A New Era for the Total Diet Study

FDA's new Total Diet Study Interface eases access to nutrient and contaminant data, supporting scientifically grounded dietary exposure assessments

The U.S. Food and Drug Administration recently released the Total Diet Study Interface (TDSi), an interactive web-based tool for accessing findings from FDA's Total Diet Study (TDS), alongside updated testing results from 2021-2022.

For decades, FDA's Total Diet Study (TDS) has served as one of the agency's primary surveillance programs for monitoring dietary exposure to nutrients and contaminants. Established in 1961, the program tracks substances such as elements (e.g., lead, cadmium, arsenic), pesticide residues, and radionuclides. FDA uses this data to monitor the food supply, estimate dietary exposure to substances, and identify foods that require more surveillance.

The introduction of the TDSi has made contaminant monitoring data more accessible and transparent than ever before, supporting broader awareness and understanding of the Total Diet Study and its findings. Easier public access creates opportunities for regulators, advocacy groups, researchers, and consumers to engage more deeply with data on nutrients and contaminants in foods. The TDSi can be used to explore both snapshots in time as well as long-term nutrient and contaminant level trends. 

Expanded access, however, also increases the risk that representative surveillance data may be mistaken for product-specific findings or used to draw conclusions about dietary exposure without the necessary scientific context. Using these data appropriately requires careful consideration of the TDS design, sampling approach, and intended applications.

What makes the new interface different?

The underlying TDS data is not entirely new; FDA has periodically updated these datasets for years. What is new is the Total Diet Study Interface (TDSi), a web-based tool that allows users to explore foods, analytes, and contaminant trends through detailed analytics, interactive visualizations, and downloadable outputs. Previously, users were required to work directly with static Excel datasets and perform their own statistical analysis. 

Released alongside updated 2021-2022 sample results, the TDSi now includes data from 2018-2022.

How are TDS foods sampled? 

The TDS uses a "market basket" approach by collecting foods intended to reflect what U.S. consumers commonly purchase and consume. As a result, the repository is designed to characterize broad population-level exposure trends rather than evaluate individual brands or products, providing a foundation for monitoring nutrients and contaminants in the national food supply.

Sample collection 

The TDS collects data on foods available in the U.S. marketplace, selected in part based on market share data to represent widely consumed products. "National" foods are defined as foods and beverages with nutrient or contaminant concentrations that are less likely to vary by location or season, while "regional" foods may vary geographically or seasonally. National food samples are obtained online or through retail outlets, while regional samples are collected from retail outlets across multiple regions, cities, and seasons throughout the year.

Sample preparation 

Many of the foods are prepared and cooked in the same way consumers would at home. For some national foods, products of different flavors or varieties are grouped to represent 50% market share in a single analytical composite. For regional foods, samples from three cities within a region are merged for analysis. The results of multiple composites are combined to estimate the average levels of nutrients and contaminants in foods consumed by the U.S. population. 

Understanding the three non-detect methodologies

As part of the TDSi, FDA provides analytical results by directly applying three different substitution methodologies for handling results below the detection limit, otherwise known as "non-detects." The three methodologies substitute non-detects based on lower bound, hybrid, and upper bound methods: 

• The lower bound sets all non-detect concentrations to zero. 
• The hybrid method sets non-detect concentrations to zero for foods with no history of detection of the substance in prior TDS data and substitutes non-detect concentrations with one-half of the reporting limit for foods with one or more detected values in prior TDS data. 
• The upper bound sets non-detect concentrations to the reporting limit. 

These assumptions can materially affect estimated contaminant concentrations and resulting exposure assessments. The upper bound produces the most conservative estimate, while the lower bound produces the lowest estimate. The hybrid method incorporates conservative assumptions while accounting for foods with no history of detection.

This is the first time FDA has released TDS analytical data with pre-processed substitution methods for non-detects. The interface improves transparency by allowing users to compare results across the three substitution approaches, highlighting the rationale for each approach and its potential influence on analytical results and subsequent exposure assessments. 

"As access to TDS data expands, it becomes increasingly important to understand what the data are designed to show — and what they are not."

What role does TDS play in dietary exposure assessments?

To conduct dietary exposure assessments, regulators, researchers, and food producers and manufacturers often combine nutrient and contaminant data from the TDS with nationally representative food consumption data collected through dietary surveys, such as the What We Eat in America (WWEIA) component of the National Health and Nutrition Examination Survey (NHANES). This combination enables estimates of dietary exposure across the U.S. population and helps inform regulatory evaluations, research, and food safety decisions.

FDA's Food Disaggregation Database is one way to streamline this process by mapping foods and ingredients reported in WWEIA/NHANES to corresponding TDS data. However, analytical results from the TDS alone are not dietary assessments. Exposure estimates depend on how TDS concentration data are combined with consumption data, how foods are mapped, how non-detects are handled, and how representative foods are selected.

For industry, that distinction matters. Food producers, manufacturers, and trade associations may use TDS data in a variety of ways including to contextualize internal testing results, compare findings across food categories, or evaluate broader marketplace trends.

Applications for industry and public health 

Although the TDSi makes contaminant monitoring data easier to access, meaningful interpretation depends on understanding how the TDS data are collected and what they are intended to represent.  Further, in the case of dietary exposure assessments, it is also important to understand how contaminant concentrations are combined with NHANES food consumption data. These exposure estimates can be influenced by food mapping decisions, treatment of non-detects, and other analytical assumptions. Food producers and manufacturers may find the data useful to guide their sourcing evaluations, ingredient selection, or sampling strategies for foods known to accumulate certain contaminants.

Because the TDS samples representative foods rather than every food consumed in the U.S., not all foods reported in NHANES have a direct counterpart in the TDS database. FDA therefore applies scientifically supported mapping approaches which can be used to estimate dietary exposure, making familiarity with these assumptions important when interpreting results. 

A more accessible but more complex data landscape

As access to TDS data expands, it becomes increasingly important to understand what the data are designed to show — and what they are not. The TDS characterizes contaminant occurrence and dietary exposure across representative food categories rather than individual brands or products. Using TDSi results to inform research, regulatory decision-making, supply chain evaluations, and risk communication requires expertise in food science, nutrition, and dietary exposure assessment, along with a firm grasp of the strengths, limitations, and intended applications of the underlying datasets.