Article

OBJECTIVE: To perform a comprehensive analysis of the psychometric properties and
dimensionality of the Upper Limb Functional Index (ULFI) using both classical
test theory and Rasch analysis (RA). DESIGN: Prospective, single-group
observational design. SETTING: Freestanding rehabilitation center. PARTICIPANTS:
Convenience sample of Italian-speaking subjects with upper limb musculoskeletal
disorders (N=174). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: The
Italian version of the ULFI. Data were analyzed using parallel analysis,
exploratory factor analysis, and RA for evaluating dimensionality, functioning of
rating scale categories, item fit, hierarchy of item difficulties, and
reliability indices. RESULTS: Parallel analysis revealed 2 factors explaining
32.5% and 10.7% of the response variance. RA confirmed the failure of the
unidimensionality assumption, and 6 items out of the 25 misfitted the Rasch
model. When the analysis was rerun excluding the misfitting items, the scale
showed acceptable fit values, loading meaningfully to a single factor. Item
separation reliability and person separation reliability were .98 and .89,
respectively. Cronbach alpha was .92. CONCLUSIONS: RA revealed weakness of the
scale concerning dimensionality and internal construct validity. However, a set
of 19 ULFI items defined through the statistical process demonstrated a
unidimensional structure, good psychometric properties, and clinical
meaningfulness. These findings represent a useful starting point for further
analyses of the tool (based on modern psychometric approaches and confirmatory
factor analysis) in larger samples, including different patient populations and
nationalities.
CI - Copyright (c) 2015 American Congress of Rehabilitation Medicine. Published by
Elsevier Inc. All rights reserved.

Langue : ANGLAIS