Abstract
To examine the validity of the Patient-Specific Functional Scale (PSFS) for the assessment of group-level change and between-group discrimination in group-level data.
We collected complete baseline and follow-up PSFS data in 1,181 consecutive patients reporting to physical therapy with a musculoskeletal disorder. Physical function was assessed at the baseline and final physical therapy visits using the PSFS and four region-specific patient-reported outcome (PRO) measures: The Neck Disability Index, Oswestry Disability Index, Upper Extremity Functional Index, and Lower Extremity Functional Scale. Global Rating of Change (GROC) was assessed at discharge. We assessed data distribution and floor and ceiling effects. Correlation and linear regression analyses assessed concurrent, convergent, and discriminant validities of PSFS baseline, final, and change scores across the cohort. One-way ANOVA was used to test for differences in PSFS scores among strata defined by region-specific PRO score and GROC. Cohen's d was used to assess responsiveness.
Results supported the concurrent, convergent, and discriminant validities (all P < 0.001), scale consistency (P < 0.001 omnibus, P < 0.05 post hoc tests), distribution, and responsiveness of the PSFS for both between-group discrimination and assessment of change over time in group-level data. The PSFS performed better than comparison PRO measures in most comparisons.
These results indicate that the PSFS is an appropriate measure for statistical comparisons in clinical research.