2017 Volume 8 Pages 21-29
Tokunaga M, Tori K, Eguchi H, Kado Y, Ikejima Y, Ushijima M, Miyabe S, Tsujimoto S, Fukuda E. The stratification of motor FIM and cognitive FIM and the creation of four prediction formulas to enable higher prediction accuracy of multiple linear regression analysis with motor FIM gain as the objective variable ― An analysis of the Japan Rehabilitation Database. Jpn J Compr Rehabil Sci 2017; 8: 21-29.
Objective: The aim of our study was to stratify the contributing factors in order to increase the prediction accuracy of the multiple linear regression analysis with motor FIM gain as the objective variable.
Methods: The subjects for our study were 2,542 stroke patients. In the multiple linear regression analysis with motor FIM gain as the objective variable, eight contributing factors were stratified. Prediction formulas were created and the correlation between the measured motor FIM gain values and the predicted values was investigated.
Results: The correlation coefficient was higher with the stratification of gender (0.509), stroke type (0.512), number of hospital days (0.516), days from onset to admission (0.518), modified Rankin Scale before onset (0.520), age (0.541), cognitive FIM at admission (0.588) and motor FIM at admission (0.641), than with the use of one prediction formula (0.507), and it was 0.653 with stratification into four groups with the two factors of motor FIM and cognitive FIM at admission.
Conclusion: By stratifying the contributing factors, we were able to increase the prediction accuracy of motor FIM gain.