Evaluation of the Structural and Construct Validity of the Credibility and Expectancy Scale for Patients With Musculoskeletal Disorders

Abstract

Background The Credibility Expectancy Questionnaire (CEQ) includes three items each on the credibility and expectancy subscales. Credibility indicates to what extent the treatment is reasonable, and expectancy indicates to what extent the treatment is expected to be effective. The CEQ has been assumed to have a two-factor structure: credibility and expectancy, among patients receiving psychotherapy. However, its internal structure has been unknown to patients receiving physical therapy for musculoskeletal disorders. This study aimed to explore the internal structure of the CEQ and preliminary investigate the construct validity of the CEQ among patients receiving physical therapy for musculoskeletal disorders. Methodology A multi-center prospective cohort study was conducted. Data from 100 patients receiving outpatient physical therapy for musculoskeletal disorders was collected using an anonymous paper-based survey. The initial survey was conducted immediately before the initial physical therapy session, and the second survey was conducted after the third to seventh physical therapy sessions. The Patient Specific Functional Scale 2.0 (PSFS 2.0) was collected in both surveys, and the CEQ and an 11-point global rating of change scale (GRCS) were collected in the second survey. Exploratory factor analysis was conducted for the CEQ, and internal consistency was assessed for each subscale and an identified factor structure. Convergent validity in construct validity was also assessed with the hypothesis that Pearson’s r values of each CEQ factor score to the PSFS 2.0 change scores and GRCS would range from 0.4 to 0.6. Results An exploratory factor analysis revealed a one-factor structure, where the percentage of the variance for the extraction sums of squared loadings was 62.8%. Cronbach’s alpha was 0.89 for all items, 0.91 for the credibility subscale, and 0.75 for the expectancy subscale. Hypothesized correlations to the PSFS 2.0 change score and GRCS were detected with the CEQ total score (r = 0.48 for the PSFS 2.0 change score and r = 0.59 for the GRCS) and each subscale score (credibility subscale, r = 0.48 for the PSFS 2.0 change score and r = 0.49 for the GRCS; and expectancy subscale, r = 0.43 for the PSFS 2.0 change score and r = 0.62 for the GRCS). Conclusion A single-factor internal structure of the CEQ was detected among patients receiving physical therapy for musculoskeletal disorders. Additionally, preliminary evidence of construct validity was detected with convergent validity between the CEQ and functional and perceived improvement.