroke patients by machine learning models

2023 Fiscal Year Final Research Report

• Project/Area Number: 22K21225
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Multi-year Fund
• Review Section: 0909:Sports sciences, physical education, health sciences, and related fields
• Research Institution: National Center of Neurology and Psychiatry
• Principal Investigator: MIYAZAKI YUTA 国立研究開発法人国立精神・神経医療研究センター, 病院 身体リハビリテーション部, 医師 (70966192)
• Project Period (FY): 2022-08-31 – 2024-03-31
• Keywords: リハビリテーション / 機械学習 / 予後予測 / 脳卒中 / FIM / ADL

Outline of Final Research Achievements

Stroke is one of the conditions requiring caregiving, underscoring the importance of prognosis. Previous research utilizing multiple regression analysis for stroke prognosis has indicated potential diminished predictive accuracy due to the non-linear nature of clinical data. Hence, this study investigated whether machine learning capable of analyzing non-linear data could enhance prognosis prediction for stroke patients compared to traditional regression analysis.
This study performed prognosis prediction of discharge Functional Independence Measure (FIM) scores, a prominent assessment index of activities of daily living (ADL), for 1046 stroke patients based on age and FIM scores at admission. The analysis utilized multiple regression analysis along with five machine learning algorithms. This study demonstrated that machine learning improves prognostic prediction accuracy compared to multiple regression analysis.

Free Research Field

リハビリテーション

Academic Significance and Societal Importance of the Research Achievements

脳卒中は、介護が必要となる代表的疾患の一つである。また、介護は退院後の患者家族の生活にも影響を与える。そのため、脳卒中発症早期から退院時ADLの予後予測が可能となれば、退院後生活の様々な準備を行うための時間的猶予を持てる。本研究は、機械学習が重回帰分析よりも脳卒中患者の予後予測精度を改善することを報告した。本研究は、機械学習が、脳卒中患者の予後予測をより高い精度で可能とし、患者や家族の退院後の生活に有益な情報を提供できる可能性が高いことが示唆した。