The Functional Independence Measure: tests of scaling assumptions, structure, and reliability across 20 diverse impairment categories.

OBJECTIVE: The analysis presented here evaluated the psychometric properties of the Functional Independence Measure (FIM) as a summated rating scale within context of the 20 impairment categories of the FIM-Function Related Group (FIM-FRG) system.
DESIGN: This study involved a cross-sectional analysis of patient records, utilizing factor analysis and techniques of multitrait scaling to verify the summative properties of the motor and cognitive dimensions of the FIM and to study the statistical properties of admission FIM scores. PATIENTS: Included were a total of 93.829 patients discharged from 252 freestanding rehabilitation hospitals and units during calendar year 1992. Cases were excluded that had missing or out-of-range values or atypical lengths of stay. These criteria were developed previously in conjunction with an expert clinical panel and confirmed through statistical analyses.
RESULTS: Factor analyses supported the motor and cognitive dimensions across all 20 impairment categories. The resulting subscales exceeded minimum criteria for item internal consistency in 96.9% of tests and item discriminant validity in 100% of tests. Reliability coefficients for each impairment category for both subscales ranged from .86 to .97. There were no major ceiling effects, but patients in certain impairment categories were unable to climb stairs at admission.
CONCLUSION: The psychometric properties of the summated FIM compare favorably to most standardized health measures used in medical practice. Findings provide support for the motor and cognitive subscales as used in the FIM-FRGs. As a unidimensional scale, the FIM quantifies care burden. Split into the motor and cognitive (as used in the FIM-FRGs) it distinguishes physical disabilities from those arising from communication or cognitive difficulties.