Objectives: To determine the prevalence and predictors of antidiabetic medication use over a 10-year period in a general population of African-American and white community-dwelling elderly.
Setting: Five adjacent counties (one urban and four rural) in the Piedmont area of North Carolina.
Participants: Those aged 65 and older present at the baseline (n=4,136), second (n=3,234), third (n=2,508), and fourth (n=1,633) in-person waves of the Duke Established Populations for Epidemiologic Studies of the Elderly.
Measurements: The use of six discrete categories of antidiabetic medications (insulin, first-generation oral sulfonylureas, second-generation oral sulfonylureas, metformin, oral combination therapy, and insulin combination therapy) was determined. Multivariate analyses, using weighted data adjusted for sampling design, were conducted to assess the association between antidiabetic medication use and race and other sociodemographic, health-status, and access-to-healthcare factors at baseline and 10 years later.
Results: Antidiabetic medications were taken by 21.4% of the population at baseline; this increased to 28.1% at the 10-year follow-up (P<.001). Insulin was the most commonly used drug at baseline (7.9%). The use of second-generation sulfonylureas increased, and use of first-generation sulfonylureas decreased over the 10-year time period. Combination antidiabetic therapy and metformin use was infrequent throughout the study. Multivariate analyses revealed that, at baseline, African Americans were nearly twice as likely (adjusted odds ratio (AOR)=1.93, 95% confidence interval (CI)=1.46–2.54) to receive any antidiabetic medication as their white counterparts. Other significant (P<.05) factors were hypertension (AOR=1.38, 95% CI=1.03–1.84), stroke (AOR=1.98, 95% CI=1.43–2.73), one or more mobility difficulties (AOR=1.29, 95% CI=1.01–1.66), continuity of care (AOR=1.74, 95% CI=1.20–2.54), and multiple doctor visits (1–4 visits, AOR=1.69, 95% CI=1.08–2.65; ≥5 visits, AOR=3.15, 95% CI=1.95–5.07). Being underweight (AOR=0.45, 95% CI=0.30–0.67) and being cognitively impaired (AOR=0.60, 95% CI=0.41–0.87) were factors significantly (P<.05) associated with a decreased risk of antidiabetic medication use. At the 10-year follow-up, similar trends were seen associating these sociodemographic, health-status, and access-to-healthcare factors with antidiabetic medication use.
Conclusion: Antidiabetic medication use is common and increases over time for community-dwelling elderly. Race is significantly associated with antidiabetic medication use, even after controlling for other sociodemographic, health-status, and access-to-healthcare variables.