(P38-12-26) More Than Just the Mean: A Distributional Approach to Dietary Surveillance

Objectives: To present an analytical framework for visualizing and interpreting cumulative distribution functions (CDFs) of usual dietary intake, and to demonstrate its application to US nutrition surveillance data.

Methods: Up to two non-consecutive 24-hour dietary recalls were obtained from the National Health and Nutrition Examination Survey (NHANES) 2011-2018. The NCI method estimated usual intake CDFs for six food groups (total vegetables, total fruit, total protein foods, total dairy, whole grains, refined grains), stratified by age and sex. CDFs were plotted with Dvoretzky-Kiefer-Wolfowitz 95% confidence bands and overlaid with recommended intake ranges from the Dietary Guidelines for Americans, 2020-2025.

Results: We present an analytical and visualization approach for applying CDFs to nutrition surveillance data. In the NHANES case study, we demonstrate that translating adherence into public health action requires considering where on the distribution the percent meeting the recommendation falls, how the distributional shape informs the scale of improvement needed, and how meaningful shifts in consumption may go undetected by conventional metrics alone. For example, 11% of 9–13-year-old males met recommendations for both refined grains and total fruit, yet these identical percentages fell at the lower tail of an approximately normal distribution for refined grains versus the upper tail of a right-skewed distribution for total fruit, each necessitating a different public health response. This example illustrates the value of examining full CDFs, an approach that extends to all six food groups and age-sex categories.

Conclusions: CDFs of usual intake data provide richer evidence for public health nutrition policy and decision-making than conventional summary statistics. This framework enables researchers to characterize consumption patterns, identify intervention priorities, and assess the magnitude of dietary change needed. This approach can be applied to other dietary components, population subgroups, or surveillance contexts.

Funding Sources: Cancer Prevention Fellowship Program, National Cancer Institute