LAUSR

OBJECTIVE To compare assay sensitivity of the Visual Analogue Scale (VAS) for global osteoarthritis pain and the Western Ontario and McMaster University (WOMAC) pain subscale, and the associated between-trial heterogeneity in effect sizes (ES). DESIGN We included trials with placebo, sham or non-intervention control that included at least 100 patients with hip or knee osteoarthritis per arm, reporting both VAS and WOMAC pain scores. ES were calculated as between-group difference in means divided by the pooled standard deviation and compared using a paired t-test. ES and τ2 as a measure of between-trial heterogeneity were combined using random-effects meta-regression with robust variance estimation to account for the correlation of data within trials and meta-analyses. RESULTS Twenty-eight trials with 44 randomized comparisons were included. In 28 comparisons (64%), ES from VAS favoured the intervention more than those from WOMAC pain (P=0.003). Twenty-six p-values (59%) were smaller according to VAS (P=0.008). The 44 comparisons contributed to 12 meta-analyses. Eleven meta-analyses (92%) showed larger benefits of interventions according to VAS, with a combined overall difference in ES of -0.08 (95% CI -0.14 to -0.02). τ2 was similar for VAS and WOMAC pain (difference in τ2, -0.003, 95% CI -0.009 to 0.004). CONCLUSION The VAS for global pain had slightly higher assay sensitivity at trial and meta-analysis levels than the WOMAC pain subscale without relevant increase in between-trial heterogeneity.