Is the Rheumatoid Arthritis Impact of Disease (RAID) score a meaningful instrument for other inflammatory rheumatic diseases? A cross-sectional analysis of data from the German National Database

Abstract

ObjectiveTo analyse the performance of the rheumatoid arthritis impact of disease (RAID) score in patients with ankylosing spondylitis, polymyalgia rheumatica, systemic lupus erythematosus, primary Sjögren's syndrome, idiopathic inflammatory myositis and systemic sclerosis, as compared with rheumatoid arthritis (RA).MethodsA total of 12 398 patients from the German National Database were included. For each diagnosis, we calculated age-adjusted/sex-adjusted partial correlation coefficients between RAID and patient global (PtGl) health, PtGl disease activity, physician global (PhGl) disease activity, Well-Being Index (WHO-5) and EuroQoL-5 Dimensions (EQ-5D). As a measure of agreement, the mean differences between the RAID and other outcomes were compared with the respective differences for RA. The effect of each diagnosis on the difference between RAID and the other scores was assessed with linear regression, with RA as the reference.ResultsAcross all diagnoses, RAID correlated strongly with PtGl health (0.71-0.83), moderately to strongly with PtGl disease activity (0.59-0.79), WHO-5 (0.65-0.81) and EQ-5D (0.68-0.73) and weakly with PhGl disease activity (0.23-0.38). Mean differences were calculated for RAID and PtGl disease activity (0 to -0.6), PtGl health (-0.4 to -0.9), WHO-5 (-0.7 to -1.3), EQ-5D (1.1 to 1.7) and PhGl disease activity (1.4 to 2.2). Discrepancies between other scores and RAID were comparable to RA. Linear regression revealed no clinically relevant effect of any of the diagnoses on the difference between RAID and the other outcomes.ConclusionThe RAID score performs comparably across all diagnoses investigated. This supports the use of RAID for measuring the impact also of other rheumatic diseases.

Keywords: arthritis, rheumatoid; outcome assessment, health care; patient reported outcome measures.

Figure 1

Violin plots displaying the distribution of the original weighted and the alternative unweighted rheumatoid arthritis impact of disease (RAID) composite score for different diagnoses. The violin plot depicts distributions of numeric data using density curves. The width of each density curve corresponds with the approximate frequency of data points in each region. The lower and upper dashed lines represent the first and third quartiles with the middle 50% of the data in between. The line in the middle shows the median. AS, ankylosing spondylitis; IIM, idiopathic inflammatory myositis; PMR, polymyalgia rheumatica; pSS, primary Sjögren’s syndrome; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SSC, systemic sclerosis.

Figure 2

Age-adjusted and sex-adjusted partial correlation coefficients between the raid score and other patient-reported and physician-reported outcomes. 0.3–0.5 weak, 0.5–0.7 moderate, >0.7 strong correlation. EQ-5D, EuroQoL-5 Dimensions; IIM, idiopathic inflammatory myositis; PhGl, physician global; PMR, polymyalgia rheumatica; pSS, primary Sjögren’s syndrome; PtGl, patient global; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SSC, systemic sclerosis; WHO-5, WHO Well-Being Index.

Figure 3

Differences between raid und five outcomes (PtGl health, PtGl disease activity, PhGl disease activity, WHO-5, EQ-5D). The difference was calculated as raid minus the respective outcome. The lower and upper dashed lines represent the first and third quartiles with the middle 50% of the data in between. The line in the middle is the median. AS, ankylosing spondylitis; EQ5-D, EuroQoL-5 Dimensions; IIM, idiopathic inflammatory myositis; PhGl, physician global; PMR, polymyalgia rheumatica; pSS, primary Sjögren’s syndrome; PtGl, patient global; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SSC, systemic sclerosis; WHO-5, WHO Well-Being Index.

Figure 4

Linear regression models. The dependent outcome variable is the difference between the RAID score and the five respective outcomes determined as: RAID minus PtGl health status (A), RAID minus PtGl disease activity (B), RAID minus PhGl disease activity (C), RAID minus EQ-5D (D), RAID minus WHO5 (E). Independent variables were sex, gender and diagnosis with RA as reference category. An effect of a diagnosis was defined as clinically relevant if the outcome deviated from the RAID more than one unit than it did for patients with RA (grey area). AS, ankylosing spondylitis; EQ5-D, EuroQoL-5 Dimensions; IIM, idiopathic inflammatory myositis; PhGl, physician global; PMR, polymyalgia rheumatica; pSS, primary Sjögren’s syndrome; PtGl, patient global; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SSC, systemic sclerosis; WHO-5, WHO Well-Being Index.