group vs. discordant group, active GS synovitis (≥ 2) was detected in 90% and 78% (P = .42) and active PD synovitis (≥) was detected in 10% and 33% (P = .17). Comparison of ultrasonographic assessments of disease activity according to the Ultrasound 7 (US7) scoring method between concordant and discordant groups with rheumatoid arthritis. US7 scores of gray scale (GS) and power Doppler (PD) synovitis, GS and PD tenosynovitis, and erosions for concordant and discordant groups, stratified by LDA (upper panel) vs MHDA (lower panel). Error bars indicate standard deviations To evaluate the potential for participation bias, the 70 participants in the prospective substudy were compared with 70 patients who declined to participate (Table 6). No significant differences were seen between participants and nonparticipants in discordance frequency, demographic characteristics, highest education level, provider type, or comorbidities. Participants were less likely than nonparticipants to have positive RF (53% vs. 76%, P = .004), anti-CCP antibodies (53% vs. 73%, P = .02), elevated CRP level (19% vs. 36%, P = .04), and radiographic joint erosions (35% vs. 54%, P = .03). However, no significant differences were found in pain, HAQ-II disability, clinical disease activity measures, or treatments between participants and nonparticipants. Comparison of clinical characteristics between participants and patients who declined participation in the prospective substudy This study is among the first to perform a comprehensive clinical evaluation of the myriad potential correlates of patient-provider discordance—including several domains not previously assessed—in a real-world RA population. Overall, the prevalence of patient-provider discordance in this study was slightly less, at 33%, than the pooled estimate of 43% reported in a recent systematic review and meta-analysis [1]. The prevalence of discordance in which patients rate their disease as more severe than their providers (i.e., positive discordance) was 29%, which is nearly identical to studies by Barton et al. [3] and Khan et al. [2] at 29% and 30% of clinical encounters. In particular, fibromyalgia, depression, and nonerosive disease were independently associated with patient-provider discordance. Barton et al. [3] have previously shown that positive discordance affects categorization of disease activity by composite measures, such that removal of the patient global assessment and calculation of the three-variable DAS28 led to shifting of patients from MHDA to LDA categories. The present study demonstrates that patients with positive discordance are less likely to be in remission and more likely to be in the LDA or MDA categories. Together, the findings underscore the difficulty in interpreting composite disease activity scores in the clinical setting of patient-provider discordance, considering the absence of meaningful correlation between discordance and inflammatory measures, as well as the uncertainties with implementation of current treat-to-target recommendations [23, 24]. Pain intensity is a key correlate of patient-provider discordance [1,2,3, 7, 25], but few studies have addressed specific characteristics or comorbidities related to pain etiologies. Based on the results of the SF-MPQ-2 analyses, continuous and neuropathic pain types are associated with patient-provider discordance in RA. The findings of this study are in agreement with Koop et al. [26], who reported neuropathic