Understanding the current landscape of heath disparities research among youths with juvenile idiopathic arthritis

The landscape of health disparities research in juvenile idiopathic arthritis (JIA), and in pediatric rheumatology overall, has grown exponentially over the past decade following the COVID-19 pandemic and the national spotlight on social justice movements. Much of the earlier research highlights racial and ethnic disparities in JIA prevalence, which tend to be racially patterned in the United States. JIA subtypes characterized by greater morbidity (e.g., rheumatoid factor (RF) positive polyarticular JIA or systemic JIA) are more prevalent among non-white youth, including those identifying as American Indian and First Nations, Black or African American, or Hispanic.^1–4 Conversely, prevalence rates of oligoarticular JIA are higher among white or non-Hispanic individuals.⁴ Although overall rates of polyarticular JIA tend to be similar among white and Black youth, previous studies report greater JIA-related morbidity among Black youth, including increased joint damage, pain, and disease activity.^4,5 While the report of racial and ethnic distributions can be useful in better understanding patterns of disease burden, these statistics may lack generalizability when stemming from single centers and may instead reflect regional differences in access to care and clinical practices. Consequently, large, multicenter cohorts have the potential to add valuable insight to the growing health disparities literature in JIA.

Race-based approaches of describing health disparities highlight racial differences in which race serves as a primary determinant of illness. However, race is a social construct that reflects differential access to power and resources within society.⁶ In the United States, socioeconomic position (SEP) is highly racialized. In other words, SEP strongly correlates with race and ethnicity, making it difficult to disentangle the “effects” of race and ethnicity from the impacts of societal barriers and hardship that disproportionally burden marginalized communities.⁷ This patterning of the effects of SEP has also been observed among youth with JIA, such that Black youth with more severe disease are more likely to come from households with annual income less than $50,000 and rely on government-assisted health insurance.^4,5 Furthermore, reliance on Medicaid health insurance, compared to private health insurance, predicts presentation with polyarthritis, systemic JIA features, active disease, and pain among youths with JIA.⁸ Similarly, Canadian youths living on a reservation, rather than youths who identified as native North American race alone, are more likely to present with greater JIA-related disability. ⁹ These studies further highlight the role of SEP and regional and local factors on racial and ethnic disparities in JIA, and the need to understand their intersecting roles. The effort to emphasize the impact of racialized determinants on health, rather than race itself, represents a growing race-conscious approach to assessing health disparities.^10,11

In the November 2025 issue of The Journal of Rheumatology, Harris et al. report findings from a cross-sectional study examining disparities in outcomes of patients with JIA across 18 children’s hospitals participating in the Pediatric Rheumatology Care and Outcomes Improvement Network (PR-COIN), a North American learning health network that monitors quality measures among youths with JIA.^12,13 The study included data from 9,601 patients seen at least once between April 2011 and March 2024. The authors focus on racial and ethnic disparities in disease outcomes, including physician and patient/caregiver assessments, active joint count, the 10-joint clinical Juvenile Arthritis Disease Activity Score (cJADAS10), and arthritis-related pain.

Approximately 62% of patients were classified as white race, 7% as Hispanic, 4% as non-Hispanic Black, and 3% as other race and ethnicity, including Asian, American Indian or Alaska Native, Native Hawaiian or other Pacific Islander, or multi-racial. Of note, approximately 25% of patients had unknown race and/or ethnicity but were maintained in the analyses as “unknown” for data completeness. Patient sociodemographic characteristics varied across the five reported racial and ethnic groups, including age at most recent visit, sex assigned at birth, and insurance status. Non-Hispanic Black patients were generally older (mean age 13.1 years) and had the highest proportion of male patients (38%) among all racial and ethnic groups. Patients identified as Hispanic (23%) or non-Hispanic Black (21%) were more likely to have Medicare or Medicaid as their primary insurance type compared to non-Hispanic white patients (8%).Consistent with findings from prior studies, RF positive polyarticular JIA (13%) and systemic JIA (12%) were most prevalent among non-Hispanic Black patients, with similar rates of RF positive polyarticular JIA among Hispanic patients (13%).

The findings by Harris et al. largely support findings from previous studies that evaluated racial and ethnic disparities in JIA outcomes.^4,14 In univariate analyses, all JIA disease outcome measures varied significantly across racial and ethnic groups. In general, non-Hispanic Black and Hispanic patients reported more active and/or severe disease. After adjusting for age, sex assigned at birth, insurance type, disease duration, JIA subtype, and study site, non-Hispanic Black patients with JIA consistently presented with worse JIA outcomes across most measures compared to non-Hispanic white patients, including arthritis-related pain score (β=0.56; 95% CI: 0.22, 0.91), physician global assessment score for disease activity (β=0.4; 95%CI: 0.22, 0.59), patient/caregiver score (β=0.68; 95%CI: 0.35, 1.02), and cJADAS10 score (β=1.4; 95%CI: 0.76, 2.04). While the authors interpret their results pertaining to age at visit, sex, and JIA subtype in relation to JIA outcome measures, we do not re-highlight them here to avoid potential misinterpretation of the estimates. In multivariable regression analyses, interpretation of estimates for individual covariates is not recommended since the prescribed models were not developed to evaluate specific covariate-outcome associations (e.g., models may be missing confounders specific to that relationship), and can result in misleading interpretations—commonly referred to as the “Table 2 Fallacy”.¹⁵

Large, multicenter patient registries like PR-COIN provide valuable insight into the epidemiology and disease course of JIA. However, when viewed through a health equity lens, it becomes critical to acknowledge and understand the inherent limitations of these data and how these limitations may impact the interpretation of results. The authors appropriately describe the identified limitations of this study, including the cross-sectional study design, the variation in the classification of race and ethnicity, the scope of missing data, potential cohort effects due to overlap with the COVID-19 pandemic, and barriers to generalizability. As the authors note, the cross-sectional study design restricts the ability to infer causal association between their demographic variables and JIA disease outcomes. While some sociodemographic variables—such as racial and ethnic self-identity—remain relatively consistent, others—such as insurance status or income—are more likely to change over time. Longitudinal studies can help capture unbiased temporal relationships between sociodemographic characteristics and disease outcomes over time. For example, a previous longitudinal study by Chang et al. reported persistent racial disparities in disease outcomes among Black versus white children with JIA following the use of clinical decision support at a single medical center, despite similar rates of improvement in JIA outcome measures across racial groups over time.⁵

The potential for misclassification of racial and ethnic groups and the high prevalence of missing race and ethnicity data represent important limitations of the study. The authors note the high variability in the methods used across the study sites to collect race and ethnicity data a, which included self-report, EMR abstraction, and assignment by clinical staff. Race and ethnicity obtained from EMRs or secondary sources must be carefully considered since there is large opportunity for misclassification, especially when race is assigned by an outside party (e.g., clinic staff) rather than self-reported by the patient/caregiver. Studies have reported the rate of discordance between EMR and parental-reported race ranging from 5% to 35% with higher rates observed for Hispanic individuals or those identifying as a combination of races and ethnicities.^16–18 In addition, higher rates of discordance have been observed among pediatric patients compared to adults, which has been attributed to the additional complexity of caregiver report versus child report.¹⁸ To address this issue, medical centers have started to emphasize the importance of implementing programs that prioritize self-reported demographics, including race and ethnicity and gender, in efforts to support inclusive and equitable practices.¹⁹

Approximately 5,000 (or one-third) of patients in the PR-COIN Registry were excluded from the analyses due to missing demographic or outcome variables or were not seen in clinic during the study period (either initial visit or follow-up). The distribution of registry patients with exclusionary missing data was not reported. Incomplete assessment of missing data in studies that utilize registry or electronic medical record (EMR) data can bias estimates and may ultimately widen health disparities through the reporting of inaccurate findings. Previous studies have demonstrated that data with socioeconomically-patterned missingness have a high potential for misclassification and misinterpretation.^20,21 This is particularly important when assessing racial and ethnic disparities, as previous studies have reported that missing data in EMR and surveillance databases occur disproportionately among Black and Hispanic/Latino patients.^21,22 Similarly, a previous study that utilized data from six PR-COIN centers reported that missing race data was correlated with missing cJADAS10 scores among registry patients with JIA.²³ After multiple rounds of audit and feedback cycles, they were able to decrease their missingness by 94% and found that recovered data were more likely to represent patients with Other race or Hispanic/Latino ethnicity. They stressed the need for data completeness and accurate assessment of race and ethnicity variables in disparities research in JIA.

While the impact of excluding participants due to missing demographics and disease outcomes (if socioeconomically patterned) may be more intuitive given the reported findings, exclusion due to visits outside the study period may be less obvious. For example, due to the long study period (13 years), the likelihood of initial visits being captured is not likely to be systematically biased. However, disparities in access to care among under resourced groups (e.g., barriers to transportation, time off from work, and decreased health literacy, etc.) could impact the rate of follow-up visits and the probability of being included at different stages of the disease course. Therefore, assessing patterns of missing data and critically interrogating plausible reasons for data missingness can be beneficial in understanding the impact of missing data on reported findings and the generalizability of results to other populations.

The identification of health disparities in JIA is still ongoing with the use of large clinical registries of patients with JIA, such as the PR-COIN Registry and the CARRA Registry.¹⁴ In the discussion, Harris et al. briefly describe race as a social construct and recommend that additional research should consider variables such as the child opportunity index and primary language when assessing JIA disparities.¹² This call aligns with a more social-ecological approach to understanding health inequities in JIA that critically examines interconnected factors related to the individual, familial, organizational, and societal environments.²⁴ To truly mitigate health disparities in JIA, clinicians, researchers, and other stakeholders will need to leverage findings from existing and novel disparities research to inform the development of meaningful interventions that address disparities at the healthcare level, as well as advocate for policy change at the local and national level. These efforts will fundamentally depend on the quality, completeness, and representativeness of collected data on disease determinants and the rigor of the health disparities research that utilize those data.

Ultimately, when documenting racial and ethnic disparities in JIA and pediatric rheumatology, it is important to consider the social context around race and ethnicity that drive these disparities. Race and ethnicity do not influence health independently but instead represent a complicated milieu of social and structural factors that are embedded in societal, environmental, political, and healthcare landscapes that disproportionally burden marginalized populations. Because “race” as a phenotypic exposure cannot be modified, taking a race-conscious approach that examines how race and ethnicity interact with socioeconomic factors as well as developing interventions that can address those intersections will be critical to reducing and ultimately eliminating observed inequities.