Category Archives: Uncategorized

Can you use NDSR data to assess intake of ultra-processed foods?

NDSR does not classify foods into the four NOVA categories. However, a researcher may carry out this classification for foods entered into NDSR dietary recall, record, and menu record types. There are multiple ways this may be done. Sneed et al. report on one approach using their NDSR data (1). These authors assess inter-rater reliability and note that their data files are available to researchers on request. Below we describe another potential approach.
Potential approach for classifying foods in NDSR dietary recall, record and menu record types into the four NOVA categories.
1) For your set of dietary recalls/records/menus identify all unique food IDs in output file 02 (file that lists foods at the whole food level).
2) Sort the list of unique food IDs by the NCC Database Food Group ID* to facilitate coding.
3) Review the unique food IDs in your dataset, and then develop coding rules that will allow for classifying each food ID into one of the four NOVA categories.
Some of the coding rules you develop may leverage the NCC Database Food Group ID. For example, you may decide to specify in your coding rules that all foods with an NCC Database Food Group ID of 63 (‘Fruits, fresh and unsweetened’) be coded into the NOVA class 1 (‘Unprocessed or minimally processed’) category.
For some types of restaurant and packaged foods you may decide to develop coding rules that require going to the food company website to locate the food’s ingredient statement so that a classification determination may be made. Note that ingredient information is generally not available in output file 01 (component ingredient file) for packaged foods. Ingredient information is generally available for restaurant foods, but it should not be relied on for determining classification because NCC formulations for both packaged and restaurant foods include only those food ingredients that contribute to the nutrient content of the food. As a result, ingredients included in small amounts that do not contribute to nutrient content, such as most food flavorings, colorings, emulsifiers, and preservatives, are generally not included in the formulation.
For some types of restaurant and packaged foods you may develop general rules that do not require locating the product ingredient statement. For example, you may establish a rule that all soft drinks be coded into NOVA class 4 because these products are generally formulated in a way that involves including one or more ingredients that conform with class 4 criteria.
For multi-ingredient home prepared foods (e.g. home-made lasagna, pizza, cookies, etc.) the ingredients used in preparing the food are available in output file 01 (component ingredient file), and may be used to guide food coding decisions for these types of foods.
4) After NOVA classification codes have been assigned to all unique food IDs in file 02, statistical analysis code may be written to assign NOVA classification codes to all foods in all records in file 02. Then, code may be written to create processed food variables of interest for your study (e.g. times per day ultra-processed foods were consumed; percent of total calories from ultra-processed foods, etc.).
* The NCC Database Food Group ID’s, categories, and names are provided in the NDSR file ‘Nccdbfg[insert version].txt in the ‘Database Documentation Files’ folder within the ‘Additional Files’ folder. For Windows installations, the Additional Files are located at C:\Users\Public\Documents\NCC\NDSR [insert version]\Additional Files\Database Documentation.
1. Sneed NM, Ukwuani S, Sommer E, Samuels L, Truesdale K, Matheson D, Noerper T, Barkin S, Heerman W. Reliability and Validity of Assigning Ultra-Processed Food Categories to 24-Hour Dietary Recall Data Collected Using the Nutrition Data System for Research (NDS-R). Current Developments in Nutrition 2022;6:778.

There are two added sugar variables in the NDSR output files- Added Sugars (by Available Carbohydrate) and Added Sugars (by Total Sugars). What is the difference between the two?

Added Sugars are those sugars and syrups added to foods during food preparation or commercial food processing. Ingredients designated as “added sugar” foods in the NCC database include: white sugar (sucrose), brown sugar, powdered sugar, honey, molasses, pancake syrup, corn syrups, high fructose corn syrups, invert sugar, invert syrup, malt extract, malt syrup, fructose, glucose (dextrose), galactose, and lactose. They do not include mono- and disaccharides occurring naturally in foods, such as lactose in milk or fructose in fruit.
The Added Sugars (by Available Carbohydrate) value assigned by NCC to foods considered to be sources of added sugars represents the amount of available carbohydrate present in the food, which includes saccharides of all types. Mono- and disaccharides along with saccharides with a higher degree of polymerization that are resistant to digestion (e.g., trehalose) are included under this definition.

For example, corn syrups with different Dextrose Equivalency (DE) contain a high amount of trisaccharides and other higher saccharides (approximately 75%) due to the incomplete hydrolysis of the cornstarch. These more complex sugars are included under Added Sugars (by available carbohydrate).
The Added Sugars (Total Sugars) value assigned by NCC to foods considered to be a sources of added sugars represents the amount of total sugars present in the food, which includes only mono- and disaccharides.

The nutrient values for brand name food products in the database don’t precisely match the values on the product’s Nutrition Facts panel. Why?

Nutrient values in the NCC Food and Nutrient Database for brand name foods do not precisely match the information on product Nutrition Facts panels for a number of reasons. One reason is that values in the database are not rounded to the nearest whole number as is allowed on the Nutrition Facts panel. Another reason is the database values may not reflect recent changes in the marketplace. For example, if General Mills reformulates Cheerios® today, the nutrient values in the current database may no longer match those on the product label. Discrepancies between database and Nutrition Facts panel values may be due to use of the nutrient composition of representative foods for some brand name product categories for which the nutrient composition across brands is similar.
As an example, although the database includes several brands of pretzel twists, the nutrient values assigned to each are based on a representative pretzel twist. It is important to note that use of a representative food is only done when variation in nutrient content across brands of a product is minimal.

The database includes several brand name products in some food product categories such as snack crackers but no brand name products in other categories such as canned and frozen vegetables. Why?

Brand name products are included if there are significant differences in the nutrient composition of food products within a category. For example, different brands of potato chips are included in the database because the fatty acid content of chips varies notably across brands. Another reason for including brands relates to how people tend to describe the food. For example, commercial cookies tend to be described by brand name (e.g., Oreo® cookie) rather than by generic food description (e.g., chocolate sandwich cookie).

How does NCC decide whether to add new nutrients or food components to the database?

The following factors are considered in deciding whether to add a nutrient or other food component to the database:

  • Scientific Interest: Is there demand for it? If there is a nutrient or food component you’d like added to the database please let us know(
  • Availability of Food Composition Information: Is there analytic composition information available for a significant proportion of core foods in the NCC Food and Nutrient Database?
  • Quality of Analytic Data: Is the analytic information available of sufficient quality (e.g., obtained using appropriate analytic methods) for use in assigning values to foods in the NCC Food and Nutrient Database?

How does NCC assign nutrient values to unknown foods, and how can I figure out what food is being used as the ‘default’ for unknown foods?

In the NCC Food and Nutrient Database there are foods defined as ‘unknown’ (e.g., ‘milk, unknown % fat’). These foods may be selected when a participant does not know the level of detail required for a food.


To assign nutrient values to unknowns NCC uses the nutrient values for the form of the food that is believed to be most commonly consumed in the U.S. For example, the nutrient values for 2% milk are utilized for ‘milk, unknown % fat’. To decide what is most common, NCC relies on scientific and food industry publications that report dietary intake patterns and product sales. Professional judgment is also used where published data is lacking.


If you need to know what food an unknown food defaults to you can look in the output files. The Food File (output file 02) lists the food as it was selected (e.g., milk, unknown % fat). The Component/Ingredient File (output file 01) lists the default food that is associated with the unknown food (e.g., milk, 2 % fat). To quickly identify unknown foods in your dataset use the column in file 2 labeled ‘Unknown (default) Food’. If a food is an unknown there will be a ‘1’ in this column.

Can NDSR be used to estimate intake of FODMAPs (fermentable oligo-, di-, and monosaccharides and polyols)?

NDSR output files include intake estimates for monosaccharides (fructose, galactose, glucose, tagatose), disaccharides (lactose, maltose, sucrose) and a variety of polyols (erythritol, inositol, isomalt, lactitol, maltitol, mannitol, pinitol, sorbitol, xylitol). Intake estimates for oligosaccharides are not available. Thus, intake of all types of FODMAPs except oligosaccharides may be estimated using data available in the output files.

How do I open my output in Microsoft Excel and view it?

If you plan to analyze your data using Excel, you may want to generate the output files with the headers. Once you have generated the output, you will first need to extract the output files. Then open Excel, select Open from the File menu, and browse to the location of the output file you wish to open. Change the “All Excel Files” option to “All Files” in the drop down menu above the Open button, then select the file you want to open. A Text Import Wizard in Excel will pop up and should already recognize that the .txt files are tab delimited. Select “finish” to open the file.

Why are the grains in ounce equivalents columns blank?

The USDA Food Pattern Equivalent grain variables (total, whole, and refined grains) will not be calculated for data collected from versions prior to NDSR 2013. If older data are rerun, the grain per ounce equivalent categories will be in the output files, but data will be missing or blank for these variables. See “The USDA Food Patterns Equivalents Grain Variables” in Appendix 10, Food Grouping” for more information on equivalent grain variables.

Can I analyze my NDSR data for high fructose corn syrup intake?

The short answer is ‘no’. Some NDSR users have wondered if they could create an estimate of HFCS by totaling fructose and glucose, two sugars that are included in the NCC Food and Nutrient Database. This approach is problematic because high fructose corn syrups are just one source of these sugars.
Added Sugars (by Available Carbohydrate) and Added Sugars (by Total Sugars) are included in the database, and depending on your research question, may be useful to you. It is also possible to estimate intake of food categories that tend to include products that contain high fructose corn syrup. For example, servings of sweetened soft drinks may be estimated using the NCC Food Groups Serving Count System.