After almost three decades of U.S. surveillance in fruit and vegetable (F&V) intake and obesity, it is important to evaluate their usefulness for monitoring prevention and health promotion efforts in public health. We used U.S. surveillance data to evaluate whether the 16-year trends of F&V intake, measured by the prevalence of eating five or more servings of fruits and vegetables a day (FV5/day) is related to obesity trend as measured by its prevalence in the same period. We also evaluated whether trends in the prevalence of FV5/day by important sociodemographic factors (age, race/ethnicity, etc.) could explain the findings.
A secondary analysis of U.S. adults (≥ 18 years) from the Behavioral Risk Factor Surveillance System (BRFSS) (1994-2009).
We categorized survey subjects for their F&V intake derived from the BRFSS six-question food frequency questionnaire into two groups: < FV5/day vs. ≥ FV5/day. Obesity was defined as BMI ≥ 30. We used logistic regressions to compute predicted prevalence of FV5/day and obesity, and to estimate the odds ratio of FV5/day by obesity and levels of sociodemographic, stratified by year.
Between 1994 and 2009, the prevalence of FV5/day hovered around 25% among U.S. adults, while the obesity prevalence steadily increased from 14.8% to 27.4%. As measured through odds ratio, an inverse association between FV5/day and obesity was only observed in 55+, but not in other age, racial/ethnic or education groups.
Between 1994 and 2009, we could not confirm a decrease in the prevalence of FV5/day associated with an increase in obesity prevalence, except for age 55+ group. Known disparities in FV5/day and obesity across sociodemographic factors persisted over the study period. FV5/day may be an inappropriate measure of total calories derived from eating fruits and vegetables. Its use to measure impact of public health strategies to improve nutrition and prevent obesity may be questionable.