FormalPara Key Summary Points

Why carry out this study?

Literature suggests healthcare delivery disparities (i.e., lower use of specialist, delays in seeking initial care resulting in greater disease severity/poorer quality of life) in patients with psoriasis.

There are limited data available quantifying inequities in psoriatic arthritis (PsA), and when reported they are aggregated with psoriasis. Because PsA affects up to 30% of psoriasis patients, studies specifically evaluating care delivery disparities in PsA are warranted.

What was learned from the study?

Care delivery disparities were observed in patients with PsA in the United States Medicaid population.

Race/ethnicity, age, Medicaid eligibility category, health insurance plan type, prior diagnosis of psoriasis, prior use of nonsteroidal anti-inflammatory drugs and corticosteroids, and comorbid conditions including anxiety, chronic kidney disease, and hepatitis C were significantly associated with time to initiation of any disease-modifying antirheumatic drug (DMARD).

After adjusting for baseline characteristics, Black patients with PsA experienced delays in the initiation of any DMARDs, specifically biologic DMARDs.

Introduction

Psoriatic arthritis (PsA) is a chronic, autoimmune form of arthritis with a substantial economic burden and a high impact on healthcare resource utilization [1, 2]. Approximately 30% of patients with psoriasis have both joint and skin involvement and may develop PsA [3]. PsA usually develops between 30 and 50 years of age and the prevalence rates are similar between men and women. In the United States (US), the prevalence of PsA is estimated as 0.06–1% [4, 5]. The disease is highly heterogeneous in terms of clinical presentation involving multiple domains including peripheral arthritis, psoriasis, axial disease, enthesitis, dactylitis, and psoriatic nail disease [6, 7], with greater severity in Black patients compared with White patients [8]. Studies have reported diagnostic delays of up to 5 years [9,10,11] in patients with PsA. Delayed treatment of PsA can result in irreversible joint damage and reduced functional capacity [12].

The treatment landscape of PsA includes nonsteroidal anti-inflammatory drugs (NSAIDs) and local glucocorticoid injections for symptomatic treatment of mild PsA to control pain and inflammation. Disease-modifying antirheumatic drugs (DMARDs) that may slow down disease progression include biologics (bDMARDs), conventional synthetic drugs (csDMARDs; methotrexate, leflunomide, and sulfasalazine), and newer oral targeted synthetic drugs (tsDMARDs; apremilast, upadacitinib, and tofacitinib). The bDMARDs are highly specific, targeted therapies. Based on their specific targeted pathways, bDMARDs are further classified as tumor necrosis factor inhibitors (TNFi; adalimumab, certolizumab pegol, etanercept, golimumab, and infliximab), interleukin 12/23 inhibitors (IL-12/23i; ustekinumab), interleukin 17 inhibitors (IL-17i; secukinumab and ixekizumab), interleukin 23 inhibitors (IL-23i; guselkumab, and rizankizumab), or selective T-cell costimulation modulators (abatacept).

Access to DMARDs for the treatment of immunologic conditions has been shown to vary by demographics, potentially contributing to disparities in outcomes and disease severity in different racial/ethnic populations [13, 14]. Literature suggests that patients with rheumatoid arthritis [15, 16] and psoriasis [17] have care delivery disparities, resulting in greater disease severity and poorer quality of life [18, 19]. An increase in healthcare costs has been reported with increasing functional disability in patients with PsA [2]. Data quantifying disparities in PsA care are an unmet need and contribute to challenges in the appropriate management of disease in historically marginalized populations [13].

The objective of this study was to assess racial/ethnic disparities in the treatment of PsA in the US, using Medicaid insurance claims data. We aimed to describe patient characteristics and use of DMARDs in patients with newly diagnosed PsA, identify patient characteristics that are associated with time to DMARD treatment initiation following initial PsA diagnosis, and report time to DMARD treatment initiation in patients according to their race/ethnicity as specified in the database.

Methods

Study Design and Population

This retrospective, non-interventional, descriptive cohort study used claims data from the IBM® MarketScan® Medicaid Multi-State Database from January 1, 2009 to December 31, 2019 (Fig. 1).

Fig. 1
figure 1

Study design. PsA psoriatic arthritis

Full size image

The Medicaid database contains medical and pharmacy claims data including information on hospital discharge and outpatient diagnoses and standard demographic variables, as well as race and ethnicity of approximately 10 million enrollees from 8 to 12 de-identified state Medicaid programs with detailed member enrollment information [20].

Patients with ≥ 2 medical claims with diagnosis codes for PsA (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] code 696.0 or International Classification of Diseases, Tenth Revision, Clinical Modification [ICD-10-CM] code L40.5x) ≥ 30 and ≤ 365 days apart during the patient identification period from January 1, 2010 to December 31, 2019 were included in this analysis. The date of the first claim with a PsA diagnosis was defined as the index date. Patients were required to have ≥ 12 months of continuous enrollment in medical and pharmacy plans prior to the index date, and ≥ 18 years of age on the index date. Patients with any claim with a PsA diagnosis code prior to the index date or with missing information on race/ethnicity were excluded. A subset of patients meeting an additional requirement, to have ≥ 12 months continuous enrollment in medical and pharmacy plans after the index date, were identified to conduct analyses on the use of DMARDs within 12 months following the initial PsA diagnosis. All database records were de-identified and compliant with US patient confidentiality requirements, including the Health Insurance Portability and Accountability Act of 1996. Because of this, institutional review board approval was not required. The datasets generated during and/or analyzed during the current study are not publicly available due to third-party license of IBM MarketScan® Multi-State Medicaid Database (2006–2021). Authors were granted permission to use the datasets for the current study.

Variables and Outcomes

Patient baseline demographics, clinical characteristics, and medication history, including the use of DMARDs, were recorded. Race/ethnicity in the database was categorized as White, Black, Hispanic, or Other. The primary outcome was the time to DMARD initiation (any DMARD, as well as csDMARDs, bDMARDs, and tsDMARDs, specifically) after PsA diagnosis. Secondary outcomes included utilization of and the type of DMARD within 12 months following the index date. The association between patient baseline characteristics and the time to initiation of DMARD therapy was evaluated.

Statistical Analysis

Descriptive analyses (means, standard deviations [SD], and percentages) were conducted for baseline patient demographic and clinical characteristics, including the use of DMARD treatment before PsA diagnosis and duration of continuous enrollment after PsA diagnosis. Data were reported for the overall group of included patients and for patients stratified by race/ethnicity. Differences in patient characteristics by race/ethnicity were evaluated using chi-square test (for categorical variables) and analysis of variance (ANOVA, for continuous variables).

Use of DMARD therapy, overall and by type of therapy, within 12 months after initial PsA diagnosis was measured in two ways. First, treatment initiation after the initial PsA diagnosis was described as a dichotomous variable as having at least one drug claim for the therapy of interest within 12 months. Chi-square tests were used to evaluate whether the percentage of patients who initiated DMARD therapy was similar across race/ethnicity categories. Logistic regression, adjusted for treatment within 12 months as a dependent variable, and race/ethnicity, patient demographics, socioeconomic status, and clinical characteristics of interest as independent variables, were then used to assess the association between race/ethnicity and the initiation of a class of therapy within 12 months after the initial PsA diagnosis.

Second, the primary endpoint, time to DMARD initiation following PsA diagnosis, was described using the Kaplan–Meier (KM) survival analysis. The association between baseline variables and treatment initiation was evaluated using multivariate Cox proportional hazards models and reported as the adjusted hazard ratio (HR) with 95% confidence interval (CI). All data analyses were performed using Statistical Analysis Software, version 7.1.

Results

Baseline Characteristics of the Overall Population

Overall, 3432 patients met the identification criteria (Fig. 2) and were included in the descriptive summary of patient characteristics and in the analysis of the primary endpoint—time to DMARD initiation after PsA diagnosis.

Fig. 2
figure 2

Patient attrition. PsA psoriatic arthritis. aThe date of the first medical claim with a diagnosis of PsA was defined as the index date

Full size image

Most patients with newly diagnosed PsA were female (69.9%); the mean (SD) age was 44.4 (12.1) years and the mean (SD) Quan-Charlson Comorbidity Index (QCI) score was 0.8 (1.4). The majority of recorded race/ethnicity was White (n = 2656, 77.4%), followed by Black (n = 347, 10.1%). We focused on the subgroups of White and Black patients because a relatively small group of Hispanic patients was identified (n = 87, 2.5%), and conclusions cannot be generalized for the group of patients classified as ‘Other’ race/ethnicity in the database (n = 342, 10.0%). Of all patients with newly diagnosed PsA, 56.2% had a diagnosis of psoriasis during the 12-month baseline period. The majority (68.3%) did not receive any DMARD treatment during the 12-month baseline period; use of csDMARDs, bDMARDs, and tsDMARDs was reported in 20.6, 15.2, and 1.6% of patients, respectively. Hypertension (45.5%), depression (35.9%), and anxiety (33.8%) were the most common comorbid conditions at baseline (Table 1).

Table 1 Baseline patient demographic and clinical characteristics
Full size table

After stratification by race/ethnicity, the White and Black patient groups were numerically similar in age. The proportion of female patients was slightly higher in the Black patient group (76.1%) than in the White patient group (70.0%), while the proportion of patients with a prior diagnosis of psoriasis was similar in the White (56.6%) and Black (55.3%) patient groups. Prior use of any DMARD and csDMARDs was numerically higher in the Black patient group (36.3% and 24.2%) than in the White patient group (31.2 and 19.9%); prior use of bDMARDs and tsDMARDs was similar between these two patient groups.

Treatment With DMARDs During the 12 months Following Initial PsA Diagnosis

Of the 3432 patients included in the overall study population, 98.1% of patients had at least 3 months of continuous enrollment, 94.7% had at least 6 months, 90.9% had at least 9 months and 87.2% (2993/3432) had at least 12 months of continuous enrollment after PsA diagnosis (Table S1 Supplementary Material). The 2993 patients with at least 12 months of continuous enrollment after PsA diagnosis were included in the analyses evaluating the proportion of patients, by race, who received DMARD therapy within 12 months after diagnosis. Of these patients, fewer Black patients compared with White patients received csDMARDs (42.0 vs. 45.2%, p = 0.289) or bDMARDs (33.6 vs. 42.6%, p = 0.003) after initial PsA diagnosis (Fig. 3). Based on a comparison of baseline characteristics for patients with at least 12 months continuous enrollment and those who do not have at least 12 months of continuous enrollment, there were no clinically meaningful differences between the groups for any of the baseline characteristics.

Fig. 3
figure 3

DMARD use within 12 months after initial PsA diagnosis. *Out of 3442 patients included in the study, 2993 (87.2%) patients had ≥ 12 months of continuous enrollment in the database following the index date. ǂAll p values shown were calculated for the group of White patients versus the other respective patient groups (Black, Hispanic, Other). b biologic, cs conventional synthetic, DMARD disease-modifying antirheumatic drug, ts targeted synthetic, PsA psoriatic arthritis

Full size image

Time From Initial PsA Diagnosis to DMARD initiation, After Adjusting for Differences in Baseline Patient Characteristics

Black race, older age, anxiety, and hepatitis C were significantly associated with a longer time to start any DMARD. White, Hispanic, and Other race/ethnicity, point of service (POS) with capitation health insurance plan type, prior diagnosis of psoriasis, comorbid chronic kidney disease, and prior use of NSAIDs and corticosteroids were significantly associated with shorter time to any DMARD initiation. Medicaid basis for eligibility was also controlled for in the model (Table 1).

Based on KM analysis, the median time to initiation of any DMARD was 1.68 months in the White patient group and 2.14 months in the Black patient group. The median time to initiation of bDMARD, csDMARD, and tsDMARD therapies was > 12 months in both patient groups. In multiple regression analysis after adjusting for baseline characteristics, Black patients had a significantly longer time to initiation of any DMARD (HR [95% CI] 0.82 [0.71–0.94], Fig. 4A) and bDMARDs (0.84 [0.71–0.99], Fig. 4C) compared with White patients. In addition, while the time to initiation of csDMARDs (0.85 [0.72–1.0], Fig. 4B) and tsDMARDs (0.85 [0.61–1.19], Fig. 4D) was not statistically significantly different, the difference appears to be growing over time for csDMARD, as seen by the separation of the KM curves (Fig. 4B).

Fig. 4
figure 4

Time to DMARD initiation after initial PsA diagnosis, overall and by type of DMARD: A Any DMARD, B csDMARD, C bDMARD, D tsDMARD. Hazard ratios and 95% CIs are from Cox proportional hazard model adjusted for baseline variables. b biologic, cs conventional synthetic, DMARD disease-modifying antirheumatic drug, ts targeted synthetic, PsA psoriatic arthritis, CI confidence interval, HR hazard ratio

Full size image

Discussion

In this large, observational Medicaid claims-based study, racial differences were observed in the use of DMARDs, with significantly longer time to treatment initiation of any DMARD, as well as bDMARDs, in Black patients compared with White patients with newly diagnosed PsA. PsA presents as a wide spectrum of clinical features and patients are often diagnosed with other chronic autoimmune disorders [21]. The clinical heterogeneity of the disease, presence of comorbidities, and lack of uniform diagnostic criteria make diagnosis and management of PsA challenging and may impact the choice and timing of treatment for individual patients [22, 23].

Patients with psoriasis often also get diagnosed with PsA. Up to 30% of psoriasis patients initially present with a skin condition and then eventually progress into joint pain over 10 years following the initial psoriasis diagnosis [3]. In patients with psoriasis, disparities in healthcare delivery and barriers contributing to these inequities vary by patient’s race, level of education, and socioeconomic status [17, 24,25,26,27]. Despite the potentially more severe disease, Black, Asian, and other non-Latinx/Hispanic, historically marginalized populations are approximately 40% less likely to see a dermatologist for psoriasis compared with White patients [28]. Moreover, the clinical presentation of psoriasis is different in people with darker skin tones compared to those with lighter skin tones [29] and contributes to delayed diagnosis in historically marginalized populations [25, 26]. In the US, inequities in psoriasis care delivery also exist with Black/African American patients having a greater level of unfamiliarity with available treatment options, particularly biologics. Considering patients with psoriasis are at an increased risk of developing PsA, these healthcare disparities likely exist in patients with PsA. An analysis of IBM® MarketScan® Commercial Database reported a significantly higher economic burden of PsA (annual all-cause healthcare cost $29,742) compared with psoriasis ($11,062) [30]. There is an unmet need to better understand the burden of PsA among different racial/ethnic groups [18].

Studies in the US assessing the prevalence of PsA in the general population and historically marginalized populations are scarce and suggest that PsA is more prevalent in White patients than in Black, Hispanic, and Asian patients [31]. The disparity in prevalence could potentially be due to underdiagnosis in these historically marginalized racial/ethnic groups [8]. A single-center study reported that PsA is more frequent in White patients compared with Black patients (64.5 vs. 30%).

A cross-sectional US claims database analysis assessed the prevalence of PsA, bDMARD/tsDMARD utilization, and visits to rheumatologists among commercial insurance, Medicare, and Medicaid enrollees [32]. The study reported differences in PsA prevalence by race, with the highest prevalence among White patients (69.5%) compared to patients of other races/ethnicities irrespective of insurance type [32]. The study also reported differences in prevalence by insurance coverage; prevalence was higher among Medicare enrollees (0.32%) compared with Medicaid enrollees (0.17%) and Black and Latinx/Hispanic patients were less likely to have Medicaid coverage. The use of DMARDs was less in Medicare enrollees (21%) compared with Medicaid enrollees (37%) [32]. An average of 55% of Medicare enrollees compared with 12% of Medicaid enrollees saw a rheumatologist for their PsA. There was no difference seen among race/ethnicity groups.

In our study, among the 3432 Medicaid beneficiaries with PsA, the factors associated with time to treatment initiation included race/ethnicity, age, Medicaid eligibility category, health insurance plan type, prior diagnosis of psoriasis, comorbidity burden, and use of NSAIDs/COX-2 inhibitors and corticosteroids during the baseline period. The time to DMARD initiation was longer in Black patients compared with White patients. In our study, younger patients had a shorter time to DMARD initiation as compared with elderly patients. Also, patients with POS with capitation health insurance plan type had a shorter time to DMARD initiation compared with patients with other insurance coverage types. Patients with anxiety and hepatitis C comorbidities reported a longer time to DMARD initiation whereas those with a prior diagnosis of psoriasis or chronic kidney disease had a shorter time to DMARD initiation. After adjusting for these baseline variables, differences in the time to treatment by race/ethnicity remained significant. Black patients were less likely to be initiated on any DMARDs, including bDMARDs than White patients, and this delay in treatment initiation may adversely impact the long-term treatment outcomes.

The findings from our study are corroborated by other studies in the US reporting disparities in healthcare delivery in PsA [8, 32]. One study reported a significantly higher degree of disease severity and lower use of biologics among Black patients compared with White patients [8]. Data from the 2018, 2019, and 2020 US National Health and Wellness Survey including 1544 PsA patients (69.5% White, 7.4% Black, 14.4% Hispanic) also highlighted significant differences in sociodemographic characteristics and disease severity between racial/ethnic groups. Additionally, Black and Hispanic patients had lower patient activation measure (PAM) scores than White patients, suggesting lower levels of engagement in managing their health [33]. These results suggest potential racial/ethnic disparities in care among patients with PsA.

Our study analyzed US Medicaid claims data from select de-identified states. Among the overall US Medicaid population, approximately 57% of adult Medicaid beneficiaries identify as racial and ethnic minorities. The 2015–2018 National Health Interview Survey found that differences in sociodemographic characteristics exist by race and ethnicity among Medicaid beneficiaries [34]. For instance, the survey data indicated that Black adults were significantly more likely to have a household income below 100% of the federal poverty level and less likely to be part of the labor force compared to White adults. Disparities in access to care may be influenced by variations in costs and covered services for adult Medicaid beneficiaries, taking into account factors such as state, population, and income [34]. Notably, socioeconomic inequalities resulting in disparities in healthcare among ethnic minorities has been reported in many European countries [34].

The findings of the present study should be interpreted from the perspective of certain limitations. First, missing information on variables that impact treatment decisions, e.g., disease severity measures, is a limitation of using claims data. However, we may assume that individuals enrolled in Medicaid are likely to have similar socioeconomic characteristics, such as low-income status. Second, as noted above, the dataset does not identify individual states and we cannot examine whether the trends identified in this study are the result of geography and state-level characteristics regarding Medicaid policies. Third, the results are not generalizable to populations using non-Medicaid or other insurance claims data as the dataset may not be a nationally representative sample. Fourth, other than White, Black, and Hispanic, all other race/ethnicity groups, which includes Asian patients, were put together in the “Others” category further limiting generalizability. Finally, given the small sample sizes of some of the subgroups of interest, such as Black patients with tsDMARD exposure, the results may not be generalizable to a larger population and the power is reduced to detect a potential difference.

Conclusions

Our Medicaid claims-based study in patients with PsA found racial differences in the use of DMARDs, with significantly longer time to treatment initiation of any DMARD, as well as bDMARDs, in Black patients compared with White patients. These findings further support the need for future studies to understand the possible reasons for racial/ethnic differences and ways to eliminate these barriers to healthcare equity.