*Original Article*
## Increasing the prediction accuracy of FIM gain by adding FIM
improvement for one month from admission to the explanatory
variables in multiple regression analyses

Makoto Tokunaga, MD, PhD, Kenichi Katsura, MD, Kaori Tokisato, MD, Shoji Honda, MD,
Toshihito Nakanishi, MD, Seiko Takai, MD, Yukihiko Nakashima, OT, Shinichi Nojiri, PT,
Susumu Watanabe, MD, PhD

Jpn J Compr Rehabil Sci 8: 16-20, 2017

Objective: The coefficient of determination in multiple
linear regression analyses using the gain in Functional
Independence Measure (FIM) as objective variables is
low. We aimed to improve the prediction accuracy of
FIM gain by adding FIM improvement for one month
from admission to the explanatory variables.

Methods: A total of 547 stroke patients admitted to a
Kaifukuki (convalescent stage) rehabilitation ward
were included. In multiple regression analyses using
motor FIM gain as the objective variable, we created a
prediction formula using only data available at
admission (model 1), and a prediction formula adding
FIM improvement for one month from admission
(model 2). The coefficient of determination adjusted
for the degrees of freedom R*2 and the residuals
obtained by subtracting the predicted value from the
measured value of motor FIM gain were investigated.

Results: The R*2 of model 1 was 0.364 and that of
model 2 was 0.711. The residual of model 1 was 0 }
12.3 and that of model 2 was 0 } 8.3. In model 2, the
standard deviation of the residual was reduced.

Conclusion: Adding FIM improvement for one month
to the explanatory variables increased the prediction
accuracy of FIM gain.

Key words: FIM gain, multiple linear regression
analysis, one month from admission, explanatory
variable, stroke

Contents (volume 8)