Reliability and construct validity of the Italian EQ-5D-5 L in Parkinson’s disease: a cross-sectional study

Abstract

Abstract Background Parkinson’s disease (PD) substantially affects health-related quality of life (HRQoL). The EQ-5D-5 L is a preference-based patient-reported outcome measure that supports cross-condition comparison and health-economic evaluation, but its measurement properties have not been established in people with PD in Italy. This study evaluated short-term reproducibility and construct validity through a priori hypothesis testing; score distributions and inter-dimensional associations were also described. Methods This cross-sectional psychometric study included a 48-hour retest component. Participants completed the EQ-5D-5 L, the Parkinson’s Disease Questionnaire-39 (PDQ-39), and the 12-item Short Form Health Survey (SF-12). Short-term reproducibility was assessed using quadratic-weighted Cohen’s kappa and exact agreement for the five dimensions and single-measures, two-way random-effects, absolute-agreement intraclass correlation coefficients for the EQ-Index and EQ-VAS. Construct validity was examined using prespecified directional and magnitude hypotheses tested with Spearman correlations; Pearson correlations were used as sensitivity analyses. Post hoc analyses examined routine-care clinical measures and ordered motor-severity categories. Results Eighty participants completed the questionnaires, and 60 underwent retesting. Fifty-nine distinct health states were observed, with one participant reporting full health. Agreement was high for the EQ-Index (ICC = 0.901, 95% CI 0.840–0.940) and EQ-VAS (ICC = 0.938, 95% CI 0.899–0.963), whereas dimension-level kappa ranged from 0.632 to 0.916. All 20 directional hypotheses were confirmed, but only 6 of the 16 prespecified magnitude criteria were met. The strongest associations were observed with PDQ-39 Mobility (ρ = −0.786 and − 0.781), PDQ-39 Activities of Daily Living (ρ = −0.671 and − 0.685), and the SF-12 Physical Component Summary (ρ = 0.730 and 0.639) for the EQ-Index and EQ-VAS, respectively. In post hoc analyses, both summary scores declined monotonically across externally defined mild, moderate, and severe motor-severity categories (both p < 0.001). Conclusions The Italian EQ-5D-5 L showed high short-term reproducibility for its summary scores and evidence supporting selected aspects of construct validity, particularly convergence with physical functioning and daily activities. Evidence for discrimination from more distant HRQoL domains was limited, and the post hoc clinical findings require prospective confirmation. The instrument may complement disease-specific measures in group-level research and health-economic evaluation, but these findings do not support its use for monitoring change in individual patients.