Patient-Reported and Economic Racial and Ethnic Disparities in Patients with Psoriatic Arthritis: Results from the National Health and Wellness Survey

Abstract

Introduction

Psoriatic arthritis (PsA) is a chronic, autoimmune form of arthritis that is associated with a substantial humanistic and economic burden. Potential differences in patient-reported outcomes (PROs) and economic outcomes among groups of varying PsA severity and different races/ethnicities have not been well studied.

Methods

This cross-sectional study assessed sociodemographic data, PROs, and economic outcomes for participants with PsA from the National Health and Wellness Survey (2018–2020). Multivariable analyses were used to assess the association of self-reported PsA severity and race/ethnicity with health-related quality of life (HRQoL), work productivity and activity impairment (WPAI), healthcare resource utilization (HCRU), and medical costs.

Results

This study included 1544 participants with PsA (1073 non-Hispanic white, 114 non-Hispanic Black, 223 Hispanic, and 134 Other). Self-reported moderate/severe PsA was associated with significantly worse HRQoL and WPAI, greater HCRU, and higher costs than self-reported mild PsA. Black participants reported more absenteeism (31.11% vs. 16.69%; P = 0.007) and activity impairment (54.27% vs. 47.96%; P = 0.047) than white participants, and fewer healthcare provider (5.93 vs. 7.42; P = 0.039) and rheumatologist visits (0.29 vs. 0.53; P = 0.028) over the past 6 months. No differences in outcomes were observed between Hispanic and white participants. Race/ethnicity moderated the association of perceived PsA severity and PROs and HCRU, such that white participants with self-reported moderate/severe PsA had a higher likelihood of depression (P < 0.001), lower HRQoL (P < 0.001), and more emergency room visits (P = 0.001) than those with self-reported mild PsA. Race/ethnicity did not moderate the relationship of PROs, HCRU, and economic outcomes among Black or Hispanic participants.

Conclusion

Participants with self-reported moderate/severe PsA reported a greater burden than those with self-reported mild PsA. Black participants had a greater humanistic burden than white participants but reported lower HCRU. Moderation results were driven by white participants, suggesting important differences in PROs, HCRU, and perception of PsA severity across race/ethnicity groups. Small sample sizes in Hispanic and non-Hispanic racial/ethnic groups limited ability to discern differences related to disease severity in these groups. Further research is needed to better understand the differential burden of PsA among individuals with varying perceptions of PsA severity across different racial/ethnic groups.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40744-024-00717-7.

Key Summary Points

Why carry out this study?

Psoriatic arthritis (PsA) is a chronic, autoimmune form of arthritis that is associated with a substantial humanistic and economic burden.

Potential differences in patient-reported outcomes (PROs) and economic outcomes among groups of varying perceived PsA severity and different races/ethnicities have not been well studied.

This cross-sectional study used sociodemographic data, PROs, and economic outcomes for participants with PsA from the National Health and Wellness Survey (2018–2020) to assess the association of self-reported PsA severity and race/ethnicity with health-related quality of life (HRQoL), work productivity and activity impairment (WPAI), healthcare resource utilization (HCRU), and medical costs.

What was learned from this study?

Self-reported moderate/severe PsA was associated with significantly worse HRQoL and WPAI, greater HCRU, and higher costs than self-reported mild PsA.

Race/ethnicity moderated the association of perceived PsA severity and PROs and HCRU among white participants but not Black or Hispanic participants, such that white participants with self-reported moderate/severe PsA had a higher likelihood of depression, lower HRQoL, and more emergency room visits than those with self-reported mild PsA.

Introduction

Psoriatic arthritis (PsA) is a chronic inflammatory arthritis that is found in one in four individuals with psoriasis [1]. The clinical presentation of PsA can involve pain, swelling, stiffness, and limited range of motion of the joints [2]. Many individuals with PsA have reported skin symptoms such as itching, redness, flaking, and scales [3]. PsA has a heterogenous disease course and can range from mild to severe [4].

PsA has a substantial negative impact on physical function and health-related quality of life (HRQoL). Individuals with PsA experience fatigue and pain, which can impact the ability to work and participate in social activities, and contributes to impaired emotional well-being, low self-esteem, and financial insecurity [5, 6]. Individuals with PsA report high disease burden and unmet need [7, 8]. The extent to which an individual is burdened by PsA is a function of condition severity, and a person’s perception of the severity of their condition may depend on various factors, including the level of skin involvement and joint pain [9]. Previous surveys have demonstrated the discordance between patient and physician perceptions of PsA severity, indicating that clinical measures of disease severity used by physicians do not fully capture the impact of PsA on patients’ lives [10–13].

Multiple studies suggest PsA is more prevalent in non-Hispanic white individuals than other race/ethnicity groups [14, 15]. However, psoriasis on darker skin colors can be difficult to diagnose because of overlapping features with other skin disorders and difficulty in identifying areas of active inflammation, which typically manifest as pink or red lesions in lighter skin and violet or hyperpigmented lesions in darker skin types [14, 15]. This is further exacerbated by the lack of medical training in disease presentation in darker skin colors, with less than 5% of images in medical textbooks being of darker skin tones [16]. Together, this may cause underdiagnosis or delayed diagnosis among individuals with darker skin colors.

Delays in PsA diagnosis are problematic, potentially contributing to inequitable treatment and care delivery among racial and minority ethnic groups [17]. Indeed, time to treatment initiation is reportedly longer among Black patients than white patients newly diagnosed with PsA [17]. Further, delayed diagnosis or misdiagnosis of PsA is associated with significantly worse clinical outcomes like irreversible joint damage and reduced functional capacity [18]. Real-world studies show that PsA severity is strongly associated with worse HRQoL and greater work productivity impairment than mild PsA [19]. Therefore, timely diagnosis and treatment of PsA is critical [20].

A better understanding of patient experiences is needed to optimize care and reduce the overall burden associated with PsA. Evaluating this in the context of perceived condition severity and race/ethnicity using patient-reported outcomes (PROs) may also lead to more effective and inclusive recommendations for improving patient-centered outcomes in PsA. Therefore, the objective of this study was to evaluate the patient-reported humanistic and economic burden associated with PsA relative to participant-reported condition severity and race/ethnicity.

Methods

Study Design

Data were from the 2018, 2019, and 2020 US National Health and Wellness Survey (NHWS), a cross-sectional, internet-based survey of PROs collected from approximately 75,000 adults (≥ 18 years) annually [21]. Potential participants were recruited through an existing, general-purpose (i.e., not healthcare-specific) web-based consumer panel via opt-in e-mails, co-registration with panel partners, e-newsletter campaigns, banner placements, and affiliate networks. All NHWS participants completed an in-depth demographic registration profile and received an honorarium for completing the NHWS in the form of points and/or prizes redeemable through panel partners online reward system. A quota sampling technique was used to ensure that the demographic composition (i.e., age, gender, and race/ethnicity) of the NHWS sample was representative of the general adult population in the USA. The NHWS was reviewed by Pearl Institutional Review Board (Indianapolis, IN) and granted exemption status.

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Study Population

Participants included in the analytic sample met the following inclusion criteria at the time of the study: ≥ 18 years old, resident of the USA, self-reported ever receiving a physician diagnosis of PsA, and had complete data for outcomes relevant to the study (see Supplementary Material).

Sociodemographic and Clinical Characteristics

We evaluated sociodemographic (i.e., sex, age, marital status, education level, employment status, household income, health insurance) and clinical characteristics (i.e., PsA severity, body mass index [BMI], current prescription medication use, sleep problems/symptoms, Charlson Comorbidity Index [CCI] [22]). We also assessed various health behaviors, including alcohol use, cigarette smoking, exercise over the past 30 days, as well as previous participation in any clinical trial and participants’ willingness to self-monitor and regulate their health behaviors using the Patient Activation Measure (PAM) [23]. See Supplementary Material for more details on self-reported PsA severity, PAM, and clinical trial participation.

Independent Variables

Participants were categorized into one of four mutually exclusive race/ethnicity groups: non-Hispanic white, non-Hispanic Black/African American, Hispanic, or those who identified as other race (Asian; biracial, multiracial or of mixed ancestry; and other ancestry). Participants’ self-reported PsA severity was captured as the participants’ rating of their PsA as either mild, moderate, or severe. As a result of sample size limitations, moderate and severe PsA were grouped into one category.

Study Outcomes

Patient-reported humanistic burden was assessed through measures of mental health, HRQoL, labor force participation, and work productivity and activity impairment (WPAI).

• Depression and anxiety severity were assessed using the Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety Disorder Assessment (GAD-7), respectively [24, 25]. Higher PHQ-9 and GAD-7 scores indicate worse depressive/anxious symptomology.

• HRQoL was assessed using the physical component summary (PCS) and mental component summary (MCS) scores of the Medical Outcomes Study 36-Item Short Form Survey Instrument (SF-36v2), the Short-Form Six-Dimension (SF-6D) health index, and the 5-level EuroQol 5-dimension (EQ-5D) health questionnaire version (EQ-5D-5L), the last of which consists of the EQ-5D descriptive system and EQ visual analog scale (VAS) [26, 27]. Higher MCS, PCS, EQ-5D, SF-6D, and EQ VAS scores indicate better HRQoL.

• Labor force participation was derived from NHWS data through coding employment status as currently in the labor force (i.e., full-time employed, part-time employed, self-employed, or not unemployed but looking for work) or not currently in the labor force (i.e., retired, disabled, homemaker, student, or not employed and not looking for work).

• Work productivity was assessed using the WPAI questionnaire [28]. Only participants who reported a work status of full-time, part-time, or self-employed provided data for absenteeism, presenteeism, and overall work impairment.

Economic burden was assessed through healthcare resource utilization (HCRU) and direct and indirect costs.

• HCRU included the number of self-reported visits to any healthcare provider (HCP), rheumatologist, and emergency room (ER), and hospitalizations over the past 6 months.

• Direct medical costs were imputed using data from the region-specific Medical Expenditure Panel Survey (MEPS) and included costs of an average HCP visit, ER visit, and hospitalization [29].

• Indirect costs were those associated with work productivity impairment and were calculated using estimated wages/salaries for each participant with data from the US Bureau of Labor Statistics. This cost analysis approach has been used in prior research in the NHWS [22, 30].

Detailed descriptions of the study outcomes and how they were calculated are included in the Supplementary Material.

Data Analysis

Descriptive statistics were used to characterize the PsA cohort; results were reported as frequencies or percentages for categorical variables and means with standard deviations (SD) for continuous variables. Bivariate analyses were used to examine sociodemographic and clinical characteristics across self-reported PsA severity groups (i.e., mild and moderate/severe disease) and race/ethnicity groups (i.e., non-Hispanic white, non-Hispanic Black/African American, Hispanic, and other race). Chi-square tests were used to determine significant differences for categorical variables, whereas analysis of variance (ANOVA) and independent sample t tests were used for continuous variables. P values (α = 0.05) were provided for the omnibus test and pairwise testing between race/ethnicity subgroups. We applied the Bonferroni correction to P values from pairwise comparison to adjust for multiplicity.

To assess associations between self-reported PsA severity and outcomes of interest, as well as race/ethnicity and outcomes of interest, multivariable analyses were conducted by constructing generalized linear regression models (i.e., linear model for SF-36 and EQ-5D, binary logistic model for labor force participation, and negative binomial model for all other outcomes). Covariates were selected on the basis of bivariate analyses (i.e., variables that were associated with the key independent variables) and theoretical importance. We used the same covariate adjustment set in each regression model to enable comparisons of the magnitude of the association for each independent variable. Models controlled for age, gender, marital status, educational attainment, household income, health insurance coverage, weight status, smoking status, alcohol use, and CCI score. Adjusted means, mean differences (MD), odds ratios (OR), rate ratios (RR), and P values were reported.

Separate covariate-adjusted models were used to assess whether the relationships between self-reported PsA severity and outcomes were modified by race/ethnicity. The heterogeneity of the association was determined using an interaction term between race/ethnicity and self-reported PsA severity, which was added to each covariate-adjusted model. The statistical significance of the interaction term was assessed using type III analyses (the F test for linear models and likelihood ratio test (LRT) for logistic or negative binomial models). For models where the interaction term was statistically significant (α < 0.05), interactions were probed via stratified estimates. All analyses were conducted using SAS 9.4.

Results

Study Population and Sociodemographic Characteristics

A total of 1544 participants with self-reported, physician-diagnosed PsA were included in this analysis (Fig. 1). Most participants identified as non-Hispanic white (69.49%), followed by Hispanic (14.44%), other race (8.68%), and non-Hispanic Black (7.38%) (Table 1). Mean age was significantly lower among non-Hispanic Black (40.95 [SD 15.19]) and Hispanic (38.08 [SD 12.70]) participants with PsA than non-Hispanic white participants (52.85 [SD 15.13]; both P < 0.001). A lower proportion of non-Hispanic Black participants with PsA were married/living with a partner (33.33%) than non-Hispanic white (63.09%) and Hispanic (65.47%) participants (both P < 0.001). In general, educational attainment was high in the sample, with 55–64% of participants across all race/ethnic groups attaining at least an Associate’s degree. However, a greater proportion of Hispanic participants with PsA graduated college than non-Hispanic white participants (34.98% vs. 23.49%; P = 0.002).

Fig. 1.

Participant flow chart. NHWS National Health and Wellness Survey, PsA psoriatic arthritis

Table 1.

Sociodemographic and clinical characteristics in participants with self-reported psoriatic arthritis

	Non-Hispanic white (n = 1073)	Non-Hispanic Black (n = 114)	Hispanic (n = 223)	Other (n = 134)	Overall (N = 1544)
Female, n (%)	555 (51.72%)	56 (49.12%)	110 (49.33%)	70 (52.24%)	791 (51.23%)
Age in years, mean (SD)	52.85 (15.13)	40.95 (15.19)*	38.08 (12.70)†	43.73 (14.35)◊#	49.04 (15.86)
Marital status, n (%)
Married/living with a partner	677 (63.09%)	38 (33.33%)*	146 (65.47%)‡	65 (48.51%)◊#	926 (59.97%)
Single, not married/divorced/separated/widowed	396 (36.91%)	76 (66.67%)*	77 (34.53%)‡	69 (51.49%)◊#	618 (40.03%)
Education, n (%)
Less than high school	12 (1.12%)	1 (0.88%)	2 (0.90%)	0 (0.0%)	15 (0.97%)
Completed some high school	34 (3.17%)	3 (2.63%)	9 (4.04%)	6 (4.48%)	52 (3.37%)
High school graduate or equivalent (e.g., GED)	166 (15.47%)	23 (20.18%)	24 (10.76%)	6 (4.48%)פ	219 (14.18%)
Completed some college, but no degree	216 (20.13%)	24 (21.05%)	47 (21.08%)	36 (26.87%)	323 (20.92%)
Associate degree	140 (13.05%)	16 (14.04%)	21 (9.42%)	14 (10.45%)	191 (12.37%)
College graduate	252 (23.49%)	26 (22.81%)	78 (34.98%)†	33 (24.63%)	389 (25.19%)
Some graduate school	53 (4.94%)	5 (4.39%)	11 (4.93%)	8 (5.97%)	77 (4.99%)
Completed graduate school	200 (18.64%)	16 (14.04%)	30 (13.45%)	31 (23.13%)	277 (17.94%)
Decline to answer	0 (0.0%)	0 (0.0%)	1 (0.45%)	0 (0.0%)	1 (0.06%)
Employment status, n (%)
Employed full time	407 (37.93%)	48 (42.11%)	138 (61.88%)†,‡	63 (47.01%)#	656 (42.49%)
Self-employed	87 (8.11%)	12 (10.53%)	6 (2.69%)†,‡	12 (8.96%)	117 (7.58%)
Employed part time	83 (7.74%)	7 (6.14%)	18 (8.07%)	15 (11.19%)	123 (7.97%)
Homemaker	54 (5.03%)	7 (6.14%)	13 (5.83%)	5 (3.73%)	79 (5.12%)
Retired	291 (27.12%)	17 (14.91%)*	15 (6.73%)†	14 (10.45%)◊	337 (21.83%)
Student	10 (0.93%)	6 (5.26%)*	10 (4.48%)†	6 (4.48%)◊	32 (2.07%)
Disability	96 (8.95%)	12 (10.53%)	15 (6.73%)	16 (11.94%)	139 (9.00%)
Not employed (whether looking for work or not)	45 (4.19%)	5 (4.39%)	8 (3.59%)	3 (2.24%)	61 (3.95%)
Household income, n (%)
< $25,000	193 (17.99%)	27 (23.68%)	27 (12.11%)‡	25 (18.66%)	272 (17.62%)
$25,000 to < $50,000	219 (20.41%)	30 (26.32%)	47 (21.08%)	34 (25.37%)	330 (21.37%)
$50,000 to < $100,000	337 (31.41%)	27 (23.68%)	63 (28.25%)	37 (27.61%)	464 (30.05%)
$100,000+	296 (27.59%)	28 (24.56%)	83 (37.22%)†	35 (26.12%)	442 (28.63%)
Decline to answer	28 (2.61%)	2 (1.75%)	3 (1.35%)	3 (2.24%)	36 (2.33%)
Health insurance, n (%)
Commercially insured	520 (48.46%)	58 (50.88%)	131 (58.74%)	85 (63.43%)	794 (51.42%)
Medicaid	115 (10.72%)	14 (12.28%)	28 (12.56%)	16 (11.94%)	173 (11.20%)
Medicare	326 (30.38%)	20 (17.54%)*	26 (11.66%)†	22 (16.42%)◊	394 (25.52%)
Other type of insurance	32 (2.98%)	7 (6.14%)	14 (6.28%)	3 (2.24%)	56 (3.63%)
Not insured	80 (7.46%)	15 (13.16%)	24 (10.76%)	8 (5.97%)	127 (8.23%)
PAM scorea, mean (SD)	62.31 (12.10)	57.83 (11.99)*	59.83 (12.69)†	60.59 (11.67)	61.47 (12.21)
PAM Level, n (%)
Level 1—Disengaged & overwhelmed	80 (7.46%)	23 (20.18%)*	39 (17.49%)†	15 (11.19%)	157 (10.17%)
Level 2—Becoming aware, but still struggling	220 (20.50%)	22 (19.30%)	32 (14.35%)	21 (15.67%)	295 (19.11%)
Level 3—Taking action	544 (50.70%)	53 (46.49%)	119 (53.36%)	76 (56.72%)	792 (51.30%)
Level 4—Maintaining behavior & pushing further	229 (21.34%)	16 (14.04%)	33 (14.80%)	22 (16.42%)	300 (19.43%)
Severity of conditionb, n (%)
Mild	585 (54.52%)	52 (45.61%)	88 (39.46%)†	61 (45.52%)	786 (50.91%)
Moderate	408 (38.02%)	49 (42.98%)	106 (47.53%)	60 (44.78%)	623 (40.35%)
Severe	80 (7.46%)	13 (11.40%)	29 (13.00%)†	13 (9.70%)	135 (8.74%)
Ever participated in any clinical trialc (yes), n (%)	183 (17.05%)	32 (28.07%)*	75 (33.63%)†	28 (20.90%)	318 (20.60%)

GED General Educational Development, PAM Patient Activation Measure, PsA psoriatic arthritis, SD standard deviation

Sociodemographic characteristics were assessed using bivariate analyses

^*P < 0.05 between non-Hispanic Black and non-Hispanic white participants. ^† P < 0.05 between Hispanic and non-Hispanic white participants. ^‡ P < 0.05 between Hispanic and non-Hispanic Black participants. ^◊P < 0.05 between other race and non-Hispanic white participants. ^§ P < 0.05 between other race and non-Hispanic Black participants. ^# P < 0.05 between other race and Hispanic participants

^aPAM scores range from 0 to 100 where higher scores indicate higher levels of activation

^bAs a result of sample size limitations, moderate and severe PsA were grouped into one category to allow for the evaluation of disease severity as an effect modifier

^cThis pertains to participation in any clinical trial, not necessarily a PsA-related trial

A greater proportion of Hispanic participants were employed full time (61.88%) than non-Hispanic white (37.93%; P < 0.001) and non-Hispanic Black (42.11%; P = 0.004) participants (Table 1). Further, the proportion of participants who were retired was significantly higher among non-Hispanic white (27.12%) participants than among non-Hispanic Black (14.91%; P = 0.034) and Hispanic (6.73%; P < 0.001) participants. A greater proportion of non-Hispanic Black participants had a household income < $25,000 US dollars (USD) (23.68%) than Hispanic participants (12.11%; P = 0.042), and a greater proportion of Hispanic participants had a household income ≥ $100,000 USD (37.22%) than non-Hispanic white participants (27.59%; P = 0.025).

Patient activation scores were, on average, lower among non-Hispanic Black (57.83 [SD 11.99]) and Hispanic (59.83 [SD 12.69]) participants than non-Hispanic white participants (62.31 [SD 12.10]; P = 0.001 and P = 0.034, respectively; Table 1). Notably, the study sample was highly engaged, with 70.7% of participants reporting a PAM level of 3 (i.e., taking action and gaining control) or 4 (i.e., maintaining behaviors and pushing further). Previous participation in any clinical trial was significantly higher among non-Hispanic Black (28.1%) and Hispanic participants (33.6%) than among non-Hispanic white participants (17.1%; P = 0.025 and P < 0.001, respectively).

Study Outcomes by Disease Severity: Main Effects

Multivariable analyses adjusting for potential confounders showed that participants with self-reported moderate/severe PsA had significantly worse depressive and anxious symptomology than participants with self-reported mild PsA based on GAD-7 scores (Fig. 2a). Similarly, participants with self-reported moderate/severe PsA had significantly worse SF-36 MCS and PCS and health utility scores than those with self-reported mild PsA (Fig. 2b, c).

Fig. 2.

Multivariable analyses of patient-reported outcomes of a GAD-7, b MCS, PCS, and EQ VAS, c SF-6D and EQ 5D, and d WPAI by self-reported PsA severity. Note: Lower PHQ-9 and GAD-7 scores and higher MCS, PCS, EQ VAS, SF-6D, and EQ-5D scores indicate better HRQoL. Error bars represent standard error. Presenteeism was only assessed among participants who worked > 0 h over the last 7 days. Mean overall WPI of 41.31% should be interpreted as, on average, patients reported being 41.31% impaired in their work productivity over the past 7 days. PHQ-9, GAD-7, and WPAI outcomes were modeled using a log link with a negative binomial distribution. MCS, PCS, EQ VAS, SF-6D, and EQ-5D were modeled using an identity link with a normal distribution. Models control for age (continuous; set to mean = 45.19 years), gender (male [ref]; female), marital status (single/never married/decline to answer; married/living with a partner [ref]), educational attainment (less than a college degree/declined to answer; college graduate or higher [ref]), household income (< $25,000; $25,000 to < $50,000; $50,000 to < $100,000; $100,000+ [ref]), health insurance coverage (Medicare; Medicaid/VA/CHAMPUS; uninsured; commercial/TRICARE/don’t know [ref]), weight status (obese; overweight; underweight/normal weight/declined to answer [ref]), smoking status (current smoker; former smoker; never smoker [ref]), alcohol use (drinks alcohol; does not drink alcohol [ref]), and Charlson Comorbidity Index score (continuous; set to mean = 1.14). CHAMPUS  Civilian Health and Medical Program of the Uniformed Services, EQ-5D  EuroQol 5-Dimension, EQ-5D-5L  EuroQol 5-Dimension 5-Level, GAD-7  Generalized Anxiety Disorder-7 item, MCS  Mental Component Summary, PCS  Physical Component Summary, PHQ-9  Patient Health Questionnaire-9 item, PsA  psoriatic arthritis, SF-6D  Short Form-6 Dimension, WPAI  Work Productivity and Activity Impairment, WPI  Work Productivity Impairment, VA  Veterans Affairs, VAS  Visual Analog Scale

Labor force participation was roughly equal among participants with self-reported moderate/severe and self-reported mild PsA (Table S1 in supplementary material). However, participants with self-reported moderate/severe PsA reported greater impairment across all WPAI measures over the past 7 days than participants with self-reported mild PsA (Fig. 2d). Additionally, participants with self-reported moderate/severe PsA reported significantly more HCP and rheumatologist visits over the past 6 months than participants with self-reported mild PsA (Fig. 3a, b). Mental health provider visits and hospitalizations were not statistically different among participants with self-reported moderate/severe or self-reported mild PsA. Direct medical costs and indirect costs were significantly higher among participants with self-reported moderate/severe PsA than participants with self-reported mild PsA (Fig. 3c, d). Multivariable results for outcomes by disease severity were generally consistent with previous bivariate analyses (Table S2 in supplementary material).

Fig. 3.

Multivariable analyses of a HCP visits, b rheumatologist visits, mental health provider visits, and hospitalizations, c annualized direct medical costs, and d annualized indirect costs by PsA severity. Note: Error bars represent standard error. Indirect costs were only calculated for participants in the labor force at the time of the survey and who had a valid response (i.e., non-missing) for the number of hours worked over the past 7 days and the number of hours missed in the past 7 days (mild n = 425; moderate/severe n = 274). Outcomes were modeled using a log link with a negative binomial distribution. Models control for age (continuous; set to mean = 45.19 years), gender (male [ref]; female), marital status (single/never married/decline to answer; married/living with a partner [ref]), educational attainment (less than a college degree/declined to answer; college graduate or higher [ref]), household income (< $25,000; $25,000 to < $50,000; $50,000 to < $100,000; $100,000+ [ref]), health insurance coverage (Medicare; Medicaid/VA/CHAMPUS; uninsured; commercial/TRICARE/don’t know [ref]), weight status (obese; overweight; underweight/normal weight/declined to answer [ref]), smoking status (current smoker; former smoker; never smoker [ref]), alcohol use (drinks alcohol; does not drink alcohol [ref]), and Charlson Comorbidity Index score (continuous; set to mean = 1.14). CHAMPUS Civilian Health and Medical Program of the Uniformed Services, HCP healthcare provider, PsA psoriatic arthritis, USD US dollars, VA Veterans Affairs

Study Outcomes by Race/Ethnicity: Main Effects

In multivariable analyses assessing study outcomes by race/ethnicity, there were no differences in depression, anxiety, or HRQoL outcomes (Fig. 4a–c). Labor force participation was not significantly different among race/ethnicity groups (Table S3 in supplementary material). However, non-Hispanic Black participants with PsA reported significantly greater absenteeism and activity impairment than non-Hispanic white participants (absenteeism: 31.11% [standard error {SE} 6.83%] vs. 16.69% [SE 1.35%]; P = 0.007 and activity impairment: 54.27% [SE 3.17%] vs. 47.96% [SE 0.91%]; P = 0.047; Fig. 4d). Non-Hispanic Black participants with PsA also reported significantly fewer HCP and rheumatologist visits than non-Hispanic white participants (HCP visits: 5.93 [SE 0.61] vs. 7.42 [SE 0.24]; P = 0.039 and rheumatologist visits: 0.29 [SE 0.08] vs. 0.53 [SE 0.04]; P = 0.028; Fig. 5a, b). There were no statistically significant differences in other HCRU outcomes or in direct medical costs or indirect costs among race/ethnicity groups (Fig. 5c, d). Bivariate analyses by race/ethnicity are presented in Table S4 in the supplementary material.

Fig. 4.

Multivariable analyses of patient-reported outcomes of a GAD-7, b MCS, PCS, and EQ VAS, c SF-6D and EQ 5D, and d WPAI by race/ethnicity. Note: Lower PHQ-9 and GAD-7 scores and higher MCS, PCS, EQ VAS, SF-6D, and EQ-5D scores indicate better HRQoL. Error bars represent standard error. Presenteeism was only assessed among participants who worked > 0 h over the last 7 days. PHQ-9, GAD-7, and WPAI outcomes were modeled using a log link with a negative binomial distribution. MCS, PCS, EQ VAS, SF-6D, and EQ-5D were modeled using an identity link with a normal distribution. Models control for age (continuous; set to mean = 45.19 years), gender (male [ref]; female), marital status (single/never married/decline to answer; married/living with a partner [ref]), educational attainment (less than a college degree/declined to answer; college graduate or higher [ref]), household income (< $25,000; $25,000 to < $50,000; $50,000 to < $100,000; $100,000+ [ref]), health insurance coverage (Medicare; Medicaid/VA/CHAMPUS; uninsured; commercial/TRICARE/don’t know [ref]), weight status (obese; overweight; underweight/normal weight/declined to answer [ref]), smoking status (current smoker; former smoker; never smoker [ref]), alcohol use (drinks alcohol; does not drink alcohol [ref]), and Charlson Comorbidity Index score (continuous; set to mean = 1.14). CHAMPUS Civilian Health and Medical Program of the Uniformed Services, EQ-5D EuroQol 5-Dimension, EQ-5D-5L EuroQol 5-Dimension 5-Level, GAD-7 Generalized Anxiety Disorder-7 item, MCS Mental Component Summary, PCS Physical Component Summary, PHQ-9 Patient Health Questionnaire-9 item, SF-6D Short Form-6 Dimension, WPAI Work Productivity and Activity Impairment, VA Veterans Affairs, VAS Visual Analog Scale

Fig. 5.

Multivariable analyses of a HCP visits, b rheumatologist visits, mental health provider visits, and hospitalizations, c annualized direct medical costs, and d annualized indirect costs by race/ethnicity.  Note: Error bars represent standard error. Indirect costs were only calculated for participants in the labor force at the time of the survey and who had a valid response (i.e., non-missing) for the number of hours worked over the past 7 days and the number of hours missed in the past 7 days (non-Hispanic white n = 501; non-Hispanic Black n = 79; Hispanic n = 119). Outcomes were modeled using a log link with a negative binomial distribution. Models control for age (continuous; set to mean = 45.19 years), gender (male [ref]; female), marital status (single/never married/decline to answer; married/living with a partner [ref]), educational attainment (less than a college degree/declined to answer; college graduate or higher [ref]), household income (< $25,000; $25,000 to < $50,000; $50,000 to < $100,000; $100,000+ [ref]), health insurance coverage (Medicare; Medicaid/VA/CHAMPUS; uninsured; commercial/TRICARE/don’t know [ref]), weight status (obese; overweight; underweight/normal weight/declined to answer [ref]), smoking status (current smoker; former smoker; never smoker [ref]), alcohol use (drinks alcohol; does not drink alcohol [ref]), and Charlson Comorbidity Index score (continuous; set to mean = 1.14). CHAMPUS Civilian Health and Medical Program of the Uniformed Services, HCP healthcare provider, USD US dollars, VA Veterans Affairs.

Interaction Between Race/Ethnicity and Disease Severity

Beyond the influence of potential confounding variables, there was evidence to suggest that associations between perceived PsA severity and several study outcomes were modified by race/ethnicity (Table S5 in supplementary material). Multivariable analyses indicated that race/ethnicity moderated the relationship between perceived PsA severity and depression status (i.e., PHQ-9 score ≥ 10) (LRT X² = 12.76; P = 0.005). These results showed that the odds of having moderate/severe depression (i.e., PHQ-9 score ≥ 10) among non-Hispanic white participants with self-reported moderate/severe PsA were 2.83 times that of those with self-reported mild PsA (Fig. 6a). In other words, the predicted probability of having moderate/severe depression among participants who were non-Hispanic white was 46.3% (SE 5.7%) for those with self-reported moderate/severe PsA and 23.4% (SE 4.1%) for those with self-reported mild PsA.

Fig. 6.

Multivariable analyses of a depression status, b SF-6D, and c number of ER visits stratified by race/ethnicity and PsA severity. Note: Results presented are from covariate adjusted models where the interaction between condition severity and race/ethnicity was statistically significant. Error bars represent standard error. Models control for age (continuous; set to mean = 45.19 years), gender (male [ref]; female), marital status (single/never married/decline to answer; married/living with a partner [ref]), educational attainment (less than a college degree/declined to answer; college graduate or higher [ref]), household income (< $25,000; $25,000 to < $50,000; $50,000 to < $100,000; $100,000+ [ref]), health insurance coverage (Medicare; Medicaid/VA/CHAMPUS; uninsured; commercial/TRICARE/don’t know [ref]), weight status (obese; overweight; underweight/normal weight/declined to answer [ref]), smoking status (current smoker; former smoker; never smoker [ref]), alcohol use (drinks alcohol; does not drink alcohol [ref]), and Charlson Comorbidity Index score (continuous; set to mean = 1.14). CHAMPUS Civilian Health and Medical Program of the Uniformed Services, ER emergency room, PHQ-9 Patient Health Questionnaire-9 Item, PsA psoriatic arthritis, SF-6D Short Form-6 Dimension, VA Veterans Affairs

Race/ethnicity also moderated the relationship between perceived PsA severity and SF-6D scores (LRT X² = 3.12; P = 0.025), such that non-Hispanic white and other race participants with self-reported moderate/severe PsA had significantly lower estimated mean SF-6D scores than those of the same race/ethnicity group with self-reported mild PsA (non-Hispanic white: 0.56 [SE 0.01] vs. 0.63 [SE 0.01]; P < 0.001 and other race: 0.57 [SE 0.01] vs. 0.61 [SE 0.02]; P = 0.049; Fig. 6b). The association between perceived PsA severity and ER visits was also moderated by race/ethnicity (LRT X² = 10.87; P = 0.013), such that non-Hispanic white participants with self-reported moderate/severe PsA had significantly more ER visits over the past 6 months than those with self-reported mild PsA (0.70 [SE 0.14] vs. 0.45 [SE 0.09]; P = 0.001; Fig. 6c).

Discussion

Psoriatic arthritis poses a considerable burden to both individuals and healthcare systems, as evidenced by our study findings. This study aimed to assess the participant-reported humanistic and economic burden of PsA, relative to participants’ perceptions of condition severity and race/ethnicity. As expected, our results showed that self-reported moderate/severe PsA was associated with significantly worse health outcomes than self-reported mild PsA. However, in the context of race/ethnicity, we noted variation in health outcomes. Specifically, non-Hispanic Black participants reported significantly more absenteeism and activity impairment, and lower HCP and rheumatologist visits than non-Hispanic white participants. Moreover, the influence of perceived PsA severity on depression, HRQoL, and HCRU depended on race/ethnicity, particularly among non-Hispanic white participants. Notably, self-reported moderate/severe PsA was associated with greater odds of depression, lower SD-6D, and more ER visits specifically among non-Hispanic white participants.

Psoriatic arthritis is associated with substantial humanistic and economic burden [7, 20, 31–40]. Worsening of individual symptoms and manifestations of PsA are associated with reduced general and PsA-specific quality of life [39–42]. Considering self-reported severity is crucial for understanding disease burden. While we used perceived PsA severity in this study rather than a more objective clinical measure, we posit that patients’ perception of their disease severity is uniquely important, and our study supports the notion that patients’ perceived PsA severity is independently related to their disease burden. Symptoms identified as most problematic by patients (i.e., painful, inflamed, or broken skin) may not be considered the most bothersome by physicians, highlighting a potential discordance in physician and patient perceptions of PsA symptoms and their impact [43]. Our findings are among the first to show that participants with self-reported moderate/severe PsA reported significantly worse HRQoL scores and greater work and activity impairment than participants with mild PsA. Mean differences in HRQoL measures of PCS and EQ-5D-5L between participants with perceived moderate/severe and mild PsA exceeded clinical significance or established minimal clinically important differences [44, 45], demonstrating the negative impact of PsA severity on HRQoL. Participants with perceived moderate/severe PsA also reported more HCP and rheumatologist visits than those with perceived mild PsA and accumulated more annual direct and indirect medical costs. Our findings are echoed by several studies that demonstrated the detrimental impact of PsA manifestations on HRQoL, functional status, and work productivity using more objective measures of PsA severity [41, 42]. Concomitant presence of multiple PsA symptoms/manifestations and worse functional status have also been shown to be associated with greater healthcare resource use and costs [46, 47]. Existing literature, coupled with the results of our study based on participant-perceived disease severity, suggest individuals with moderate/severe PsA represent an underserved population with substantial unmet needs. Still, we recognize that our measure of PsA severity may not fully reflect the clinical severity of their disease and may be confounded by other socio-emotional and/or health considerations.

There is a paucity of data on racial and ethnic disparities in patients with PsA with respect to PROs, including HRQoL, WPAI, and economic outcomes. In our study that sought to quantify the humanistic and economic burden of PsA by measuring 21 PROs, self-reported moderate/severe disease was associated with worse burden than self-reported mild disease across all race/ethnic groups evaluated. Of note, our results suggest non-Hispanic Black participants may experience greater absenteeism and overall activity impairment than other race/ethnicity groups. Non-Hispanic Black participants also reported fewer HCP and rheumatologist visits over the past 6 months in our study, even after adjusting for disease severity. Barriers to access to care among non-Hispanic Black patients could prevent adequate and timely treatment of PsA, potentially leading to less controlled disease and increased interference with workplace productivity and activity [17, 48]. Efforts to improve access to care are needed to achieve better outcomes for patients of minority racial/ethnic groups.

Our findings among non-Hispanic Black patients differed from that reported by Ogdie et al. [49], which showed similar rates of rheumatologist visits across race/ethnicity groups with PsA by insurance coverage. It is important to note that it is difficult to make apples-to-apples comparisons between studies given differences in the datasets used, study methods, and participant populations. Specifically, the Ogdie study was descriptive in nature and did not conduct any statistical analyses to identify differences between race/ethnicity groups. And as a claims-based analysis, patients insured with Medicare other than Medicare Advantage or the uninsured were not included. In contrast, the current study conducted statistical comparison between race/ethnicity groups and adjusted for potentially confounding variables across these groups. While the NHWS dataset was designed to be nationally representative of the general US population, and not for any specific disease cohort, participants were not restricted on the basis of presence or type of coverage. Therefore, our study provides unique information on how patient experiences may compare between race/ethnicity groups and additional perspectives for patients who are frequently excluded from clinical studies relying on other secondary data sources such as claims data.

Evidence suggests that non-Hispanic Black, Asian, and Hispanic populations are approximately 40% less likely to see a dermatologist for psoriasis than non-Hispanic white individuals [50]. The clinical presentation of psoriasis in skin of color can contribute to delayed diagnosis and greater disease severity, leading to untreated or undertreated disease [14, 15, 51]. Additionally, among individuals with newly diagnosed PsA, time to treatment initiation is significantly longer among non-Hispanic Black patients than among non-Hispanic white patients [17]. Other factors that may impact access to care include skepticism about the effectiveness of clinicians, perception of symptom severity, satisfaction with current treatment, and financial- or insurance-related issues [42]. For instance, Hispanic patients with lighter skin tones may exhibit symptoms more congruent with those of non-Hispanic white patients, potentially facilitating timely diagnosis and adequate care. To comprehensively understand these nuances, future analyses should incorporate factors such as socioeconomic status, cultural variation, and regional influences.

Our moderation analyses suggest that the influence of perceived PsA severity on outcomes may vary across race/ethnicity groups. Specifically, our results show that race/ethnicity moderated the relationship between self-reported PsA severity and mental health, such that moderate/severe PsA was associated with a greater odds of having moderate/severe depression compared to mild PsA, but only among non-Hispanic white participants. For all other racial/ethnic groups, we found no evidence of a significant association between perceived PsA severity and depression status. It is unclear why more severe PsA would be linked to poorer mental health among one racial/ethnic group, but not others. While the prevalence of depression is lower among non-Hispanic Black individuals than non-Hispanic white individuals in the USA, rates of depression differ when accounting for ethnicity and immigration status [52, 53]. Further, current instruments used to assess mental health may not adequately capture the experiences of non-Hispanic Black individuals [53]. Similarly, the prevalence of major depression differs among Hispanic ethnic subgroups, despite having similar lifetime prevalence of depressive disorder as non-Hispanic whites overall [53–55]. Cultural beliefs around the existence and causes of mental health issues may influence the observed disparity in prevalence [53]. The complex relationship between familial, contextual, and social factors, as well as other factors such as age of arrival in the USA, degree of acculturation, and language proficiency, may influence the risk of mental health issues such as depression.

Our results also suggest that race moderates the interaction between disease severity and ER visits. Healthcare resource use in individuals with PsA is high, irrespective of disease severity [38]. In a matched analysis of patients with and controls without PsA and psoriasis, the rate of annual ER visits and the proportion of patients with ER visits were higher among patients with PsA [38]. Additionally, racial/ethnic differences in ER visits, although not specific to PsA, are well documented. According to the National Center for Health Statistics, in 2021, non-Hispanic Black individuals had the highest overall ER visit rate, at double that for non-Hispanic white and Hispanic individuals [56]. These higher rates of ER visits among non-Hispanic Black people have been previously attributed to receiving routine healthcare in the emergency department [57–59]. Our findings show that non-Hispanic white participants with self-reported moderate/severe PsA reported more frequent ER visits than those with self-reported mild PsA, whereas non-Hispanic Black and Hispanic people with self-reported moderate/severe PsA reported a reduction in the rate of ER visits compared with those with self-reported mild PsA, though the estimates were not significant. While the results for non-Hispanic white participants align with expectations for more severe PsA, the observed trends in non-Hispanic Black and Hispanic groups raise the need for further investigation. It is possible that the increase in ER visits among non-Hispanic white participants is attributable to impaired mental health and HRQoL rather than PsA severity, which might suggest that those with more severe disease are more likely or willing to seek healthcare [60] or have access to more healthcare resources, the latter of which is supported by the greater overall HCP visits among those with self-reported moderate/severe PsA observed in this analysis. These results provide insights on racial/ethnic differences in outcomes and propose variations in self-perception of PsA severity based on race/ethnicity (i.e., people of color may underestimate the severity of their disease or overestimate their mental health). Future efforts exploring nuances within non-Hispanic Black and Hispanic populations and/or using more objective measures of condition severity may help to better understand these results.

Our study had several strengths, including the use of PROs, which provide participant perspectives on disease outcomes over third-party perceptions of disease burden. Racial/ethnic PRO data collection can aid in understanding the impacts of PsA and help inform clinical care. Further, this study used NHWS data, and provides PRO data on a diverse group of participants with PsA. The extensive roster of outcomes assessed in this study provides a robust contribution to the literature on a variety of PROs, increases the validity of the study findings through consistency across measures, and contributes to a holistic understanding of patient burden in PsA.

Limitations of this study include the self-reported nature of race/ethnicity and disease severity, which could not be independently verified [61–63]. Because race and ethnicity are social constructs that categorize people on the basis of perceived shared physical traits and culture [64], self-reporting of race/ethnicity is generally considered the gold standard [65–67]. And, while potentially confounding covariates were adjusted for in multivariable analyses, participant perception of disease severity may differ even with known clinical and social characteristics being equal. Another limitation was the relatively small sample sizes of non-Hispanic Black and Hispanic groups, despite pooling data across three survey years. Additionally, other than non-Hispanic white, non-Hispanic Black, and Hispanic groups, participants of all other races/ethnicities were included in the “Other” category, limiting the conclusions we could draw from our study. The survey was conducted in English within a virtual setting and may therefore introduce selection bias, as non-English-speaking individuals and individuals not appropriately targeted by a web-based survey could be underrepresented in our sample [68]. The study population was also generally more educated and affluent than historical national averages and may not therefore be reflective of the overall PsA population within the USA, as the use of data from the NHWS dataset was designed to be nationally representative of the general US population, and not for any specific disease cohort. The study sample may also not be representative of the diverse Hispanic population in the USA, as the proportion of uninsured Hispanics in our dataset contrasted with 2022 US Census data [69]. Notably, the NHWS dataset did include participants with diverse experiences and sociodemographic characteristics such as patients who were uninsured and, therefore, frequently omitted from secondary data analyses using real-world data. Further, cost data may underestimate the financial impact in the current post-COVID era, as medical costs were calculated using 2018 data. Additionally, residual confounding may have biased multivariable models, despite attempts to adjust for other potential explanatory variables. As a result of the cross-sectional nature of the study, causal inference was not possible.

Conclusions

This study contributes to literature showing that individuals with self-reported moderate/severe PsA experience greater patient-reported humanistic and economic burden than those with self-reported mild PsA. Further, non-Hispanic Black participants experienced greater humanistic and economic burden across some measures, suggesting racial/ethnic variation in the experiences of patients with PsA. Additionally, the observation that race/ethnicity moderates the relationship between self-reported PsA severity and outcomes such as depression status, HRQoL, and ER visits is significant, particularly as these associations were pronounced among non-Hispanic white participants. This underscores the need for nuanced understanding of the intricate interplay between self-reported PsA severity and race/ethnicity. Further research is needed to explore the potential factors influencing the relationship between self-reported PsA severity and race/ethnicity, with an overall goal of addressing unmet needs in PsA management.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

All authors (Iris Lin, Kathryn Krupsky, Nate Way, Aarti A Patel, Arlene Tieng) made substantial contributions to the study conception and design. Data collection and analysis were performed by Kathryn Krupsky and Nate Way. All authors contributed to the interpretation of the data and provided critical feedback on the drafted manuscript. All authors provided final approval of the version to be published.

Funding

This study and the journal’s Rapid Service Fee were funded by Janssen Pharmaceuticals.

Data Availability

The datasets generated during and/or analyzed during the current study are not publicly available because the participants of this study did not give written consent for their data to be shared publicly.

Declarations

Conflict of Interest

Iris Lin and Aarti A. Patel are employees of Janssen Scientific Affairs (JSA), LLC (a Johnson & Johnson company) and hold stock in Johnson & Johnson. Kathryn Krupsky and Nate Way are employees of Oracle Life Sciences and paid consultants of JSA. Dr. Arlene Tieng is an employee of BronxCare Health System and Icahn School of Medicine at Mount Sinai, and is a paid consultant of JSA.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.