Food Supply Chain Analytics and Sensing Initiative

Searchable Database of Packaged Foods and Beverages

The vast majority of food and beverages consumed by Americans today are sold in packages. Beyond highly processed items, even foods such as beans, grains like flour and oats, and washed salad greens are canned, bagged, or boxed. The FDA’s Nutrition Facts Label requirement mandates that all of these packages display the ingredients and major nutrient contributions of all packaged foods and beverages. Thus, there is a vast array of publicly available but widely dispersed data on what Americans are eating. In this project, we aim create a structured database that consolidates and standardizes such information. From there, we will be able to study the potential health impacts of the ingredients in the food supply more rigorously than has been previously possible, as well as analyze the prevalence of potentially harmful ingredients in the U.S. food supply and compare that prevalence against that of other countries.

The project aims to develop research capabilities to answer a range of research questions, such as:

Quantifying the prevalence of food ingredients in branded food products.
Assessing the nutritional values and potential harm of ingredients in branded food products.
Assessing the exposure of households to food ingredients and their nutritional consumption and potential health implications and understanding the household characteristics that affect exposure and consumption.
Comparing 1-3 above in the US to results from other countries.

A critical enabler to address these research questions is accessible data. Thus, a central part of the current effort is focused on developing large scale searchable data infrastructure on branded food products. Our team is leveraging prior experience creating innovative data pipelines, as well as publicly available data, to develop advanced search capabilities in addition to summarize existing research on health risks caused by exposure to food ingredients.

The deliverables of the proposed project will be made available for public use. We believe that the proposed research effort will enable new and more quantitative (and rigorous) research on many research questions with high relevance to public health policy.

Preliminary Findings

In this report, we used a prototype of our searchable database of the ingredients of packaged foods & beverages to estimate the number and proportion of foods sold in the U.S. containing the 6 petroleum-based color additives that are being phased out by the U.S. Department of Health and Human Services (HHS) and FDA.

CLICK HERE TO READ THE REPORT