Relative efficiency and validity properties of a visual analogue vs a categorical scaled version of the Western Ontario and McMaster Universities Osteoarthritis (WOMAC) Index: Spanish versions

Abstract

Objective: To compare the performance of visual analogue (VAS) vs categorical (CT) scaled versions of the three subscales (pain, stiffness and difficulty) of the WOMAC Index in patients with knee osteoarthritis.

Material and methods: Patients with knee OA (at least grade II of Kellgren & Lawrence classification) were treated for a 6-weeks period with an NSAID. The following measures were applied at baseline and after treatment: VAS and CT WOMAC scales, Lequesne Index, pain on VAS, and global assessment by patient and observer. Construct Validity was determined by correlation of each of the subscale scores with the other outcomes both at baseline and after treatment (Pearson's test); responsiveness comparing baseline vs final status by Wilcoxon's test; reliability by analysis of the internal consistency using Cronbach's alpha at baseline and after the treatment period; and test-retest reliability by Kendall's Tau-c statistics. Finally, correlation and analysis of the relative efficiency (RE) between the subscales of both formats were tested.

Results: Seventy-three (8 men and 65 women) patients with a mean disease duration from first diagnosis of 69 months (3 to 254) completed the study. The following were the observed values for the instruments' psychometric properties: (1) construct validity: correlations ranged from 0.30 to 0.84 for VAS and 0.27 to 0.77 for CT subscales; (2) responsiveness: achieved P values for the pain, stiffness and difficulty scales were P<0.0001, P=0.002 and P<0.0001 in VAS and P=0.003, P<0.0001 and P=0.001 in CT format respectively; (3) internal consistency: the obtained Cronbach's alpha coefficients ranged from 0.71 to 0.97 for the VAS and 0.64 to 0.95 for the CT subscales; (4) test-retest reliability: correlation coefficients ranged from 0.36 to 0.76 for VAS and 0.34 to 0.52 for CT subscales; and (5) the relative efficiency of the subscales in VAS vs CT format were 2.20, 0.91 and 1.91 for pain, stiffness and difficulty respectively. Significant correlations between subscales in both formats ranging from 0.72 to 0.86 were observed.

Conclusion: We have shown that both the VAS version of WOMAC have adequate evaluative and discriminative properties. We found the pain and physical function scales in VAS format and the stiffness scale in CT format to have a slightly better performance in this sample.