Effect of racial differences on ability to afford prescription medications

Abstract

Purpose: The association of race with not filling prescription medications because of cost for African-American and white patients 65 years or older was examined.

Methods: African-American and white patients age 65 years or older were recruited from the practices of 48 Alabama primary care physicians participating in the Alabama Nonsteroidal Antiinflammatory Drug Patient Safety Study. All eligible patients were asked questions related to their ability to pay for prescription medications, comorbidities, insurance status, and socioeconomic status. Baseline and follow-up telephone surveys were completed between August 2005 and April 2006. Mediation analysis was conducted to determine whether patients' perceived income inadequacy mediated the association between race and not filling medications using staged logistic regression models and adjusting for age, comorbidities, and traditional markers of socioeconomic position (income, education, and insurance status).

Results: Of 399 participants, 32% were African-American, 74% were women, and 53% had an annual household income of <$15,000. Patients not filling prescription medications were more likely to be African-American (50% versus 25%) and to report inadequate income to meet basic needs (61% versus 17%) (p < 0.001 for both comparisons). After adjusting for all covariates except the mediator, the odds ratio (OR) for African Americans not filling a prescription medication was 2.3 when compared with white patients. Adding the mediator (perceived income inadequacy) to the model reduced the OR to 1.4.

Conclusion: African Americans reported markedly greater difficulty in affording prescription medications than did white patients, even after accounting for income, education, health insurance status, and comorbidities. The inability of African Americans to afford prescription medications may be better predicted by perceived income inadequacy than more traditional measures of socioeconomic status.