Ab0579 association of Race or Ethnicity With Medication Adherence in Systemic Lupus Erythematosus

Abstract

BackgroundSystemic lupus erythematosus (SLE) causes significant disease burden. Racial and ethnic differences in outcomes have been recognized, with increased morbidity and mortality in Black and Hispanic individuals, compared to non-Hispanic White referents. Financial barriers to treatment, such as copayments (or copays), may affect adherence to essential pharmacologic therapies, contributing to increased health adversity. Determining the contribution of copays to adherence may identify disparities in medication access and pathways to improve adherence.ObjectivesTo examine the association between race or ethnicity and medication adherence.MethodsWe used administrative health claims data (Optum’s De-identified Clinformatics® Data Mart Database, 2007-2021) to identify individuals with incident SLE (2 ICD-9/10 codes within 30-365 days) between 2010-19 receiving hydroxychloroquine (HCQ), azathioprine (AZA), mycophenolate mofetil (MMF), methotrexate (MTX), or combination therapy (HCQ with AZA, MMF, or MTX). Race or ethnicity were categorized as Asian, Black, Hispanic, White, or unknown. We dichotomized copays as low (<$10/30 days) or high (≥$10/30 days). Adherence was defined as proportion of days covered (PDC) ≥80%. We examined the association between race or ethnicity and adherence with multivariable-adjusted logistic regression models, using White individuals as referent. Results were then stratified by copay categories.ResultsWe identified 12,510 individuals: age 54.2±15.5 years; 88.2% women; 63.2% (7910) White, 14.9% (1860) Black, and 13.8% (1729) Hispanic. A total of 9510 (76%) were on HCQ and 15% on combination therapy. Median (IQR) 30-day copays were $8 (4, 10) for HCQ, $7 (2, 10) for AZA, $10 (5, 20) for MMF, and $8 (3, 11) for MTX. Black individuals had 51% (95% CI 0.31-0.80), 30% (95% CI 0.61-0.81), and 35% (95% CI 0.45-0.94) lower odds of adhering to AZA, HCQ, and combination therapy, respectively. Hispanic individuals had 27% (95% CI 0.63-0.94) lower odds of adhering to HCQ. When stratified by low or high copay, the association between race or ethnicity and adherence did not change, and the interaction of copayment on the association between race or ethnicity and adherence was not significant (Table 1).ConclusionIn this large cohort, Black and Hispanic individuals had decreased odds of adherence to SLE medications even after adjusting for several socioeconomic factors and comorbidities. The detrimental associations between copay and adherence were consistent across copay strata, suggesting that other factors contribute to racial and ethnic disparities. Factors such as polypharmacy, pharmacy deserts, transportation, mental health, cultural barriers, and health care access merit examination for their relation to adherence.Table 1.Association of Race or Ethnicity and Odds of Medication AdherenceAsianBlackHispanicOR (95% CI); P valueAZA (n=690)0.67 (0.23-1.94);0.460.50 (0.30-0.80);0.0041.00 (0.57-1.75);0.99HCQ (n=8629)1.15 (0.87-1.53);0.330.70 (0.61-0.81); <0.00010.73 (0.63-0.85); <0.0001MMF (n=756)1.53 (0.66-3.53); 0.320.73 (0.50-1.13);0.160.82 (0.52-1.30); 0.40MTX (n=518)0.75 (0.20-3.43); 0.711.11 (0.57-2.15);0.771.36 (0.67-2.76); 0.41Combination Therapy* (n=1720)1.04 (0.53-2.02); 0.910.65 (0.55-0.94);0.020.70 (0.48-1.03); 0.97Reference group: White individuals.Table 2.Stratification by Copayment≥$10 vs <$10 CopayOR (95% CI); P valueAsianBlackHispanicInteraction p valueAZA0.18 (0.02-1.54); 0.180.37 (0.17-0.81); 0.010.43 (0.17-1.11); 0.080.79HCQ0.52 (0.30-0.90); 0.020.72 (0.56-0.93); 0.010.57 (0.44-0.73); <0.00010.51MMF0.68 (0.13-3.44); 0.640.85 (0.42-1.75); 0.660.83 (0.39-1.75); 0.620.78MTX0.41 (0.02-9.51); 0.580.75 (0.25-2.30); 0.622.00 (0.50-8.04); 0.330.67Combination Therapy1.70 (0.45-6.45); 0.431.25 (0.69-2.27); 0.460.61 (0.32-1.16); 0.130.14Adjusted for age, sex, Elixhauser comorbidities, tobacco use, antiphospholipid syndrome, educational attainment, annual household income, insurance type, and US geographic region.AcknowledgementsThis work is supported by K24HL160527.Disclosure of InterestsRaisa Lomanto Silva: None declared, Gretchen M Swabe: None declared, Sebastian E. Sattui Grant/research support from: Bristol Myers Squibb Foundation Robert A. Winn Diversity in Clinical Trials Career Development Award, outside of the submitted work., Jared W Magnani: None declared.