Medicaid beneficiaries found poorly adherent to lupus medication

By Ted Bosworth

Adherence to a primary therapy for systemic lupus erythematosus (SLE) among Medicaid beneficiaries is low overall and even lower among some specific groups identified by sociodemographic factors or other variables, according to an analysis of a Medicaid database.

The focus in this study was on adherence to hydroxychloroquine, which the investigators identified as a standard of care for SLE. Of 10,268 adult Medicaid beneficiaries started on this therapy over a 10-year period, only 15% met the study threshold of adherence, defined as taking hydroxychloroquine 80% or more of the time during first 365 days of treatment.

After researchers adjusted for multiple factors, including education attainment, health care resources, and poverty, the odds ratio (OR) of adherence was significantly lower in ZIP codes with a higher percentage of blacks. Conversely, there was a significantly greater likelihood of adherence in areas that had a higher concentration of hospitals (HR. 1.30; 95% CI, 1.07-1.58).

Data generated by Medicaid Analytic eXtract (MAX) during 2000-2010 in the 29 most populated states, which captured more than 85% of all Medicaid beneficiaries, were evaluated. MAX included data on drug dispensing, billing claims, and health care utilization. Only data generated by patients with 12 months of continuous enrollment in Medicaid and for whom complete drug-dispensing data were available were included in this analysis.

A social ecological model was employed to consider environmental factors, such as poverty or limited health care resources, on adherence. Covariates in the analysis, in addition to race, included age, sex, educational attainment, and comorbidities. ZIP code of residence, analyzed with by Ameri-can Community Survey data, which tracks poverty level, race, and education by ZIP codes, was identified as a “contextual covariate.”


In a regression model, lower levels of adherence were associated with younger versus older age, black versus Hispanic or white race, and greater versus fewer emergency department visits. There also was lower adherence among those patients who had diabetes, compared with those who did not, and among those taking antidepressants relative to those who did not.

The authors emphasized that there was a significant dose-response relationship “between ZIP code–level percent black and poor adherence to hydroxychloroquine” even after researchers adjusted for individual race, area poverty, and education. These associations were not seen with analyses by ZIP–code percent white or percent Hispanic.


“We hypothesize that in these areas [ZIP codes with predominantly black population], there may be reduced access to high-quality hospitals, outpatient providers, pharmacy services, and transportation. There may be increased stress and depression from racial discrimination and crime, which may also contribute to nonadherence,” according to the investigators.

Other studies also have demonstrated high rates of nonadherence in SLE patients, particularly among lower-income individuals and those from racial or ethnic minorities, according to the authors. However, they maintained, “this is the first study to examine the role of contextual factors on adherence in a high-risk SLE population.” They further speculated that poor adherence among Medicaid beneficiaries in this study may be relevant to previous evidence that minorities and low-income individuals suffer from the highest burden of adverse outcomes.

“Our findings should pave the way for further work examining the importance of social determinants, including racial residential segregation and racial discrimination, as well as neighborhood-specific [health care] quality, on health behaviors and outcomes in vulnerable populations,” the authors of this study suggested.

Most importantly, the authors indicated that the variables they identified for nonadherence “should be considered in the design of interventions that aim to reduce the racial and ethnic disparities.” Although they cautioned that the findings should be considered “hypothesis generating” only, the data generated “provide an impetus to explore the mechanisms behind the associations uncovered.”

Ted Bosworth is a freelance reporter for MDedge News.