@article{oai:kanazawa-u.repo.nii.ac.jp:00016174,
 author = {Yamada, Tomomi and Someya, Fujiko},
 issue = {1},
 journal = {金沢大学つるま保健学会誌 = Journal of the Tsuruma Health Science Society Kanazawa University},
 month = {Jul},
 note = {The aim of this study was to clarify the usefulness of assessing dual task performance for fall prediction. In experiment 1, we prospectively investigated the new fall experiences of 108 participants (50 elderly daycare users and 58 fall prevention class participants) and produced a fall prediction model. The fallers represented the subjects who suffered a new fall experience during the follow up 12- month period, whereas the non-fallers did not fall during this period. The two groups were compared with regard to their age, sex, history of falling, medication related to falling, grip strength, gait speed, step length, dynamic standing balance, the Timed Up and Go (TUG) test score, dt-TUG score, %TUG, and mini－mental state examination score. Multivariate logistic regression analyses were performed to produce a fall prediction model. In experiment 2, we investigated the accuracy of the fall prediction model using 39 newly recruited individuals. In experiment 1, 8 (13.8%) of the fall prevention class participants fell within 12 months. Among the fall prevention class participants, there were significant differences in grip strength, gait speed, step length, TUG, and dt-TUG between the fallers and non-fallers. In stepwise logistic regression analysis, two factors were found to have value for fall prediction ; i.e., falling within one year preceding the baseline and TUG. In experiment 2, 12 people were predicted to fall within the 12-month follow-up period ; however, only 5 people actually fell. Three of them were predicted to fall, but 2 of them were not. The success ratio of the fall prediction model was 60%. Nine of the 12 people that were predicted to fall did not. Our result showed that TUG was the most important factor for fall prediction; however, dt-TUG was also related to fall prediction in individuals that were not in receipt of care services. Therefore, dual task performance could be a useful reference factor for fall prediction.　高齢在宅生活者を対象に、転倒予測における二重課題の評価は重要因子となるかを研究 目的とした。今回は、以前評価を行った転倒予防教室参加者５８名、デイケア利用者５０名に 対し、実験１ではその後１年間の転倒有無を追跡調査し、年齢、性別、服薬、歩行速度、 歩幅、動的立位バランス、Timed Up＆Go（以下TUG）、TUGと加算の二重課題（以下dt- TUG）を因子として検討し、転倒予測モデルを作成した。実験２では新たに３９名の対象者 に対して調査し、転倒予測モデルの有用性を検証した。 　結果、実験１では転倒予防教室参加者で転倒した者は８名であり、転倒と各項目の分析 では、握力、歩行速度、歩幅、TUG、dt-TUGの５項目に有意差がみられた。デイケア利用 者は５名の追跡調査ができず、転倒した者は１６名であり、転倒と項目の有意な関連がな かった。転倒予測モデルについて転倒予防教室参加者を用い多重ロジスティック回帰分析 をした結果、TUGと過去１年間の転倒経験が抽出された。実験２では３９名のうち、予測モ デルで転倒が予測された者は１２名であった。追跡調査で転倒した者は５名で、そのうち３ 名は予測できたが、２名は予測できなかった。 　以上より、転倒予測に有用な因子はTUGであり、dt-TUGは参考程度の因子であった。ま た、転倒予測については、今回の項目の身体機能と転倒経験だけでは高い抽出率を得るの は困難と示唆された。, [原著:Originals]},
 pages = {19--25},
 title = {Is dual task performance the most useful factor for fall prediction ?},
 volume = {35},
 year = {2011}
}