Psychometric Properties of the Obstacles to Return-to-Work Questionnaire in a Brazilian context

Abstract

Purpose To evaluate the psychometric properties of the Obstacles to Return-to-Work Questionnaire (ORTWQ) among Brazilian workers on sick-leave due to musculoskeletal disorders. Methods Confirmatory factor analysis was conducted to evaluate the factor structure validity of the ORTWQ. Model fit indices and salience of factor loadings were assessed. The convergent validity was estimated using the Average Variance Extracted (AVE) and Composite Reliability (CR). The correlational analysis was verified using the Spearman Correlation between the ORTWQ and other specific tools. Discriminant Validity, internal consistency, stability (test–retest) and floor/ceiling effect were also assessed. Results A total of 301 participants completed the ORTWQ with a mean age of 45.0 (9.9) years. After refinement, the factor structure indexes of the oblique model were [χ2/df = 1.8; CFI = 0.9; TLI = 0.9; PGFI = 0.7; PCFI = 0.8; RMSEA = 0.05 (90% CI 0.05–0.06)]. Only Depression, Physical Workload and Perceived Prognosis subscales presented suitable AVE indices: 0.63, 0.51 and 0.52 respectively. The correlations between ORTWQ and the other questionnaires were appropriate for almost all subscales. Reliability evaluation showed adequate estimates for all subscales except for the Worry Due to Sick-Leave (CR = 0.45; α = 0.44; ICC = 0.69). A higher order hierarchical model is suggested, in order to estimate an overall score to ORTWQ in a Brazilian population. Conclusions The psychometric properties of the Brazilian version of the ORTWQ were evaluated and after refinement, the validity, reliability and floor/ceiling effects indexes were suitable when applied to a sample of Brazilian workers on sick-leave due to musculoskeletal disorders. However, the factor structure presented some issues regarding convergent and discriminant validity.