OBJECTIVE: To develop a short version of the motor Functional Independence Measure (FIM) for use in long-term care settings. PARTICIPANTS: For model construction, the participants were 398 community-dwelling persons with disability (mean age 79.3 years (SD 10.3)) who were receiving visiting nurse services. For cross-validation, 169 patients with stroke (mean age 78.0 years (SD 11.2)) in the chronic phase and 187 patients with stroke (mean age 63.4 years (SD 12.7)) in the recovery phase. DESIGN: Model construction and cross-validation study. MAIN OUTCOME MEASURES: The second power of correlation coefficient (R2) was used for agreement analysis between the short and the full version. Cross-validation of the models was estimated with the intraclass correlation coefficient (ICC). RESULTS: Five to 7 motor FIM items were selected for the models based on Rasch calibration and consideration of internal consistency. Total motor FIM was estimated with the 6-item and 7-item models with regression analysis, which yielded high correlations with the original 13-item motor FIM score (R2>0.95). Regression formulas derived from the models could estimate total motor FIM scores accurately in the 2 cross-validation samples (ICC > 0.98). CONCLUSION: The short version of the motor FIM developed is a useful measure of functional status, not only in long-term care but in the recovery phase rehabilitation settings.
OBJECTIVE: To develop a short version of the motor Functional Independence Measure (FIM) for use in long-term care settings. PARTICIPANTS: For model construction, the participants were 398 community-dwelling persons with disability (mean age 79.3 years (SD 10.3)) who were receiving visiting nurse services. For cross-validation, 169 patients with stroke (mean age 78.0 years (SD 11.2)) in the chronic phase and 187 patients with stroke (mean age 63.4 years (SD 12.7)) in the recovery phase. DESIGN: Model construction and cross-validation study. MAIN OUTCOME MEASURES: The second power of correlation coefficient (R2) was used for agreement analysis between the short and the full version. Cross-validation of the models was estimated with the intraclass correlation coefficient (ICC). RESULTS: Five to 7 motor FIM items were selected for the models based on Rasch calibration and consideration of internal consistency. Total motor FIM was estimated with the 6-item and 7-item models with regression analysis, which yielded high correlations with the original 13-item motor FIM score (R2>0.95). Regression formulas derived from the models could estimate total motor FIM scores accurately in the 2 cross-validation samples (ICC > 0.98). CONCLUSION: The short version of the motor FIM developed is a useful measure of functional status, not only in long-term care but in the recovery phase rehabilitation settings.