Background Inflammatory pain is considered one of the most limiting symptoms in patients with rheumatoid arthritis (RA). Accordingly, there is a variety of patient reported outcomes available to evaluate different facets of this complex phenomenon. However, to date, the most common single items measuring overall pain (i.e. the visual analog scale, the numerical rating scale and the verbal rating scale) have not been validated against each other yet.
Objectives To assess the construct validity, retest-reliability and sensitivity to change of the visual analog scale (VAS), the numerical rating scale (NRS) and the verbal rating scale (VRS) measuring overall pain in RA-patients.
Methods 62 outpatients with definite diagnosis of RA according to the 1987 ACR classification criteria1 were asked to voluntarily complete a questionnaire containing the VAS, a 0-10 NRS, and five-category VRS evaluating overall pain during the last week. Each patient had to complete the corresponding questionnaire three times: Before and immediately after medical consultation (time points 1 & 2) as well as one week after his visit (time point 3). The pain rating scales were spread across each questionnaire which also included the HAQ and three items of the BASDAI for distracting the patient’s attention from the last rating on one of the pain scales. Additionally, the patient’s current DAS28 score and information on gender, age, duration of disease, and a possible change of medication were also taken into account. Both, construct validity and retest-reliability were evaluated by partial correlation analysis controlling for the influence of demographic variables (construct validity: correlation across pain scales at each time point, retest reliability: correlation across time points for each pain scale in patients with stable medication). Sensitivity to change of each pain scale was assessed by calculating the standardized response mean (SRM) between time point 2 and time point 3 in a subgroup of patients that had undergone a change in antirheumatic treatment.
Results Partial correlation coefficients of overall pain VAS, NRS and VRS were reaching from r =0.79 to 0.89 indicating a valid measurement of the same construct (i.e. overall pain) at each of the three time points. Retest reliability in the stable medication subgroup (N=54) again revealed satisfactory results with correlation coefficients being consistently r ≥0.68 (range r: 0.68-0.97) and a common proportion of variance of at least 46% at all time points. SRM as an indicator of sensitivity to change was indicating small effects for VAS (SRM=0.46) and NRS (SRM=0.42), respectively, and a medium effect in view of the VRS (SRM=0.79).
Conclusions According to our results, satisfactory construct validity and retest reliability can be assumed for VAS, NRS, and VRS measuring overall pain in RA. To date, results on sensitivity to change should be interpreted cautiously as the corresponding subgroup was rather small but will be enlarged in the future.
Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 1988 Mar;31(3):315-24.
Disclosure of Interest None Declared