Significance of updating discharge FIM prediction at one month after admission using multiple regression analysis.

Tokunaga M, Sannomiya K, Imada Y. Significance of updating discharge FIM prediction at one month after admission using multiple regression analysis. Jpn J Compr Rehabil Sci 2026; 17: 16-23. Objective: To clarify the significance of updating predictions at one month after admission in multiple regression analyses that predict the motor component of the Functional Independence Measure (mFIM) at discharge in stroke patients. Methods: A total of 849 stroke patients admitted to a convalescent rehabilitation ward were included. The dependent variables were discharge mFIM (S prediction) and mFIM effectiveness (E prediction). Independent variables consisted of either admission data alone or admission data plus mFIM improvement during the first month after admission (plus prediction). Four models were constructed: S prediction, S plus prediction, E prediction, and E plus prediction. Absolute residuals were compared among the four groups using the Kruskal-Wallis test. When significant differences were observed, multiple comparisons were performed using the Steel-Dwass test. Results: The absolute residuals were 9.3 ± 7.2 for S prediction, 6.8 ± 5.3 for S plus prediction, 7.6 ± 7.0 for E prediction, and 6.3 ± 5.7 for E plus prediction. Significant differences were observed among the four models. Multiple comparisons revealed that S plus prediction had significantly smaller residuals than S prediction; E plus prediction was smaller than E prediction; E prediction was smaller than S prediction; and E plus prediction was smaller than S plus prediction. Conclusion: Predicting mFIM effectiveness and converting it to discharge mFIM yielded more accurate predictions than directly predicting discharge mFIM. Updating predictions at one month after admission further improved the accuracy of discharge mFIM prediction.