Abstract
PURPOSE: Medications can have unintended effects. High medication use populations may benefit from increased regimen oversight. Limited knowledge exists concerning racial and regional polypharmacy variation. We estimated total medication distributions (excluding supplements) of American black and white adults and assessed racial and regional polypharmacy variation. METHODS: REasons for Geographic And Racial Differences in Stroke (REGARDS) cohort data (n = 30,239 U.S. blacks and whites aged ≥45 years) were analyzed. Home pill bottle inspections assessed the last two weeks' medications. Polypharmacy (≥8 medications) was determined by summing prescription and/or over-the-counter ingredients. Population-weighted logistic regression assessed polypharmacy's association with census region, race, and sex. RESULTS: The mean ingredient number was 4.12 (standard error = 0.039), with 15.7% of REGARDS using 8 ingredients or more. In crude comparisons, women used more medications than men, and blacks and whites reported similar mean ingredients. A cross-sectional, logistic model adjusting for demographics, socioeconomics, and comorbidities showed increased polypharmacy prevalence in whites versus blacks (OR [95% CI]: 0.63, [0.55-0.72]), women (1.94 [1.68-2.23]), and Southerners (broadly Southeasterners and Texans; 1.48 [1.17-1.87]) versus Northeasterners (broadly New England and upper Mid-Atlantic). Possible limitations include polypharmacy misclassification and model misspecification. CONCLUSION: Polypharmacy is common. Race and geography are associated with polypharmacy variation. Further study of underlying factors explaining these differences is warranted.