Ascertainment of arthritis can be challenging in large-scale studies, and self-reported case definitions are often used to facilitate the evaluation of disease burden. Case definitions used in most studies involve participant reports of any prior receipt of an arthritis or rheumatic disease diagnosis from a health professional; self-reported chronic arthritis symptoms have also been used, rarely as a stand-alone criterion and typically together with a self-reported doctor diagnosis. There is evidence that self-reported case definitions for arthritis have satisfactory sensitivity for population surveillance purposes, but suffer from low specificity, particularly among individuals less than 65 years old (Bombard, Powell, Martin, Helmick, & Wilson, 2005). Nevertheless, there is no evidence that instrument validity for self-reported arthritis definitions varies by SES, as could systematically bias estimates of the SES and self-reported arthritis association. To the extent that most self-reported arthritis involves OA of the hips, knees, or hands, the evidence for its differential prevalence by SES complements that presented in our section on radiographic OA.

We summarize some of the work on socioeconomic differences in the risk of self-reported arthritis in Table 6.2. These studies span decades, geographic regions, and demographic groups, yet almost unanimously find meaningful differences in the risk of self-reported arthritis by a number of socioeconomic characteristics, including education, income, neighborhood disadvantage, or occupational class. There is equivocal evidence for effect-measure modification by race/ethnicity. Cunningham and colleagues notably found gradient patterns of increased self-reported arthritis susceptibility with lower SES, but only among non-indigenous participants in a sample of Australian participants (Cunningham, 2011); in comparison, data from a cohort study in rural North Carolina (USA) did not suggest significantly different patterns of inequities between Whites and African Americans (Callahan et al., 2008).

With a prevalence between 0.5 and 1 %, rheumatoid arthritis (RA) is the most common autoimmune disorder; women represent two thirds of all cases (Gabriel, 2001) and the disease typically occurs between the age of 50 and 75. Current genetic heritability estimates for RA do not exceed 50 % (Frisell et al., 2013), leaving ample room for modifiable determinants which, with the exception of smoking (Silman, Newman, & Macgregor, 1996), remain largely unknown.

We provide an overview of studies on socioeconomic differences in RA risk in Table 6.3. While early studies found no clear associations of socioeconomic status with RA development (Bankhead et al., 1996; Uhlig et al., 1999), more recent work consistently suggests that people with low SES are at an increased susceptibility for RA. Intriguing data further indicate that socioeconomic disadvantage chiefly increases susceptibility to seropositive disease (Bengtsson et al., 2005; Pedersen et al., 2006), as characterized by rheumatoid factors, supporting the increasingly adopted notion that seropositive and seronegative RA constitute etiologically distinct phenotypes.