This presentation is part of: Poster Session - CEA: Methods and Applications; Health Services Research
RACIAL DISPARITIES IN THE PRESCRIBING OF WARFARIN FOR NONVALVULAR ATRIAL FIBRILLATION
Daniel P. Schauer, MD1, Joseph A. Johnston, MD, MSc2, Mark Wess, MD1, Charles Moomaw, PhD1, and Mark Eckman, MD, MS1. (1) University of Cincinnati, Internal Medicine, Cincinnati, OH, (2) Eli Lilly, US Outcomes Research, Indianapolis, IN
PURPOSE: Warfarin has been shown to decrease the rate of thromboembolic events in patients with nonvalvular atrial fibrillation (AF) but is frequently under-prescribed. Our goal was to establish whether there are racial disparities in the prescribing of warfarin for patients with newly incident nonvalvular AF in the Ohio Medicaid population. METHODS: A retrospective cohort of Ohio Medicaid recipients with newly incident nonvalvular AF between January 1, 1998, and May 31, 2002, was identified. Patients were included if they had at least one year of continuous enrollment in Medicaid prior to the diagnosis of atrial fibrillation and two or more International Classification of Diseases, Ninth Revision, Clinical Modification codes for atrial fibrillation. Exclusions included valvular heart disease and warfarin prescriptions prior to the diagnosis of atrial fibrillation. Race was identified from the demographic information in the database, and the analysis was limited to White and African-American patients. Possible confounders included age, sex, history of hypertension, diabetes mellitus, congestive heart failure, renal disease, liver disease, previous stroke, previous bleeding, risk factors for non-adherence and increased risk of falling. Univariate logistic regression was used to evaluate the unadjusted association of potential confounders with warfarin prescribing. To evaluate the independent role of race in warfarin prescribing, we created a multivariable logistic regression model incorporating all predictors significant at p<0.10 in univariate models. RESULTS: 6,283 patients were identified as having newly incident nonvalvular AF, 18.5% of which were African-Americans. In general, African-American patients had a higher rate of comorbid illness, with significantly higher rates of risk factors for both stroke (i.e., hypertension, diabetes and congestive heart failure) and bleeding (i.e., renal disease and prior bleeding). 9.4% of White patients and 7.7% of African-American patients were prescribed warfarin. In the univariate analysis African-Americans had an odds ratio of 0.80 (95% CI 0.63, 1.01) for receiving warfarin when compared to White patients. When controlling for significant confounders in the multivariable logistic regression model, African-Americans had an odds ratio for receiving warfarin of 0.74 (95% CI 0.58, 0.94) when compared to White patients. CONCLUSION: African-American patients in the Ohio Medicaid population between 1998 and 2002 were significantly less likely than White patients to be prescribed warfarin for newly incident nonvalvular AF.
