Abstract
Objective:
Medication nonadherence challenges the management of systemic autoimmune rheumatic diseases (SARDs). We investigated cost-related medication behaviors among patients with SARDs, and compared to those of patients without SARDs, in a large diverse cohort across the US.
Methods:
As part of the All of Us Research (All of Us) Program (version 7), a nationwide diverse adult cohort with linked electronic health records (EHR) begun in 2017, participants completed questionnaires concerning cost-related medication behaviors. Chi-square tests compared responses between patients with SARDs, by disease and medication type, and to those without SARDs. Logistic regression analyses were used to calculate odds ratios (ORs, 95% confidence intervals).
Results:
We analyzed data from 3,997 patients with SARDs and 73,990 participants without SARDs. After adjustment, patients with vs. without SARDs had 1.56 (95%CI 1.43-1.70) times increased odds of reporting unaffordability of prescription medicines, 1.43 (95%CI 1.31-1.56) times increased odds of cost-related medication nonadherence, and 1.23 (95%CI 1.14-1.32) times increased odds of using cost-reducing strategies. Patients with SARDs who reported unaffordability were 16.5% less likely to receive a disease-modifying drug (95%CI 0.70-0.99), but 18.1% more likely to receive glucocorticoids (95%CI 0.99-1.42). In addition, unaffordability of prescription medicines was likely to have a 1.27 (95%CI 1.03-1.57) and 1.38-fold (95%CI 0.96-1.99) increased odds of 1-2 and ≥3/year emergency room visits.
Conclusion:
In this large diverse cohort, patients with vs. without SARDs had more self-reported cost-related medication behaviors and those who reported medication unaffordability used fewer disease modifying drugs and had more emergency room visits.
Keywords: access to care, medication adherence, cost, systemic autoimmune rheumatic diseases
Introduction
Systemic autoimmune rheumatic diseases (SARD), such as rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and psoriatic arthritis (PsA), are chronic inflammatory disorders that mainly, but not exclusively, affect connective tissues like joints and muscles.1,2 With the introduction of newer biological agents, significant advances have been achieved in the treatments of SARDs in recent years.3 However, the high cost of biologic and targeted synthetic disease-modifying anti-rheumatic drugs (DMARDs) has restricted their use and can lead to care inequities.4–6
Medication nonadherence is a challenging issue in the management of chronic diseases, in particular SARDs.7–8 The prevalence of medication nonadherence in patients with SARDs has varied widely, reported to range from 6% to 90% in past studies.9–11 Suboptimal adherence has been associated with worse clinical outcomes, including higher disease activity, and increased hospitalization and emergency department visits, all of which can contribute to increasing health care cost and disease burden.12–15 Many factors are reported to be associated with medication nonadherence, including patient-related factors (e.g., age, gender, race, and ethnicity), therapy-related factors (e.g., complexity, side effects), and condition-related factors (e.g., mental health, prognosis); medication cost is one of the most important components contributing to poor adherence.9,10,16,17 Thus, understanding cost-related medication nonadherence is important, but has less well studied among patients with SARDs in recent years. Several studies have been performed among individuals with RA or SLE,18–20 but not among those with other SARDs. Therefore, we aimed to investigate the prevalence of cost-related medication behaviors in patients with and without SARDs and their associations with prescription rates for DMARDs and glucocorticoids and health care utilization in a large, diverse cohort across the United States.
Patients and Methods
Study population
The All of Us Research Program (All of Us) is a nationwide longitudinal cohort study aiming to enroll 1 million diverse participants across the U.S. The procedures of All of Us have been previously described in detail.21,22 The institutional review board of the All of Us approved the operational protocol, and all participants provided informed consent at the time of enrollment. Briefly, adults aged 18 years or older are eligible to participate through All of Us affiliated healthcare provider organizations or directly through the All of Us website or certain events. After enrollment, participants were asked to complete several surveys, which are provided in both English and Spanish at a fifth-grade reading level. Each survey is intended to be completed in about 15 minutes. They were asked to agree to share their electronic health records (EHR), undergo physical measurements and donate biospecimens.
Among the participants with baseline survey and linked EHR data in All of Us version 7 (C2022Q4R9; released April 2023), participants who fully answered the seven questions on cost-related medication behaviors and received at least one prescription of medications within one year of enrollment were included in the study. Among them, patients with SARDs, including RA, SLE, PsA, ankylosing spondylitis, Sjögren’s disease, systemic sclerosis, dermatomyositis/polymyositis, mixed/undifferentiated connective tissue diseases, and vasculitis, were identified if they had at least two International Classification of Diseases (ICD)-9, ICD-10 and Systematized Nomenclature of Medicine (SNOMED) codes at least 60 days apart within two years before their enrollment.23,24 The comparison group for the first part of the analysis consisted of those who did not have any code for SARDs before enrollment.
Questionnaires of cost-related medication behaviors
From the Health Care Access and Utilization survey in All of Us, derived from the National Health Interview Survey, we used the seven questions including three indices of cost-related medication behaviors: unaffordability of prescription medicine, cost-related medication nonadherence, and cost-reducing strategies .25,26 Unaffordability of prescription medicine was assessed by asking whether there was a time when they needed prescription medication, but did not get them because they could not afford them. Cost-related medication nonadherence, a binary variable, was defined as an indication of “yes” to any of three questions: (1) skipping medications, (2) taking less medicines, and (3) delaying filling a prescription to save money. Following cost-reducing strategies was defined as answering “yes” to any of three questions: (1) asking their doctor for a lower cost medication, (2) buying prescription drugs from another country, and (3) using alternative therapies to save money.
Variables
From the basic survey in All of Us, we extracted the following data: age, gender, race and ethnicity, household income, education level, insurance type, employment status, and smoking status. Data on primary spoken language was not available. Geographic region based on state of enrollment was categorized as Northeast, Midwest, West, and South. The area deprivation index was derived from the 2017 American Communities Survey measures and calculated by taking the population-weighted average of the index for the US Census tracts based on their three-digit zip code.27 Obesity was defined by body mass index (BMI) ≥30 kg/m2 from the baseline in-person physical measurement data. The Charlson Comorbidity Index (CCI) was assessed using ICD-9, ICD-10, and SNOMED codes from the EHR within the year prior to the enrollment.28 Medication history was collected using RxNorm from the EHR within the year prior to the enrollment, and the number of medications was defined as the maximum number of prescribed medications on a single day, and categorized into three groups: 1-4, 5-9 (i.e., polypharmacy), and >=10 (i.e., excessive polypharmacy).29 DMARD and glucocorticoid use was coded by their ingredients using RxNorm. Health care utilization, including inpatient visits, emergency room visits, and outpatient visits, were identified within the 1 year prior to the enrollment.
Statistical analysis
The Chi-square test compared self-reported responses regarding cost-related medication behaviors. We compared responses among patients with and without SARDs, and additionally, we compared each type of SARD with those without. Unadjusted and adjusted odds ratios (ORs) with 95% confidence intervals (95% CIs) were calculated using logistic regression for cost-related medication behaviors, and analyzed in three models, the first adjusting for number of medications, the second further adjusting for age, gender, race and ethnicity, and geographical region, and the last further adjusting for annual household income, education level, insurance type, employment status, area deprivation index, and CCI. A sensitivity analysis was conducted to recategorize patients with 2 or more SARDs into the SARD category for which they had the highest number of codes (ICD-9, ICD-10 or SNOMED). In this analysis, we excluded patients with an equal number of codes for more than 1 SARD as it was still difficult to distinguish which was the primary disease.
We used logistic regression models to examine unadjusted and adjusted associations of each cost-related medication behavior with the prescriptions of DMARDs (any vs. none) and glucocorticoids (any vs. none) among the patients with SARDs. Covariates were number of medications including DMARDs and/or glucocorticoids, age group, gender, race and ethnicity, insurance type, CCI, and types of SARDs.
Among the patients with SARDs, we also compared annual healthcare utilization in those with or without each type of cost-related medication behavior. Here, in each model, the independent variable was the cost-related medication behavior and the dependent variable was (1) baseline annual inpatient visits (0 (ref), 1-2, 3+), (2) baseline annual emergency department visits (0 (ref), 1-2, 3+), or (3) baseline annual outpatient visits (0 (ref), 1-9, 10+). For these healthcare utilization outcomes, multinomial logistic regression analyses were performed. Models were adjusted for number of medications, age group, gender, race and ethnicity, insurance type, CCI, and types of SARDs.
A p-value <0.05 was considered statistically significant. All statistical analyses were performed using R version 4.1.0 in a Jupyter Notebook contained in the All of Us workbench.
Results
Among 185,902 participants with at least 1 prescription within 1 year of enrollment, 80,663 participants answered all seven questions of interest (Figure 1). Those who answered the survey tended to be older and more educated than those who did not, and more respondents were female and non-Hispanic White (Supplementary Table 1). Among them, we identified 3,997 patients with SARDs and 73,990 participants without any SARD diagnosis. Of them, 1,330 (33.3%) had RA, 468 (11.7%) had SLE, 373 (9.3%) had Sjögren’s disease, 282 (7.1%) had PsA, 279 (7.0%) had ankylosing spondylitis, 441 (11.0%) had other SARDs including systemic sclerosis (72; 1.8%), dermatomyositis/polymyositis (41; 1.0%), mixed/undifferentiated connective tissue diseases (163; 4.1%), and vasculitis (165; 4.1%), and 824 (20.6%) had two or more type of SARDs. The characteristics of study population were described in Table 1. The patients with SARDs had a mean age of 57.5 years, and 78.4% were women, whereas the participants without SARDs had a lower mean age of 53.7 years and a lower percentage were women (63.3%). In both groups, approximately 71% self-identified as Non-Hispanic White Americans, 10% as Non-Hispanic Black Americans, and 11% as Hispanic Americans. Patients with SARDs were more likely than those without to have lower annual household income, be less educated, have public insurance, and be unemployed. Compared to those without SARDs, the patients with SARDs had higher mean CCI (1.5 vs. 1.0), calculated without rheumatic disease (potential range 0-32). As expected, more patients with than without SARDs were prescribed conventional DMARDs (41.7% vs. 2.8%), biologic and/or targeted synthetic DMARDs (19.0 % vs. 0.7%), and glucocorticoids (45.9% vs. 23.5%), along with 5 or more medications (56.6% vs. 41.6%).
Figure 1. Flow chart of study population.
SARDs: Systemic autoimmune rheumatic diseases.
Table 1.
Demographic and survey data of participants with and without systemic autoimmune rheumatic diseases (SARDs) in All of Us Research Program who responded to the Health Care Utilization survey (data release version 7)
| Characteristics | Participants with SARDs (n=3997) | Participants without SARDs (n=73990) |
|---|---|---|
| Age category, years, (%) | ||
| 18-49 | 1061 (26.5) | 27821 (37.6) |
| 50-64 | 1491 (37.3) | 22823 (30.8) |
| ≥65 | 1445 (36.2) | 23346 (31.6) |
| Gender | ||
| Women | 3133 (78.4) | 46830 (63.3) |
| Men | 749 (18.7) | 25089 (33.9) |
| Others/Prefer not to answer/Missing | 115 (2.9) | 2071 (2.8) |
| Self-reported race and ethnicity (%) | ||
| Hispanic Americans | 424 (10.6) | 7917 (10.7) |
| Non-Hispanic Black Americans | 437 (10.9) | 7307 (9.9) |
| Non-Hispanic White Americans | 2806 (70.2) | 52824 (71.4) |
| Non-Hispanic Others/Prefer not to answer/Missing | 330 (8.3) | 5942 (8.0) |
| Enrollment year | ||
| 2018 | 1105 (27.6) | 19391 (26.2) |
| 2019 | 1336 (33.4) | 24767 (33.5) |
| 2020 | 417 (10.4) | 7981 (10.8) |
| 2021 | 701 (17.5) | 13606 (18.4) |
| 2022 | 438 (11.0) | 8245 (11.1) |
| Annual household income, $ | ||
| <35k | 1058 (26.5) | 16671 (22.5) |
| 35k-<100k | 1294 (32.4) | 24667 (33.3) |
| ≥100k | 1072 (26.8) | 22973 (31.0) |
| Prefer not to answer/Missing | 573 (14.3) | 9679 (13.1) |
| Education level | ||
| High school graduate or less | 657 (16.4) | 11183 (15.1) |
| Some college | 1163 (29.1) | 18318 (24.8) |
| College graduate or above | 2082 (52.1) | 42766 (57.8) |
| Prefer not to answer/Missing | 95 (2.4) | 1723 (2.3) |
| Insurance typea | ||
| Private | 1926 (48.2) | 39045 (52.8) |
| Public | 1510 (37.8) | 23352 (31.6) |
| Uninsured | 41 (1.0) | 2011 (2.7) |
| Prefer not to answer/Missing | 520 (13.0) | 9582 (13.0) |
| Employment status | ||
| Employed/Self-employed | 1559 (39.0) | 38397 (51.9) |
| Unemployed | 163 (4.1) | 3659 (4.9) |
| Othersb | 2160 (54.0) | 30051 (40.6) |
| Prefer not to answer/Missing | 115 (2.9) | 1883 (2.5) |
| Region | ||
| Northeast | 1649 (41.3) | 26129 (35.3) |
| Midwest | 1102 (27.6) | 23373 (31.6) |
| West | 733 (18.3) | 14998 (20.3) |
| South | ≤513 (≤12.8) | 9453 (12.8) |
| Missing | <20 | 37 (0.1) |
| Area deprivation index, median (IQR)c | 0.30 (0.07) | 0.30 (0.06) |
| Smoking | ||
| Ever | 1590 (39.8) | 27170 (36.7) |
| Never | 2344 (58.6) | 45558 (61.6) |
| Missing | 63 (1.6) | 1262 (1.7) |
| BMI (kg/m2) | ||
| ≥30 | 1799 (45.0) | 29950 (40.5) |
| <30 | 2083 (52.1) | 41159 (55.6) |
| Missing | 115 (2.9) | 2881 (3.9) |
| Charlson comorbidity indexd | 1.5 ± 2.1 | 1.0 ± 1.9 |
| Myocardial infarction | 109 (2.7) | 1266 (1.7) |
| Congestive heart failure | 264 (6.6) | 2569 (3.5) |
| Peripheral vascular disease | 256 (6.4) | 2265 (3.1) |
| Cerebrovascular disease | 253 (6.3) | 2282 (3.1) |
| Diabetes | 681 (17.0) | 9820 (13.3) |
| Chronic pulmonary disease | 980 (24.5) | 9644 (13.0) |
| Liver disease | 329 (8.2) | 3577 (4.8) |
| Renal disease | 420 (10.5) | 3518 (4.8) |
| Number of medications | ||
| 1-4 | 1736 (43.4) | 43247 (58.4) |
| 5-9 | 983 (24.6) | 14598 (19.7) |
| ≥10 | 1278 (32.0) | 16145 (21.8) |
| Medications | ||
| Conventional DMARDse | 1666 (41.7) | 2070 (2.8) |
| Biologic DMARDs and/or targeted synthetic DMARDsf | 760 (19.0) | 540 (0.7) |
| Glucocorticoidsg | 1834 (45.9) | 17362 (23.5) |
| Types of SARDs | ||
| Rheumatoid arthritis | 1330 (33.3) | - |
| Systemic lupus erythematosus | 468 (11.7) | - |
| Sjögren’s disease | 373 (9.3) | |
| Psoriatic arthritis | 282 (7.1) | - |
| Ankylosing spondylitis | 279 (7.0) | - |
| Other SARDs | 441 (11.0) | - |
| 2 or more SARDs | 824 (20.6) | - |
BMI: Body mass index; DMARD: disease modifying anti-rheumatic drug; SARD: Systemic autoimmune rheumatic disease; SMD: Standardized mean difference. According to All of Us Research Program data sharing policies, cells with less than 20 participants were suppressed.
Reports of both private and public insurance considered as private. Reports of having only Indian Health Service coverage considered uninsured.
Others including homemaker, students, retired, and unable to work (disabled).
Area deprivation index was derived from the 2017 American Communities Survey based on their three-digit zip code. There were 33 participants with missing data.
Charlson comorbidity index calculated using 16 comorbidities, excluding rheumatic disease.
Conventional DMARDs included azathioprine, cyclophosphamide, cyclosporine, hydroxychloroquine, leflunomide, methotrexate, mycophenolate/mycophenolic acid, sulfasalazine, and tacrolimus.
Biologic DMARDs included abatacept, adalimumab, anakinra, belimumab, brodalumab, canakinumab, certolizumab, etanercept, golimumab, guselkumab, infliximab, ixekizumab, ocrelizumab, rilonacept, risankizumab, rituximab, sarilumab, secukinumab, tildrakizumab, tocilizumab, and ustekinumab. Target synthetic DMARDs included baricitinib, tofacitinib, upadacitinib.
Glucocorticoids included dexamethasone, methylprednisolone, prednisolone, and prednisone.
Table 2 shows the prevalence of cost-related medication behaviors among patients with and without SARDs. Patients with SARDs were significantly more likely than those without SARDs to report unaffordability of prescription medicines: 18.3% of patients with SARDs, and 11.9% of those without (p < 0.001). Patients with SARDs were 1.5-fold more likely to experience cost-related medication nonadherence than those without (18.1% vs. 13.1%, p < 0.001), and similar trends were observed in all items of cost-related medication nonadherence, as follows: skipping medication doses (9.9% vs. 6.9%, p < 0.001), taking less medicines (10.8% vs. 7.7%, p <0.001), and delaying filling a prescription (15.6% vs. 11.0%, p < 0.001). The self-reported prevalence of using cost-reducing strategies was also significantly higher among patients with SARDs than those without (28.7% vs. 24.1%, p < 0.001). There were significant differences in the prevalence of requesting a lower cost medication (24.3% vs. 19.4%, p < 0.001) and using alternative therapies (8.1% vs. 7.1%, p 0.02), but no difference in buying prescription drugs from another country (3.2% vs. 3.3%, p 0.97).
Table 2.
Prevalence of cost-related medication behaviors in patients with and without systemic autoimmune rheumatic diseases (SARDs) in All of Us Research Program (data release version 7)
| Patients reported cost-related medication behavior | Participants with SARDS (n=3997) | Participants without SARDs (n= 73990) | P valuea |
|---|---|---|---|
| Unaffordability of prescription medicines | 731 (18.3) | 8837 (11.9) | <0.001 |
| Cost-related medication nonadherence | 725 (18.1) | 9678 (13.1) | <0.001 |
| Skipped medication doses | 396 (9.9) | 5092 (6.9) | <0.001 |
| Took less medicine | 433 (10.8) | 5661 (7.7) | <0.001 |
| Delayed filling a prescription | 622 (15.6) | 8138 (11.0) | <0.001 |
| Cost-reducing strategies for prescription medication | 1148 (28.7) | 17803 (24.1) | <0.001 |
| Asked for a lower cost medication | 971 (24.3) | 14391 (19.4) | <0.001 |
| Bought prescription drugs from another country | 129 (3.2) | 2407 (3.3) | 0.97 |
| Used alternative therapies | 322 (8.1) | 5251 (7.1) | 0.02 |
SARD: systemic autoimmune rheumatic disease.
Chi-squared test
The demographic and survey responses of the patients with SARD by disease are presented in Supplementary Table 2 and Figure 2. The prevalence of reporting unaffordability of prescription medicines was 16.3%, 22.2%, 16.1%, 17.7%, 17.9%, 15.4%, and 22.1% among patients with RA, SLE, Sjögren’s disease, PsA, ankylosing spondylitis, other SARDs, and two or more type of SARDs, respectively. For cost-related medication non-adherence, 16.7%, 20.1%, 18.2%, 16.3%, 17.9%, 15.4%, and 21.5% of patients with RA, SLE, Sjögren’s disease, PsA, ankylosing spondylitis, other SARDs, and two or more type of SARDs answered yes, whereas 25.8%, 30.3%, 33.5%, 34.8%, 23.3%, 27.4%, and 30.8% of patients with each disease reported yes for cost-reducing strategies. Among those with these diseases, the highest proportion of patients reporting unaffordability of prescription medicines were those with SLE (22.2%), followed by those with two or more types of SARDs (22.1%), ankylosing spondylitis (17.9%), and PsA (17.7%). Patients with two or more types of SARDs and those with SLE also reported the highest cost-related medication nonadherence (21.5% and 20.1%, respectively), whereas those with PsA reported the most cost-reducing strategies (34.8%).
Figure 2. Prevalence of cost-related medication behaviors in patients with different types of systemic autoimmune rheumatic diseases (SARDs) in All of Us Research Program (data release version 7).
SARDs: Systemic autoimmune rheumatic diseases. Cost-related medication nonadherence included skipping medication doses, taking less medicine, and delayed filling a prescription to save money. Cost-reducing strategies for prescription medication included asking for a lower cost medication, buying prescription drugs from another country, and using alternative therapies to save money.
Table 3 displays the results of univariable and multivariable-adjusted logistic regression analyses for cost-related medication behaviors. After sequentially adjusting for number of medications (minimally-adjusted), adding demographic factors (demographics-adjusted), and adding socioeconomic factors and CCI (fully-adjusted), the patients with SARDs had 1.56 (95% CI: 1.43-1.70) times increased odds of unaffordability of prescription medications, 1.43 (95% CI: 1.31-1.56) times increased odds of cost-related medication nonadherence, and 1.23 (95% CI: 1.14-1.32) times increased odds of cost-reducing strategies. When comparing types of SARDs to those without any SARD in unadjusted models, all types of SARDs had significantly increased odds of having all three cost-reducing behaviors than did non-SARDs patients, with the exception of cost-related medication non-adherence among those with PsA and cost-reducing among those with RA and ankylosing spondylitis, among whom the odds were not significantly elevated. After multivariable adjustment, all types of SARDs were seen to have increased odds of reporting unaffordability of prescription medications by at least 40%. Patients with PsA had the highest adjusted OR for medication unaffordability (adjusted OR: 1.81, 95% CI: 1.30-2.47), followed by those with two or more types of SARDs (adjusted OR: 1.79, 95% CI: 1.50-2.13). Having two or more types of SARDs was also most strongly associated with cost-related medication nonadherence (adjusted OR: 1.62, 95% CI: 1.36-1.93), followed by Sjögren’s disease (adjusted OR: 1.57, 95% CI: 1.18-2.05). PsA was the SARD most strongly associated with reporting cost-reducing strategies in our fully adjusted models (adjusted OR: 1.68, 95% CI: 1.30-2.15), but was less strongly associated with cost-related non-adherence (adjusted OR: 1.39, 95% CI: 0.99-1.91). Full adjustment reduced the odds of all three cost-related medication behaviors the most for those with SLE: after full adjustment, the odds ratio for unaffordability of medication, cost-related medication non-adherence, and cost-reducing strategies decreased from 2.11 to 1.45, from 1.67 to 1.29, and from 1.37 to 1.28, respectively, for those with SLE.
Table 3.
Unadjusted and adjusted odds ratio (95% CI) for cost-related medication behaviors in patients with systemic autoimmune rheumatic diseases (SARDs) compared to patients without SARDs in All of Us Research Program (data release version 7).
| (a) Unaffordability of prescription medicines | ||||
|---|---|---|---|---|
| Unadjusted Odds ratio | Adjusted Odds Ratio | |||
| Minimally-adjusteda | Demographics-adjustedb | Fully-adjustedc | ||
| Any SARDs | 1.65 (1.52-1.79) | 1.56 (1.43-1.69) | 1.64 (1.50-1.79) | 1.56 (1.43-1.70) |
| Rheumatoid arthritis | 1.44 (1.24-1.66) | 1.36 (1.17-1.57) | 1.54 (1.32-1.79) | 1.43 (1.22-1.66) |
| Systemic lupus erythematosus | 2.11 (1.68-2.61) | 2.01 (1.61-2.49) | 1.50 (1.19-1.87) | 1.45 (1.15-1.82) |
| Sjögren’s disease | 1.41 (1.06-1.85) | 1.36 (1.02-1.78) | 1.57 (1.17-2.07) | 1.52 (1.12-2.01) |
| Psoriatic arthritis | 1.59 (1.16-2.14) | 1.52 (1.11-2.05) | 1.86 (1.35-2.53) | 1.81 (1.30-2.47) |
| Ankylosing spondylitis | 1.61 (1.17-2.17) | 1.55 (1.13-2.09) | 1.65 (1.19-2.24) | 1.58 (1.14-2.17) |
| Other SARDs | 1.34 (1.03-1.73) | 1.27 (0.97-1.64) | 1.51 (1.15-1.96) | 1.54 (1.17-2.01) |
| 2 or more SARDs | 2.09 (1.77-2.46) | 1.92 (1.62-2.27) | 1.92 (1.61-2.27) | 1.79 (1.50-2.13) |
| (b) Cost-related medication nonadherenced | ||||
| Unadjusted Odds Ratio | Adjusted Odds Ratio | |||
| Minimally-adjusteda | Demographics-adjustedb | Fully-adjustedc | ||
| Any SARDs | 1.47 (1.35-1.60) | 1.41 (1.30-1.53) | 1.47 (1.35-1.60) | 1.43 (1.31-1.56) |
| Rheumatoid arthritis | 1.33 (1.15-1.54) | 1.27 (1.10-1.47) | 1.43 (1.23-1.65) | 1.37 (1.18-1.59) |
| Systemic lupus erythematosus | 1.67 (1.32-2.09) | 1.61 (1.28-2.01) | 1.28 (1.01-1.61) | 1.29 (1.01-1.62) |
| Sjögren’s disease | 1.48 (1.13-1.92) | 1.44 (1.10-1.86) | 1.61 (1.22-2.10) | 1.57 (1.18-2.05) |
| Psoriatic arthritis | 1.30 (0.93-1.76) | 1.25 (0.90-1.70) | 1.44 (1.03-1.97) | 1.39 (0.99-1.91) |
| Ankylosing spondylitis | 1.45 (1.06-1.95) | 1.41 (1.03-1.90) | 1.48 (1.07-2.00) | 1.44 (1.04-1.97) |
| Other SARDs | 1.21 (0.93-1.56) | 1.16 (0.89-1.50) | 1.31 (1.00-1.69) | 1.36 (1.03-1.76) |
| 2 or more SARDs | 1.82 (1.53-2.14) | 1.71 (1.44-2.01) | 1.68 (1.41-1.99) | 1.62 (1.36-1.93) |
| (c) Cost-reducing strategies for prescription medicatione | ||||
| Unadjusted Odds ratio | Adjusted Odds Ratio | |||
| Minimally-adjusteda | Demographics-adjustedb | Fully-adjustedc | ||
| Any SARDs | 1.27 (1.18-1.36) | 1.24 (1.16-1.33) | 1.25 (1.16-1.34) | 1.23 (1.14-1.32) |
| Rheumatoid arthritis | 1.10 (0.97-1.24) | 1.07 (0.94-1.21) | 1.09 (0.96-1.24) | 1.07 (0.95-1.22) |
| Systemic lupus erythematosus | 1.37 (1.12-1.67) | 1.35 (1.10-1.64) | 1.29 (1.05-1.57) | 1.28 (1.05-1.57) |
| Sjögren’s disease | 1.59 (1.28-1.97) | 1.57 (1.26-1.94) | 1.55 (1.24-1.92) | 1.51 (1.21-1.87) |
| Psoriatic arthritis | 1.68 (1.31-2.14) | 1.65 (1.29-2.10) | 1.71 (1.33-2.19) | 1.68 (1.30-2.15) |
| Ankylosing spondylitis | 0.96 (0.72-1.26) | 0.94 (0.71-1.24) | 0.94 (0.71-1.24) | 0.92 (0.69-1.22) |
| Other SARDs | 1.19 (0.96-1.47) | 1.17 (0.94-1.44) | 1.19 (0.96-1.47) | 1.21 (0.98-1.50) |
| 2 or more SARDs | 1.41 (1.21-1.63) | 1.36 (1.17-1.57) | 1.34 (1.15-1.56) | 1.31 (1.13-1.53) |
SARD: Systemic autoimmune rheumatic disease.
Minimally-adjusted model adjusted for number of medications.
Demographics-adjusted model additionally adjusted for age, gender, race and ethnicity, and geographical region (based on 4 U.S. census regions).
Fully-adjusted model III additionally adjusted for annual household income, education level, insurance type, employment status, area deprivation index, and Charlson Comorbidity Index.
Cost-related medication nonadherence included skipping medication doses, taking less medicine, and delayed filling a prescription to save money.
Cost-reducing strategies for prescription medication included asking for a lower cost medication, buying prescription drugs from another country, and using alternative therapies to save money.
Results of a sensitivity analysis in which patients with 2 or more SARDs were categorized into the group with the highest number of codes (ICD-9, ICD-10 or SNOMED) are shown in Supplementary Figure 1 and Supplementary Table 3. As in the main analysis, patients with PsA had the highest adjusted OR for medication unaffordability (adjusted OR: 1.87, 95% CI: 1.40-2.46) after adjustment. For cost-related medication nonadherence, patients with Sjögren’s disease showed the highest adjusted OR (adjusted OR: 1.49, 95% CI: 1.15-1.90), whereas those with PsA had the highest adjusted OR for cost-reducing strategies (adjusted OR: 1.65, 95% CI: 1.32-2.06).
Figure 3 displays associations between cost-related medication behaviors and prescription rates for DMARDs and glucocorticoids and health care utilization among patients with SARDs. The odds of receiving DMARDs were similar or slightly lower among patients with SARDs who reported each of the cost-related medication behaviors compared to those who did not. However, the odds of receiving glucocorticoids was significantly higher among patients reporting each of the cost-related medication behaviors than those who denied these. After adjusting for number of medications, age, gender, race and ethnicity, CCI, and type of SARD, compared with reporting affordability of prescription medicines, unaffordability was significantly associated with 0.83 (95% CI: 0.70-0.99) times decreased odds of DMARDs prescription, but marginally associated with 1.18 (95% CI: 0.99-1.42) times increased odds of glucocorticoid prescription. In terms of health care utilization, annual outpatient visits were not significantly different between patients reporting or not reporting each type of cost-related medication behavior. However, after adjustment, compared with reporting affordability of prescription medicines, patients reporting medication unaffordability were more likely to have 1.27 (95% CI: 1.03-1.57) times increased odds of having 1-2 emergency department visits per year and 1.38 (95% CI: 0.96-1.99) times increased odds of ≥3 emergency department visits per year. Experiencing cost-related medication nonadherence was marginally significantly associated with a 1.44 (95% CI: 0.99-2.08) times increased odds of having ≥3 emergency department visits per year, whereas experiencing cost-reducing strategies was significantly associated with a 1.63 (95% CI: 1.26-2.11) times increased odds of having ≥3 inpatient visits per year.
Figure 3. Associations between cost-related medication behaviors and baseline medication prescriptions and health care utilization in patients with systemic autoimmune rheumatic diseases (SARDs) in All of Us Research Program (data release version 7).
For prescription of DMARDs (any vs. none) and prescription of glucocorticoids (any vs. none), unadjusted and adjusted logistic regression analyses were performed compared to non-user group. For baseline annual inpatient visit (0 (ref), 1-2, 3+), emergency department visit (0 (ref), 1-2, 3+), and outpatient visit (0 (ref), 1-9, 10+), unadjusted and adjusted multinomial logistic regression analyses were performed. Adjusting factors included number of medications, age group, gender, race and ethnicity, insurance type, Charlson’s Comorbidity Index and types of systemic autoimmune rheumatic diseases.
Discussion
In this large, diverse cross-sectional study, the prevalence of cost-related medication behaviors among patients with SARDs self-reported on three survey domains, unaffordability of prescription medicines, cost-related medication adherence, and cost-reducing strategies, was higher than among those without SARDs. This large nationwide study reveals that approximately 1 in 5 patients with SARDs reported unaffordability of medication and cost-related medication nonadherence. This overall percentage of 18.1% of patients with SARDs reporting cost-related medication non-adherence, higher than among those without SARDs (13.1%) exceeds that previously reported in other disease populations such as hypertension (10.7~12.7%), cardiovascular disease (12.6~15.1%), and diabetes (15.9~16.5%).30–33
Our study aligns with previous findings. According to Harrold et al.18 and Minhas et al.,20 the prevalence of cost-related medication nonadherence among patients with RA and SLE was 18.4% and 21.7%, compared to 16.3% and 22.2% as reported in the current study. Specifically, in a past study by Minhas et al.,20 13.4% of patients with SLE skipped medication doses to save money, 15.2% took less medication to save money, and 16.2% delated filling a prescription to save money. Similarly, in our study, we found that 13.0% of patients with SLE skipped medication doses to save money, 12.0% took less medication to save money, and 16.7% delated filling a prescription to save money.
The high prevalence of cost-related medication problems experienced by this population may be explained by both costly medications and the demographics of the populations with SARDs. Several past studies have found a large economic burden of medications in patients with SARDs. A recent systematic review of the economic burden of RA found that the main component of the direct cost to patients with RA was that of their medications, which accounted for up to 87.2% of direct costs.34 According to Clarke et al., outpatient pharmacy cost was the third largest driver of high healthcare costs among SLE patients.35 Similarly, Walsh et al. showed that patients with ankylosing spondylitis spent 8.1 times more on total outpatient pharmacy costs than the matched controls.36
Among different types of SARDs, PsA had the highest adjusted ORs for self-reported cost-related medication problems, particularly for unaffordability of medication and use of cost-reducing strategies, such as requesting a lower cost medication or switching medications. This may point to the high costs of biologic DMARDs for PsA, including newer biologics such as those targeting interleukin-17 and interleukins 12 and 23, median point-of-sale price per fill of which is up to $23,417 in the U.S. in 2020.37,38 In our study, approximately half of patients (45.4%) with PsA were receiving a prescription of a biologic DMARD, the highest proportion of those with SARDs. According to a past study investigating Medicare plan coverage for PsA therapies, coverage for biologic DMARDs varied by drug, ranging from 10.0% to 99.8%, and only 2.4% to 5.5% of plans provided co-pay assistance plans for their medications.38 Low coverage and use of co-pay assistance programs from pharmaceutical companies for some of these drugs may explain the high odds of reporting unaffordability and use of cost-reducing strategies, but non-significantly elevated odds of medication nonadherence among PsA patients.
We found the largest effect of adjustment for sociodemographic differences among those patients with SLE, who reported the most medication nonadherence and unaffordability of prescription medication. However, after sequentially adjusting for number of medications, demographic factors, socioeconomic factors, and CCI, the ORs for cost-related medication nonadherence decreased by 22.7% and for unaffordability of medications by 32.3%. The younger, more non-White, and less insured populations most affected by SLE may thus explain the high prevalence of medication unaffordability and nonadherence among these patients, a very important consideration in their clinical care.39,40
Cost-related medication behaviors can affect patient outcomes. Recently, two studies have demonstrated significant associations between cost-related medication nonadherence and patient-reported outcomes among patients with SLE.41,42 Patients with cost-related medication nonadherence or using cost-reducing strategies had worse patient-reported outcomes, not only for SLE disease activity and damage, but also for depressive symptoms and health-related quality of life. Further studies are needed to determine how cost-related medication nonadherence influences clinical outcomes for those with other SARD diagnoses and to develop strategies to mitigate cost-related nonadherence for patients with SARDs.
Importantly, our cross-sectional analysis found that patients with SARDs who reported unaffordability of their medications were 16.5% less likely to be prescribed DMARDs and 18.1% more likely to receive glucocorticoids. Although DMARDs are standard treatment of SARDs, prescription drug unaffordability appears to be a barrier to DMARD use. Glucocorticoids rapidly alleviate symptoms, but use should be as limited as much as possible, due to their many negative effects on patient outcomes, including increased risks of cardiovascular disease, infections, and osteoporosis.43,44 However, as glucocorticoids are relatively inexpensive and readily accessible treatments, they appear to be overused for those patients with medication cost barriers. In addition, we observed that patients with SARDs who experienced unaffordability of their medications had similar outpatient visits, but more frequent inpatient emergency department visits than those who did not. This finding is in line with previous studies which have observed that problems in medication adherence may increase acute health care utilization in chronic diseases, including SLE.13,45,46 Our cross-sectional finding should be pursued in a longitudinal cohort study.
We acknowledge several limitations of this study. First, participants who did not answer the Health Care Utilization survey with questions on cost-related medication behaviors were excluded in the study, which may affect the results. The All of Us Research Program has aimed to include underrepresented people in biomedical research, with 45% racial and ethnic minorities in the version 7 release studied here. However, 56.6% of enrolled participants did not answer the Health Care Utilization survey on cost-related medication behaviors and thus were not included in our study. In the current study, 71.3% of participants were non-Hispanic White Americans, and those who were older, non-Hispanic White, female, and who had higher education levels, were overrepresented with higher survey response rates. This trend may have led to underrepresentation of other sociodemographic groups and influenced results. It is likely in this case that our results are really an underestimate of medication cost-related non-adherence among SARDs patients. Second, cost-related nonadherence rates among SARD patients may in fact be underestimated due to low response rates and uneven distribution of respondents and non-respondents, and the possibility of social desirability bias on self-reported questionnaires.48 Third, as we identified patients by diagnosis codes, there may be misclassification errors, although these diagnostic algorithms have been developed, validated, and are widely used in EHR datasets.23,24 Similarly, drug prescriptions may be underestimated due to possible incomplete EHR records. Fourth, SARD and non-SARD patients were not matched, but adjusted for possible confounders in the analysis instead because we were specifically interested in how they affected cost-related nonadherence behaviors. There may be possibilities that more comorbidities in SARD patients could affect their cost-related behaviors. Further studies that match comorbid conditions are needed to validate our findings in more balanced cohorts. Lastly, we did not measure actual pill counts for adherence to anti-rheumatic medications nor perform cost analyses. Lastly, due to the cross-sectional nature of the study, we could not determine causality. Despite these limitations, to our knowledge, this is the first large US survey study to investigate the prevalence of self-reported cost-related medication behaviors among patients with different types of SARDs, and compared to patients followed for other diverse clinical indications. This study includes a nationally representative and diverse patient population with a large sample size, making the findings generalizable.
In summary, patients with SARDs reported 1.56 times increased odds of unaffordability of prescription medicines, 1.43 times increased odds of cost-related medication nonadherence, and 1.23 times increased odds of cost-reducing strategies than did patients without SARDs. Among those with SARDs, we observed variation in these measures, with the highest prevalence of nonadherence among those with SLE and the highest prevalence of using cost-reducing strategies among those with PsA. Ongoing research is investigating how cost-related medication behaviors influence medication adherence and clinical outcomes for patients with SARDs. Clinicians and other stakeholders should take the financial burden of medications faced by patients with SARDs into account to enhance medication adherence and clinical outcomes.
Supplementary Material
Significance and Innovations.
This large cross-sectional survey within the All of Us Research Program found that patients with systemic autoimmune rheumatic diseases (SARDs) reported more cost-related medication behaviors than did those without SARDs.
When compared by type of SARDs, patients with systemic lupus erythematosus reported the highest cost-related medication nonadherence, whereas those with psoriatic arthritis reported using the most cost-reducing strategies.
Among patients with SARDs, unaffordability of prescription medicines was more likely to have a 1.27 (95%CI 1.03-1.57) and 1.38 (95%CI 0.96-1.98) times increased odds of having 1-2 and ≥3/year emergency room visits in the same year.
Acknowledgement:
The All of Us Research Program is supported by the National Institutes of Health, Office of the Director: Regional Medical Centers: 1 OT2 OD026549; 1 OT2 OD026554; 1 OT2 OD026557; 1 OT2 OD026556; 1 OT2 OD026550; 1 OT2 OD026552; 1 OT2 OD026553; 1 OT2 OD026548; 1 OT2 OD026551; 1 OT2 OD026555; IAA #: AOD 16037; Federally Qualified Health Centers: HHSN 263201600085U; Data and Research Center: 5 U2C OD023196; Biobank: 1 U24 OD023121; The Participant Center: U24 OD023176; Participant Technology Systems Center: 1 U24 OD023163; Communications and Engagement: 3 OT2 OD023205; 3 OT2 OD023206; and Community Partners: 1 OT2 OD025277; 3 OT2 OD025315; 1 OT2 OD025337; 1 OT2 OD025276. In addition, the All of Us Research Program would not be possible without the partnership of its participants.
Funding:
This research was funded by NIH K24AR066109 and R03AR083661.
Reference
- 1.McGonagle D, McDermott MF. A proposed classification of the immunological diseases. PLoS Med. 2006;3:e297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Goldblatt F, O’Neill SG. Clinical aspects of autoimmune rheumatic diseases. Lancet. 2013;382:797–808. [DOI] [PubMed] [Google Scholar]
- 3.Sarzi-Puttini P, Ceribelli A, Marotto D, et al. Systemic rheumatic diseases: From biological agents to small molecules. Autoimmun Rev. 2019. Jun;18(6):583–592. [DOI] [PubMed] [Google Scholar]
- 4.Hresko A, Lin TC, Solomon DH. Medical care costs associated with rheumatoid arthritis in the US: A systematic literature review and meta-analysis. Arthritis Care Res (Hoboken). 2018;70:1431–1438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.D’Angiolella LS, Cortesi PA, Lafranconi A, et al. Cost and cost effectiveness of treatments for psoriatic arthritis: A systematic literature review. Pharmacoeconomics. 2018;36:567–589. [DOI] [PubMed] [Google Scholar]
- 6.Bergstra SA, Branco JC, Vega-Morales D, et al. Inequity in access to bDMARD care and how it influences disease outcomes across countries worldwide: results from the METEOR-registry. Ann Rheum Dis. 2018;77:1413–1420. [DOI] [PubMed] [Google Scholar]
- 7.Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353:487–97. [DOI] [PubMed] [Google Scholar]
- 8.Harrold LR, Andrade SE. Medication adherence of patients with selected rheumatic conditions: a systematic review of the literature. Semin Arthritis Rheum. 2009;38:396–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vangeli E, Bakhshi S, Baker A, et al. A systematic review of factors associated with non-adherence to treatment for immune-mediated inflammatory diseases. Adv Ther. 2015;32:983–1028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Anghel LA, Farcaş AM, Oprean RN. Medication adherence and persistence in patients with autoimmune rheumatic diseases: a narrative review. Patient Prefer Adherence. 2018;12:1151–1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Murage MJ, Tongbram V, Feldman SR, et al. Medication adherence and persistence in patients with rheumatoid arthritis, psoriasis, and psoriatic arthritis: a systematic literature review. Patient Prefer Adherence. 2018;12:1483–1503. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li L, Cui Y, Yin R, et al. Medication adherence has an impact on disease activity in rheumatoid arthritis: a systematic review and meta-analysis. Patient Prefer Adherence. 2017;11:1343–1356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Feldman CH, Yazdany J, Guan H, et al. Medication nonadherence is associated with increased subsequent acute care utilization among medicaid beneficiaries with systemic lupus erythematosus. Arthritis Care Res (Hoboken). 2015;67:1712–1721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Chambers SA, Rahman A, Isenberg DA. Treatment adherence and clinical outcome in systemic lupus erythematosus. Rheumatology (Oxford). 2007;46:895–898. [DOI] [PubMed] [Google Scholar]
- 15.De Vera MA, Mailman J, Galo JS. Economics of non-adherence to biologic therapies in rheumatoid arthritis. Curr Rheumatol Rep. 2014;16:460. [DOI] [PubMed] [Google Scholar]
- 16.Goh H, Kwan YH, Seah Y, et al. A systematic review of the barriers affecting medication adherence in patients with rheumatic diseases. Rheumatol Int. 2017;37:1619–1628. [DOI] [PubMed] [Google Scholar]
- 17.de Achaval S, Suarez-Almazor ME. Treatment adherence to disease-modifying antirheumatic drugs in patients with rheumatoid arthritis and systemic lupus erythematosus. Int J Clin Rheumtol. 2010;5:313–326. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Harrold LR, Briesacher BA, Peterson D, et al. Cost-related medication nonadherence in older patients with rheumatoid arthritis. J Rheumatol. 2013;40:137–143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Heidari P, Cross W, Weller C, et al. Medication adherence and cost-related medication non-adherence in patients with rheumatoid arthritis: A cross-sectional study. Int J Rheum Dis. 2019;22:555–566. [DOI] [PubMed] [Google Scholar]
- 20.Minhas D, Marder W, Harlow S, et al. Access and cost-related nonadherence to prescription medications among lupus patients and controls: The Michigan Lupus Epidemiology and Surveillance Program. Arthritis Care Res (Hoboken). 2021;73:1561–1567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.All of Us Research Program Investigators; Denny JC, Rutter JL, Goldstein DB, et al. The ‘All of Us’ Research Program. N Engl J Med. 2019;381:668–676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.National Institutes of Health All of Us Research Program Investigators. All of Us Research Program Operational Protocol. https://allofus.nih.gov/sites/default/files/All%20of%20Us%20Research%20Program%20Operational%20Protocol%202022.pdf
- 23.Bernatsky S, Linehan T, Hanly JG. The accuracy of administrative data diagnoses of systemic autoimmune rheumatic diseases. J Rheumatol. 2011;38:1612–1616. [DOI] [PubMed] [Google Scholar]
- 24.D’Silva KM, Jorge A, Cohen A, et al. COVID-19 Outcomes in patients with systemic autoimmune rheumatic diseases compared to the general population: A US multicenter, comparative cohort study. Arthritis Rheumatol. 2021;73:914–920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.National Institutes of Health All of Us Research Program Investigators. Survey explorer. https://www.researchallofus.org/data-tools/survey-explorer/
- 26.National Center for Health Statistics. National Health Interview Survey. https://www.cdc.gov/nchs/nhis/index.htm
- 27.Brokamp C, Beck AF, Goyal NK, et al. Material community deprivation and hospital utilization during the first year of life: an urban population-based cohort study. Ann Epidemiol. 2019;30:37–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care. 2005;43:1130–9. [DOI] [PubMed] [Google Scholar]
- 29.Masnoon N, Shakib S, Kalisch-Ellett L, et al. BMC Geriatr. 2017;17:230. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Van Alsten SC, Harris JK. Cost-related nonadherence and mortality in patients with chronic disease: a multiyear investigation, National Health Interview Survey, 2000-2014. Prev Chronic Dis. 2020;17:E151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Fang J, Chang T, Wang G, et al. Association between cost-related medication nonadherence and hypertension management among US adults. Am J Hypertens. 2020;33:879–886. [DOI] [PubMed] [Google Scholar]
- 32.Khera R, Valero-Elizondo J, Das SR, et al. Cost-related medication nonadherence in adults with atherosclerotic cardiovascular disease in the United States, 2013 to 2017. Circulation. 2019;140:2067–2075. [DOI] [PubMed] [Google Scholar]
- 33.Kang H, Lobo JM, Kim S, et al. Cost-related medication non-adherence among U.S. adults with diabetes. Diabetes Res Clin Pract. 2018;143:24–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hsieh PH, Wu O, Geue C, et al. Economic burden of rheumatoid arthritis: a systematic review of literature in biologic era. Ann Rheum Dis. 2020;79:771–777. [DOI] [PubMed] [Google Scholar]
- 35.Clarke AE, Yazdany J, Kabadi SM, et al. The economic burden of systemic lupus erythematosus in commercially- and medicaid-insured populations in the United States. Semin Arthritis Rheum. 2020;50:759–768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Walsh JA, Song X, Kim G, et al. Healthcare utilization and direct costs in patients with ankylosing spondylitis using a large US administrative claims database. Rheumatol Ther. 2018;5:463–474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ayan G, Ribeiro A, Macit B, et al. Pharmacologic treatment strategies in psoriatic arthritis. Clin Ther. 2023;45:826–840. [DOI] [PubMed] [Google Scholar]
- 38.Pourali SP, Nshuti L, Dusetzina SB. Out-of-pocket costs of specialty medications for psoriasis and psoriatic arthritis treatment in the medicare population. JAMA Dermatol. 2021;157:1239–1241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Buie J, McMillan E, Kirby J, et al. Disparities in lupus and the role of social determinants of health: current state of knowledge and directions for future research. ACR Open Rheumatol. 2023;5:454–464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Williams JN, Drenkard C, Lim SS. The impact of social determinants of health on the presentation, management and outcomes of systemic lupus erythematosus. Rheumatology (Oxford). 2023;62:i10–i14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Minhas D, Marder W, Hassett AL, et al. Cost-related prescription non-adherence and patient-reported outcomes in systemic lupus erythematosus: The Michigan Lupus Epidemiology & Surveillance program. Lupus. 2023;32:1075–1083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Aguirre A, DeQuattro K, Shiboski S, et al. Medication cost concerns and disparities in patient-reported outcomes among a multiethnic cohort of patients with systemic lupus erythematosus. J Rheumatol. 2023;50:1302–1309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Fraenkel L, Bathon JM, England BR, et al. 2021 American College of Rheumatology guideline for the treatment of rheumatoid arthritis. Arthritis Care Res (Hoboken). 2021;73:924–939. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.W J Bijlsma J, Buttgereit F. Adverse events of glucocorticoids during treatment of rheumatoid arthritis: lessons from cohort and registry studies. Rheumatology (Oxford). 2016;55:ii3–ii5. [DOI] [PubMed] [Google Scholar]
- 45.Sokol MC, McGuigan KA, Verbrugge RR, et al. Impact of medication adherence on hospitalization risk and healthcare cost. Med Care. 2005;43:521–530. [DOI] [PubMed] [Google Scholar]
- 46.Ho PM, Rumsfeld JS, Masoudi FA, et al. Effect of medication nonadherence on hospitalization and mortality among patients with diabetes mellitus. Arch Intern Med. 2006;166:1836–1841. [DOI] [PubMed] [Google Scholar]
- 47.Julian LJ, Yelin E, Yazdany J, et al. Depression, medication adherence, and service utilization in systemic lupus erythematosus. Arthritis Rheum. 2009;61:240–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Althubaiti A Information bias in health research: definition, pitfalls, and adjustment methods. J Multidiscip Healthc. 2016;9:211–217. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.