Structural validity of the Brazilian version of the Shoulder Pain and Disability Index in patients with chronic shoulder pain

Abstract

BackgroundThe Shoulder Pain and Disability Index (SPADI) is a measuring instrument for pain and disability in patients with shoulder dysfunction. However, its structure is controversial (ie, 1 or 2 domains). PurposeTo investigate the structural validity of the Brazilian version of the SPADI in patients with chronic shoulder pain. Study DesignCross-sectional study. MethodsBrazilian Portuguese speakers of both genders aged 18 years or older with shoulder pain for at least 90 days (chronic) were included in this study. The structure of the SPADI was assessed by means of exploratory factor analysis with the implementation of parallel analysis and by confirmatory factor analysis. For the latter, the following goodness-of-fit indices were considered: root mean square error of approximation with 90% confidence interval, comparative fit index, Tucker–Lewis index, standardized root mean square residual, and chi-square/degrees of freedom. ResultsA total of 103 participants were included in the study. The original version of the SPADI with 2 domains and 13 items showed high covariance between domains (0.98), suggesting one-dimensionality. After exploratory factor analysis with the implementation of parallel analysis, one-dimensionality was found in the SPADI. Comparing models with 1 and 2 domains by means of confirmatory factor analysis, both presented adequate values (chi-square/degrees of freedom < 3; comparative fit index and Tucker–Lewis index > 0.90; root mean square error of approximation and standardized root mean square residual < 0.08). However, the model with one domain is more suitable for presenting lower values of Akaike information criterion (3308.713) and Bayesian information criterion (3377.216). ConclusionThe Brazilian version of the SPADI has a one-dimensional structure with 13 items (“shoulder impairment” construct), in contrast to the version with 2 domains.