Age, Period and Cohort Effects on Overweight and Obesity in a Nationally Representative Sample of Adults

Eric N. Reither, University of Wisconsin at Madison

Although “epidemic” is a technically correct description of rapidly increasing obesity rates in the U.S., it is premature to conflate the obesity epidemic with secular trends (i.e., period effects), since no study to date has disentangled period effects from age and birth cohort effects. This study estimated the relative importance of age, period and cohort (APC) effects by utilizing data from 1.6 million adults in the National Health Interview Survey, 1976-2002. Both logistic and continuous APC models supported the assumption that period effects are primarily responsible for the obesity epidemic. For instance, net of education, age and cohort, the odds of class II obesity in 2002 were 5.3 times greater than in 1976. Models stratified by race and sex revealed many similarities but some differences; e.g., unlike white male cohorts, white female and all black cohorts born after 1955 exhibited increasing risks of obesity. Implications of these results are discussed.