The 2001 Behavioral Risk Factor Surveillance System (BRFSS) physical activity module and the International Physical Activity Questionnaire (IPAQ) are used in population studies to determine the prevalence of physical activity. The comparability of the prevalence estimates has not been compared in U.S. adults. This study compares the physical activity prevalence estimates from the BRFSS and the IPAQ. A telephone survey was administered to a random sample of 11,211 U.S. adults aged 18-99 yr who were enrolled in the National Physical Activity and Weight Loss Survey. Data were analyzed from 9945 adults who provided complete data on the BRFSS and the IPAQ. Prevalence estimates were computed (1) applying the BRFSS scoring scheme for both questionnaires (2). Kappa statistics were used to compare prevalence estimates generated from the BRFSS and the IPAQ. When scored using the BRFSS protocol, agreement between physical activity categories was fair (kappa = 0.34-0.49). Prevalence estimates were higher on the IPAQ than the BRFSS for the lowest category (inactive) by 0.1-3.9% and for the highest category (meets recommendations) by 0.2-9.7%. When scored using their own scoring, agreement between physical activity categories was lower (kappa = 0.26-0.39). The prevalence estimates on the IPAQ were higher than on the BRFSS for the lowest physical activity category by 0.2-13.3% and for the highest physical activity category by 0-16.4%. Differences in physical activity categories were observed for sex, age, income, education, and body mass index on both questionnaires. Because of differences in the physical activity prevalence estimates, direct comparison of the BRFSS and IPAQ prevalence estimates is not recommended.
