National Health and Nutrition Examination Survey

The objective of the dietary interview component is to obtain detailed dietary intake information from NHANES participants. The dietary intake data are used to estimate the types and amounts of foods and beverages (including all types of water) consumed during the 24-hour period prior to the interview (midnight to midnight), and to estimate intakes of energy, nutrients, and other food components from those foods and beverages. Following the dietary recall, participants are asked questions on salt use, whether the person’s overall intake on the previous day was much more than usual, usual or much less than usual, and whether the respondent is on any type of special diet. Questions on frequency of fish and shellfish consumed during the past 30 days were asked of survey participants 1 year or older, with the use of proxies for young children (see the MEC In-Person Dietary Interviewers Procedures Manual for more information on the proxy interview).

The dietary interview component, called What We Eat in America (WWEIA), is conducted as a partnership between the U.S. Department of Agriculture (USDA) and the U.S. Department of Health and Human Services (DHHS). Under this partnership, DHHS' National Center for Health Statistics (NCHS) is responsible for the survey sample design and all aspects of data collection and USDA’s Food Surveys Research Group (FSRG) is responsible for the dietary data collection methodology, maintenance of the databases used to code and process the data, and data review and processing.

All NHANES participant are eligible for two 24-hour dietary recall interviews. The first dietary recall interview is collected in-person in the Mobile Examination Center (MEC) and the second interview is collected by telephone 3 to 10 days later.

As in previous years, two types of dietary intake data are available for the 2011-2012 survey cycle: Individual Foods files and Total Nutrient Intakes files.

What's New with the 2011-2012 WWEIA Release:

In the Individual Foods files, the range of response options regarding the source from which each food/beverage was obtained (e.g., from a store, fast food restaurant, cafeteria, etc.) has been expanded. The former “store” code has been divided into three separate codes to specify whether the source was a grocery store/supermarket, a convenience store, or store—no additional information.

Appendix 1 provides a summary of changes among the 5 latest cycles of data collection.

New additions for analyzing WWEIA data include:

Expanded documentation on Combination codes.

The What We Eat in America Food Categories, available on the FSRG website (www.ars.usda.gov/ba/bhnrc/fsrg), are a new grouping scheme that combines foods and beverages together that have similar usage and nutrient content with the emphasis on how they are commonly consumed in the American diet. There are approximately 150 unique categories and each is assigned a 4-digit number and description. The WWEIA Food Categories contain discrete food items and are not disaggregated (e.g., pizza vs. grain, cheese, tomatoes, etc.). Designed to be flexible, the categories can be combined as needed to address specific research questions. A new version of the WWEIA Food Categories is produced for each 2-year release cycle of WWEIA, NHANES.

Distinguishing Between Foods/Beverages and Dietary Supplements in NHANES

The 24-hour dietary supplement use component is administered after the 24-hour dietary recall. All NHANES participants responding to the 24-hour dietary recall interview are eligible for the dietary supplement and non-prescription antacid use questions. Information is obtained on all vitamins, minerals, herbals, and other dietary supplements as well as non-prescription antacids that were consumed during a 24-hour time period (midnight to midnight), including the name and the amount of supplement or antacid taken.

Distinguishing between foods/beverages and supplements can be challenging. NCHS and FSRG review questionable items reported in the dietary supplement and dietary recall components to resolve disposition of these items into the appropriate component. Products that are labeled as a dietary supplement, that have a supplement facts panel on the label, and are in tablets, capsules, softgels, gelcaps, or other pill forms, are considered dietary supplements. Items that are powders or liquids can be hard to distinguish. General guidelines used state that if powders and liquid concentrates have product directions stating that they be added to a liquid, they are classified as beverages. Examples are teas and protein powders. An exception is made for fiber products, which are classified as dietary supplements. Along this same guideline, energy drinks are considered beverages, but “energy shot” type products are considered dietary supplements.

It is best to refer to the two databases that detail every food/beverage and dietary supplement reported in NHANES to identify exact determination used. The databases are:

• 2011-2012 Food and Nutrient Database for Dietary Studies 
• NHANES Dietary Supplement Database 

Dietary Interview Data Files: Four data files were produced from the information collected in the dietary interviews: two Individual Foods files and two Total Nutrient Intakes files. Each file includes one day of intake data. The number “1” or “2” in the file name identifies the day (and mode) of the interview: 1 = first day (in-person), 2 = second day (phone). File names are as follows:

All NHANES participants are eligible for the dietary interview component. However, only participants one year or older are eligible for the frequency of fish and shellfish consumption questions following the 24-hour recall.

The examination protocol and data collection methods are fully documented in the NHANES dietary interviewer procedures manuals (In-person interview: https://wwwn.cdc.gov/nchs/data/nhanes/2011-2012/manuals/mec_in_person_dietary_procedures_manual_jan_2012.pdf
phone follow-up interview: https://wwwn.cdc.gov/nchs/data/nhanes/2011-2012/manuals/phone_follow-up_dietary_procedures.pdf).

Interviews were conducted for survey participants less than six years of age with a proxy (who was generally the person most knowledgeable about the survey participant’s intake). Interviews of children ages 6 to 8, were conducted with a proxy and with the child present to assist in reporting intake information. Interviews of children ages 9-11, were conducted with the child and the assistance of an adult familiar with the child’s intake. Participants 12 years or older answered for themselves. Dietary interviewers conducted interviews in English and Spanish. Translators were used to conduct interviews in other languages.

The in-person interview was conducted in a private room in the NHANES mobile examination center (MEC). A set of measuring guides (various glasses, bowls, mugs, drink boxes and bottles, household spoons, measuring cups and spoons, a ruler, thickness sticks, bean bags, and circles) was available in the MEC dietary interview room for the participant to use for reporting amounts of foods (NHANES Measuring Guides for the Dietary Recall Interview: https://www.cdc.gov/nchs/nhanes/measuring_guides_dri/measuringguides.htm). Upon completion of the in-person interview, participants were given measuring cups, spoons, a ruler, and a food model booklet, which contained two-dimensional drawings of the various measuring guides available in the MEC, to use for reporting food amounts during the telephone interview. Telephone dietary interviews were collected 3 to 10 days following the MEC dietary interview, but not on the same day of the week as the MEC interview. Any participant who did not have a telephone was given a toll-free number to call so that the recall could be conducted.

What We Eat in America data were collected using USDA's dietary data collection instrument, the Automated Multiple Pass Method (AMPM) (http://www.ars.usda.gov/ba/bhnrc/fsrg). The AMPM was designed to provide an efficient and accurate means of collecting intakes for large-scale national surveys. The AMPM is a fully computerized recall method that uses a 5-step interview outlined below:

1. Quick List - Participant recalls all foods and beverages consumed the day before the interview (midnight to midnight).
2. Forgotten Foods - Participant is asked about consumption of foods commonly forgotten during the Quick List step.
3. Time and Occasion - Time and eating occasion are collected for each food.
4. Detail Cycle - For each food, a detailed description, amount eaten, and additions to the food are collected. Eating occasions and times between eating occasions are reviewed to elicit forgotten foods.
5. Final Probe - Additional foods not remembered earlier are collected.

The AMPM includes an extensive compilation of standardized food-specific questions and possible response options. Routing of questions is based on previous responses. The AMPM is updated for each 2-year collection of WWEIA to reflect the changing food supply and to address research needs from the data user community. Additional information about the AMPM is provided in Raper et al. (Raper et al., 2004).

The AMPM was validated in a large study and shown to be an effective method for collecting accurate group energy intake of adults. Completed in 2004, this extensive research project included 524 healthy, weight-stable volunteers, aged 30-69 years. The accuracy of the AMPM was evaluated by comparing reported energy intake (EI) to total energy expenditure (TEE) using the doubly labeled water technique (Moshfegh et al., 2008). Among the findings were that EI compared to TEE was under-reported by 11% overall, by less than 3% for normal weight subjects with body mass index (BMI) < 25 and 16% for overweight subjects with BMI ≥ 25.

Additional studies provide evidence that the AMPM accurately measures group energy intake. Blanton (Blanton et al., 2006) reported that EI was not significantly different from TEE for a sample of 20 adult females. Rumpler and colleagues found that mean EIs were accurately reported for a sample of 12 adult males (Rumpler et al., 2008).

Additional evidence for the accuracy of AMPM has been provided by analysis of the 24-hour urinary sodium data collected in the AMPM Validation Study, which suggest the AMPM is a valid measure for estimating sodium intake in adults. Dietary sodium intake calculated from 24-hour recall data of 465 subjects collected via AMPM was compared with sodium values from 24-hour urine collections measured during the same 24-hour period. The AMPM-derived mean dietary sodium estimates reflected over 90% of the biomarker-based estimates (Rhodes et al., 2013).

For additional information about the dietary interview component and related survey protocols, please go to the Survey Operations Manuals, Brochures, Consent Documents site at: https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/default.aspx?BeginYear=2011.

All dietary interviewers were required to complete an intensive one-week training course and to conduct supervised practice interviews before working independently in the field. Retraining sessions were conducted annually to reinforce the proper protocols and technique.

Interviewers were monitored throughout the data collection period. Monitoring consisted of the following:

• Reviews of audio recorded interviews or in-person observations were conducted for approximately 5% of each interviewer's work.
• Interviews were checked for completeness of the recalls, missing information, inconsistent reports, and unclear notes. Written notification and feedback were provided to the interviewers.

Interview data files were sent electronically from the field and were imported into Survey Net, a computer-assisted food coding and data management system developed by USDA (Raper et al., 2004).

USDA's Food and Nutrient Database for Dietary Studies (FNDDS) 2011-2012 was used for processing the 2011-2012 intakes (http://www.ars.usda.gov/ba/bhnrc/fsrg). The FNDDS includes comprehensive information that can be used to code individual foods and portion sizes reported by participants and also includes nutrient values for calculating nutrient intakes. The underlying nutrient values for FNDDS 2011-2012 were based on values in the USDA National Nutrient Database for Standard Reference, release 26, produced by USDA’s Nutrient Data Lab (http://www.ars.usda.gov/nutrientdata). FNDDS values are updated for every 2-year WWEIA, NHANES release cycle.

FNDDS 2011-2012, which corresponds with WWEIA 2011-2012, now includes the addition of enhanced characterization of food and beverages indicating if the item is fortified or contains fortified ingredients. Another enhancement for FNDDS 2011-2012 is the incorporation of the Recipe Protocol Project, which provides a framework for the selection of ingredients and amounts across similar foods. The protocols standardize and modernize new and existing recipes for about 2,500 items in FNDDS 2011-2012. Additional information about the FNDDS and related tools is available on the Food Surveys Research Group website (Raper et al., 2004; Anand et al., 2006; Ahuja et al., 2008).

Coders were required to pass a certification test after the initial training. They were routinely monitored to ensure the quality and completeness of their work. Approximately 10 percent of the coder's work was randomly selected to be independently coded by another coder. Results from the two codings were compared and adjudicated, if necessary.

After intake data were coded, various types of reviews were conducted to ensure the quality of the data. An overview of quality assurance procedures conducted during the data processing stage is available in Anand, et al. (Anand et al., 2006). Examples of reviews include the following:

• Overall acceptability of each recall. This review determined if the recall met the two minimum criteria listed below. A recall was considered unacceptable if it failed to meet either of these criteria:
  1. The first 4 steps of the 5-step AMPM were completed. Failure to meet this criterion occurs infrequently and is usually due to the participant stopping the interview before completion of the fourth step. This step collects the details (description and amount consumed) for each reported food.
  2. Foods consumed for each reported eating occasion must be identified. A report of an eating occasion with missing foods occurs most often when an adult serving as a proxy, usually for a child, has reported that the survey participant ate a meal, usually at daycare or school, but does not know what was served or how much was consumed (Anand et al., 2006).
• Interviewers' and coders' questions and comments are reviewed to ensure that they have been accounted for in coding.
• Decisions are made about how to code new or unusual foods or food quantities reported by participants. 
  Foods/beverages or portions that could not be matched to items in the database were resolved by FSRG scientists. New items and new portion sizes are added as needed to the FNDDS. Information about new foods/beverages and package sizes were collected using internet resources, direct contact with manufacturers, or food labels. Unusual food mixtures were coded using multiple food codes to represent the mixture, linking those foods with a combination food number, and specifying the type of combination.
• Specific data integrity checks for reasonableness, consistency, and logic.
  Many quality control features are built into the AMPM software, including automated routing based on the participant's previous answers and extensive checks which prevent illogical responses. Nevertheless, over 30 unique checks were conducted across all dietary data.
• Intakes with extreme levels for individual nutrients.
  Nutrient intakes were reviewed separately for various age and gender groups to determine if corrections were necessary.

During data processing, the following edit was made to ensure the logical consistency and analytic usefulness of the data:

• Modified nutrient values for some food mixtures.
  During the food coding process, predefined recipes for some food mixtures were modified to match more closely the food as described by the respondent. Nutrients were modified by substituting ingredients in a predefined recipe for the mixture. An example of a modified recipe is spinach cooked with butter instead of margarine. Each modification was assigned a unique 6-digit identification number. Recipe modification numbers appear in the variable DR1MC in the DR1IFF_G file and in the variable DR2MC in the DR2IFF_G file. Descriptions for each modification are provided in a separate file called DRXMCD.
  
  Starting in 2011, the following foods were not modified because new items are now included in FNDDS 2011-2012 to capture this detail: eggs; fish; macaroni and cheese; oatmeal; rice, select salads—such as potato salad, coleslaw, tuna salad, etc.; rice.

A data processing step called salt adjustment that was based on survey participants' responses to selected questions was discontinued in 2009-2010 (Sebastian et al. 2013; Sebastian et al. 2012). For participants whose households used salt in cooking occasionally or less often, some or all of the salt attributable to home preparation was removed from foods that typically have salt added during preparation and were obtained from the store.
Factors leading to the discontinuation of the data processing step of salt adjustment, include the following:

• The use of store purchase as a proxy indicator of home preparation is no longer appropriate, due to the proliferation of ready-to-eat, ready-to-heat, and other convenience foods that are available in stores.
• Results from the AMPM Validation Study demonstrate that the AMPM is a valid measure for estimating sodium intakes at the group level when the data processing step of salt adjustment is not applied. 

For more detail about the data processing step of salt adjustment and its discontinuation, see Sebastian et. al. (Sebastian et al. 2013; Sebastian et al. 2012).

Each Individual Foods file (Day 1 and Day 2) is comprised of food records. For most survey participants, there are multiple records in each file. For each Total Nutrient Intakes file (Day 1 and Day 2) there is one record for each participant. These files can be linked with other NHANES files by the respondent sequence number (SEQN).

Variable names: For data collected on both Day 1 and Day 2, variable names are differentiated by having the number "1" or "2" in the third position of the variable name to identify the collection day. For example, the USDA food code variable (in the Individual Foods File), which identifies the food reported by the participant, is named DR1IFDCD in the Day 1 file and DR2IFDCD in the Day 2 file. Appendices 2 and 5 list the Day 1 and Day 2 variable names for the Individual Foods file and the Total Nutrient Intakes file, respectively.

Names for the following variables are the same for both days in the Individual Foods file and the Total Nutrient Intakes file:

Number of days of intake: A variable has been included to indicate the number of days of intake collected from each participant. The variable name is DRDINT. In 2011-2012, 8,519 participants provided complete dietary intakes for Day 1. Of those providing the Day 1 data, 7,605 provided complete dietary intakes for Day 2.

Dietary recall status code: A status code (DR1DRSTZ or DR2DRSTZ) is used in both the Individual Foods and Total Nutrient Intakes files to indicate the quality and completeness of a survey participant's response to the dietary recall section. The codes are the following:

1 = Reliable and met the minimum criteria

For individuals with a code 1, all relevant variables associated with the 24-hour dietary recall contain a value.

2 = Not reliable or did not meet the minimum criteria

Individuals with a code 2 have incomplete records. No data on total nutrient intakes and the total number of foods reported are provided for these cases. These individuals have no records in the Individual Foods files.

3     [Code 3 is not included in the current datasets. It was only used for data from the 1999-2000 survey cycle.]

4 = Reported consuming breast milk

For infants and children who consumed human milk, there is a record in the Individual Foods files for each report of human milk. However, because amounts of human milk intake are not quantified, these records contain missing values for the amount consumed and for the amounts of energy and nutrients from human milk. Also, records of human milk have a missing value for the food source variable (DR1FS, DR2FS) and the eaten at home variable (DR1_040Z, DR2_040Z) in the Individual Foods files. Records for any other foods and beverages consumed by breast-fed infants and children are included in the Individual Foods files along with their amounts and nutrient information. Because of the missing amount or quantity information for human milk, no total nutrient intakes (contained in the Total Nutrient Intakes files) were computed for participants with a code 4.

A variable that identifies breast-fed children, DRABF, is included. This variable has a code of 1 if a child consumed human milk in either intake day.

5 = Not done

This code is assigned when the dietary recall section of the interview did not take place due to various reasons (such as arrived late/left early, refusal, illness, emergency, or equipment failure). These individuals have no records in the Individual Foods files. These individuals have a record in the Total Nutrients file with values only for the following variables: the respondent sequence number (SEQN), the dietary recall status code (DR1DRSTZ or DR2DRSTZ) and for participants one year or older, the fish and shellfish questions in the DR1TOT_G file (DRD340, DRD350A-K, DRD350AQ-JQ, DRD360, DRD370A-V, and DRD370AQ-UQ).

Although there are four possible values, only codes 1 and 4 appear in the Individual Foods file. In addition to the status code described above, the variable DR1_300 and DR2_300 in the Total Nutrients file, denotes the participant's assessment of whether the amount of food consumed on the recall day was usual, much more than usual, or much less than usual.

Participants who reported consuming only water, no food or other beverages: Records are included in the Individual Foods file for participants who consumed only water. There are 7 such individuals in the 2011-2012 datasets, 1 in the Day 1 data and 6 in the Day 2 data. Their dietary recall status variable for the day is coded as "1" (complete and reliable) in the Total Nutrients file and the total number of items is coded based on how many times water was reported. Individuals with just water intake and no food intake will have zero energy intake for the day.

Participants who reported consuming no water, food or other beverages: There can be survey participants whose intakes are determined to be complete even though they reported no water, food, or other beverage records for the day. For such participants there are no records in the Individual Foods file but their dietary recall status is coded as complete and reliable and the Total Nutrients file will include records with zero values for all nutrients. In the 2011-2012 datasets, there are no such participants.

Number of days between the intake day and the day of family interview: Each of the four intake files includes a variable (DR1DBIH for Day 1 files and DR2DBIH for Day 2 files) to indicate the number of days between the intake day (i.e., the period covered by the 24-hour recall) and the day that the family questionnaire was administered in the household. A positive value in DR1BHIH or DR2BHIH indicates the family interview occurred prior to the intake day. In the survey, most of the family interviews were done before the participant came to the MEC and participated in the dietary interview. A value of "0" in DR1BHIH or DR2BHIH indicates the family interview occurred on the same date as the intake day. A negative value (i.e., DR1BHIH<0 or DR2BHIH<0) means that the family interview occurred after the intake day.

Food source: The source from which each food was obtained (e.g., from a store, fast food restaurant, cafeteria) is identified by the variables DR1FS (day 1) and DR2FS (day 2) in the Individual Foods files. 

In 2011-2012, the code for “store” (value=1) has been divided into three separate codes: “Store - grocery/supermarket” (value=1), “Store - convenience type” (value=27), and “Store - no additional info” (value=28). Additionally, descriptions “Cafeteria not at school” and “Cafeteria at school” have been revised to “Cafeteria NOT in a K-12 school” and “Cafeteria in a K-12 school.”

The code descriptions for this variable are:

Eating occasion: The variables DR1_030Z and DR2_030Z are located in the Individual Foods file. The code descriptions for the eating occasion variables are shown in the table below.

Eating occasion was designated by the respondent. During the interview, a list of eating occasion names was available to the respondent for selection. However, eating occasion names were not defined for the respondent and the interpretation may differ from one person to another.

Foods and beverages coded as part of a combination: Forty-three percent of foods and beverages reported in WWEIA, NHANES 2011-2012 were identified as items consumed together as combinations. Items consumed as a combination were identified by one of fifteen unique "combination food types". Foods and beverages not coded in combination have the code "0" for the combination food type variable.

The combination types provide a linkage for:

• Foods or beverages with additions, such as cereal with milk, coffee with cream;
• Multi-component foods that have specific protocol for collection such as some salads and sandwiches (primarily those that are not from fast food establishments); and
• Other combinations that do not have a unique code in the FNDDS.

All items given a combination food type are given an additional variable to identify each of the items within the combination. This variable is the "combination food number" that is unique to the combination food type within the individual intake.

Special diet: Information on whether the participant is currently on any kind of diet to lose weight or for other health-related reason and, if so, the type of diet, was provided by the variable DRQSDIET identifies whether a participant was on a special diet. The variables DRQSDT1 through DRQSDT12 and DRQSDT91 identify the type of diet(s) that the participant was following. These variables can be found in the Total Nutrient Intakes file.

Note: A participant could report more than one type of diet, and all the responses were recorded. For researchers studying this topic using data from more than one survey cycle, it is important to note that the number of response options increased from 2003-2004 to 2009-2010. There was no change in response options between 2009-2010 and 2011-2012.

Sample weights for dietary intake data: The NHANES participants were selected on the basis of a national probability design. In order to increase the number of participants for specific demographic groups, a multi-stage, unequal probability of selection design was implemented. Beginning with 2011-2012 collection the sample design includes an oversample of Asian Americans.

Sample weights are constructed that encompass the unequal probabilities of selection, as well as adjustments for non-participation by selected sample persons. In order to produce national, representative estimates, the appropriate sample weights must be used.

For the 2011-2012 NHANES, there were 13,431 persons selected; of these 9,338 were considered participants to the MEC examination and data collection. 8,519 of the MEC participants provided complete dietary intakes for Day 1 and of those providing the Day 1 data, 7,605 provided complete dietary intakes for Day 2.

Most analyses of NHANES data use data collected in the MEC and the variable WTMEC2YR should be used for the sample weights. However, for the WWEIA dietary data, different sample weights are recommended for analysis. Although attempts are made to schedule MEC exams uniformly throughout the week, proportionally more exams occur on weekend days than on weekdays. Because food intake can vary by day of the week, use of the MEC weights would disproportionately represent intakes on weekends.

A set of weights (WTDRD1) is provided that should be used when an analysis uses the Day 1 dietary recall data (either alone or when Day 1 nutrient data are used in conjunction with MEC data). The set of weights (WTDRD1) is applicable to the 8,519 participants with Day 1 data. Day 1 weights were constructed by taking the MEC sample weights (WTMEC2YR) and further adjusting for (a) the additional non-response and (b) the differential allocation by day of the week for the dietary intake data collection. These Day 1 weights are more variable than the MEC weights, and the sample size is smaller, so estimated standard errors using Day 1 data and Day 1 weights are larger than standard errors for similar estimates based on MEC weights.

When analysis is based on both days of dietary intake, only 7,605 sample participants have complete data. The NHANES protocol requires an attempt to collect the second day of dietary data at least 3 days after the first day, but the actual number of days between the two interviews is variable. A set of adjusted weights, WTDR2D, is to be used only when an analysis uses both Day 1 and Day 2 dietary data. This two-day weight was constructed for the 7,605 participants by taking the Day 1 weights (WTDRD1) and further adjusting for (a) the additional non-response for the second recall and (b) for the proportion of weekend-weekday combinations of Day 1 and Day 2 recalls.

Note that all sample weights are person-level weights and each set of dietary weights should sum to the same population control total as the MEC weights (WTMEC2YR). In addition, the MEC weights (WTMEC2YR) are appropriate for use in the analysis of the fish and shellfish consumption data (i.e., variables DRD340, DRD350A-K, DRD350AQ-JQ DRD360, DRD370A-V, and DRD370AQ-UQ) located in the Day 1 Total Nutrient Intake File (DR1TOT_G), if no other dietary data are included in the analysis. Additional explanation of sample weights and appropriate uses are included in the NHANES Analytic Guidelines. Please also refer to the on-line NHANES Tutorial for further details on other analytic issues.