Arthritis care & research | 2019
Intersectional inequalities and individual heterogeneity in chronic rheumatic diseases: An intersectional multilevel analysis.
Abstract
OBJECTIVE\nTo examine how intersections of multiple sociodemographic variables explain the individual heterogeneity in risk of being diagnosed with any of following chronic rheumatic diseases (CRDs): osteoarthritis (OA), gout, rheumatoid arthritis (RA), or spondyloarthritis (SpA).\n\n\nMETHODS\nWe identified people aged 40-65 years residing in Skåne, Sweden, by 31st December 2013 and having done so from 1st January 2000 (N=342,542). We used Skåne healthcare register to identify those with a diagnosis of the CRD of interest between 1st January 2014 and 31st December 2015 with no previous such diagnosis during 2000-2013. We created 144 intersectional social strata (ISS) using categories of age, gender, education, income, civil status, and immigration. With individuals nested within ISS, we applied multilevel logistic regression models to estimate: 1) variance partition coefficient (VPC) as a measure of discriminatory accuracy of the ISS and 2) predicted absolute risks and 95% credible intervals for each stratum.\n\n\nRESULTS\nIn overall, 3.5%, 0.5%, 0.2%, and 0.2% of the study population were diagnosed with OA, gout, RA, and SpA, respectively. The VPC ranged from 16.2% for gout to 0.5% for SpA. Gender explained the largest proportion of between-strata variation in risk of RA, gout, and SpA while age was the most important factor for OA. The most between-strata differences in risk of these CRDs were due to the additive main effects.\n\n\nCONCLUSION\nDespite meaningful between-strata inequalities in the risk of being diagnosed with CRDs (except SpA), there were substantial within-strata heterogeneities that remains unexplained. There were limited evidence of intersectional interaction effects.