The Cohort Dynamics of Diabetes in the United States

Ezra Fishman, University of Pennsylvania
Andrew Stokes, University of Pennsylvania
Sam Preston, University of Pennsylvania

Using a representative sample of the civilian noninstitutionalized U.S. population, we estimate trends in diabetes prevalence across cohorts born 1910 to 1989 and provide the first estimates of age-specific diabetes incidence using nationally representative, measured data. Data were from 40,130 individuals aged 20-79 years who participated in the third National Health and Nutrition Examination Survey (NHANES III), 1988-1994, and the continuous 1999-2010 NHANES. We estimate age-specific diabetes prevalence for the five-year age groups 20-24 through 75-79 for cohorts born 1910-1919 through 1980-1989 and periods 1988-1994, 1999-2002, 2003-2006, and 2007-2010. We model diabetes prevalence as a function of age, period, and cohort and use our cohort model to estimate age-specific diabetes incidence. Diabetes prevalence rose with age, for all ages within each cohort, and age-specific diabetes prevalence rose across cohorts. Cohort-based models show a steeper age-pattern of diabetes prevalence than do period-based models. Incidence was highest for ages 55 to 64.