Predicting FIM gain in stroke patients by adding median FIM gain stratified by FIM score at hospital admission to the explanatory variables in multiple regression analysis—An analysis of the Japan Rehabilitation Database

Stroke survivors don't give a crap about what the predictions are except that they are used to not do aggressive treatments. And I bet the patient and family are not told that the treatments are being reduced.
https://www.jstage.jst.go.jp/article/jjcrs/7/0/7_13/_article

http://doi.org/10.11336/jjcrs.7.13

Original Article

Original Article

Makoto Tokunaga¹⁾, Yoshitaka Mori²⁾, Yoshitaka Ogata²⁾, Yasunori Tanaka²⁾, Kouichi Uchino³⁾, Yuki Maeda³⁾, Machiko Kamiyoshi³⁾

1) Department of Rehabilitation, Kumamoto Kinoh Hospital 2) Department of Physical Therapy, Kumamoto Kinoh Hospital 3) Department of Occupational Therapy, Kumamoto Kinoh Hospital

Released on J-STAGE 20160513  

Keywords: Functional Independence Measure, FIM gain, multiple regression analysis, explanatory variable, stroke

• Full Text PDF [1101K]

• Abstracts
• References(13)
• Supplementary materials(1)

Tokunaga M, Mori Y, Ogata Y, Tanaka Y, Uchino K, Maeda Y, Kamiyoshi M. Predicting FIM gain in stroke patients by adding median FIM gain stratified by FIM score at hospital admission to the explanatory variables in multiple regression analysis —An analysis of the Japan Rehabilitation Database—. Jpn J Compr Rehabil Sci 2016; 7: 13-18.
Objective: To clarify whether the accuracy of predicting motor Functional Independence Measure (FIM) gain in stroke patients can be improved by calculating median values of motor FIM gain (median mFIM gain) stratified by motor FIM score at hospital admission, then inserting these standard gain values in multiple regression analysis.
Methods: The subjects were 2,542 stroke patients registered in the Japan Rehabilitation Database. Motor FIM score at admission was stratified into 39 groups at 2-point intervals and “median mFIM gain” was calculated for each group. With motor FIM gain as the objective variable, multiple regression analysis was performed with and without median mFIM gain in the explanatory variables. Then, correlations were examined between measured values and predicted values of motor FIM gain.
Results: Adding median mFIM gain to the explanatory variables increased the correlation coefficient of measured values and predicted values of motor FIM gain from 0.507 to 0.638.
Conclusion: Adding median mFIM gain to the explanatory variables can improve the accuracy of multiple regression analyses to predict motor FIM gain.