Food frequency questionnaire
Food frequency questionnaire is a questionnaire used to obtain frequency and, in some cases, portion size information about food and beverage consumption over a specified period of time, typically the past month or year. FFQs are a common dietary assessment tool used in large epidemiologic studies of diet and health. Examples of usage include assessment of intake of vitamins and other nutrients, assessment of the intake of toxins, and estimating the prevalence of dietary patterns such as vegetarianism. It is an example of a dietary assessment tool.
Structure and data collected
Questions
A FFQ includes questions on a set of food and beverage items. For each food or beverage item, there may be questions asking about the following:- Frequency of consumption.
- Portion size information, which may be framed in terms of portion size for each round of consumption or total portion size. Respondents may be asked to translate usual consumption amount to number of specified units, such as a cup of rice. Some questionnaires include portion size images in an attempt to enhance reporting accuracy.
- Other information, such as frequency of consumption at particular times of the year, or the subtype or brand that people consume.
Length and administration format
A FFQ that aims to capture total dietary intake includes questions on 80 to 120 food and beverage items, and takes 30 to 60 minutes to complete.FFQs may be interviewer-administered in case of low literacy as well as when being conducted on children.
FFQs are intended for individual rather than household use. For very young children, the FFQ may be answered by a parent or guardian instead.
Types of FFQs
FFQs are classified as:- Non-quantitative FFQs if they collect only frequency information, and no portion size information, and
- Semi-quantitative FFQs if they collect information on both frequency and portion size.
Relation with other assessment tools
Similar diet assessment tools
FFQ is a common method for dietary assessment, i.e., for constructing the respondent's diet history. Some other methods include:- 24-hour diet recall: This is a structured interview intended to capture detailed information about all the foods and beverages consumed by the respondent in the last 24 hours. For young children, the parent or guardian may respond on the child's behalf. A variant of this method is the multiple-pass recall. Here, after an initial recall, the interviewer goes back to previous questions to clarify further details.
- Other kinds of diet history interviews, such as narrative diet histories.
- Weighted food diary : The individual records details of food and beverage in the diary at the time of consumption. A variant of this method is estimated food records, where the quantification of food and drink is estimated rather than weighed.
Advantages and disadvantages relative to weighted food records
One disadvantage of FFQs and other retrospective methods is that they may be less accurate because participants are relying on memory rather than information recorded in real time.
Advantages and disadvantages relative to the 24-hour diet recall
FFQs have the following advantages:- They can better capture consumption of foods consumed occasionally or episodically than the 24-hour diet recall. For instance, if people habitually consume alcohol more on Fridays, then a 24-hour diet recall conducted on Tuesday will underestimate the extent of alcohol consumption, whereas a well-designed FFQ can capture the information.
- They are easier to administer to literate populations at large scale and also easier to process than diet recalls.
- Due to the longer time period involved, FFQs are more dependent on people having good memory as well as being able to accurately estimate the frequency of rare events. These are cognitively complex tasks that require good generic memory.
- FFQs lack detailed information on how the food was prepared as well as information on exactly how much was consumed and when.
- Since FFQs are limited to a fixed list of foods, they may not capture eating patterns of people who have very different eating patterns.
- FFQs are more subject to social desirability bias, with people over-reporting the extent of consumption of "healthy" foods such as fruits and vegetables.
Validation of FFQs using other methods
Food frequency questionnaires in use
Specificity to populations
Since FFQs list specific food and beverage items, a FFQ designed for and validated against one population is not valid for other populations. Therefore, FFQs must be appropriately modified and revalidated against new populations. A few key challenges include:- Different populations consume different sets of foods and beverages, so the list of items in a FFQ validated against one population may not cover all the items for the other population.
- Even when the items match in name, the methods of preparation may differ significantly enough that results from one FFQ may not be applicable to a different FFQ.
- Different populations use different languages, so the FFQ needs to be translated. Even when populations use the same language the jargon for specific food items can differ. Thus, for instance, in order to administer the Diet History Questionnaire in Canada, a Canadian version, C-DHQ-II, was created.
Common FFQs in the United States
The list below is drawn from existing lists of food frequency questionnaires. All questionnaires in the list below are validated through multiple research studies.| Questionnaire name | Developer | Format of administration | Type of FFQ | Length |
| Harvard FFQ, also known as the Harvard Service FFQ and the Willett FFQ | Walter Willett, M.D., and his colleagues at Harvard University | Pen and paper version only; booklet plus analysis cost roughly $15.00-20.00 per questionnaire | Semi-quantitative | The booklet has 19 pages and 18 questions, with some of the questions having a large number of parts. |
| Diet History Questionnaire | National Cancer Institute's Risk Factor Assessment Branch. The first version, DHQ I, was developed by a team including Fran Thompson and Amy Subar and described in a paper published in 2001. | DHQ I had a single version and was paper-and-pencil-only. DHQ II has four versions, each of which can be taken online or using paper-and-pencil. A Canadian version is also available. Sample PDFs for paper versions are freely available. Other languages supported are Spanish and French. | Semi-quantitative | DHQ I had 124 questions and took an hour to complete. DHQ II has 134 questions in each of its four versions. |
| Block FFQ | National Cancer Institute, under the direction of Gladys Block. The questionnaire design was described in a 1986 paper and the first research paper validating the questionnaire was published in 1990. The FFQ was subsequently modified and a web version was created. | Pen-and-paper and web version, both available at cost. Cost is $2 per respondent for pen-and-paper version, with a minimum of $100. | Semi-quantitative | The 2014 questionnaire has questions of 127 food and beverage items, plus additional questions to adjust for fat, protein, carbohydrate, sugar, and whole grain content. |
| NHANES | National Cancer Institute | Pen-and-paper version mailed to large numbers of respondents periodically. Results are used as comparison baselines for other FFQs. Sample is available online. | Non-quantitative | The 2003/2004 booklet has 139 questions and 24 pages. |
Analysis of data
Calculation of nutrient intake (for semi-quantitative FFQs)
Calculations for nutrient intake can be estimated via computerized software programs that multiply the reported frequency of each food by the amount of nutrient in a serving of that food. References databases commonly used for this purpose are listed below. Note that to estimate total nutrient intake, it is necessary to include dietary supplements in the FFQ and add the nutrient intake from these, particularly when dealing with populations where the consumption of dietary supplements is common.- The United States Department of Agriculture Food and Nutrient Database for Dietary Studies, which in turn is based on the USDA National Nutrient Database. To incorporate the effect of dietary supplements, the USDA Dietary Supplement Integrated Database can be used.
- The University of Minnesota Nutrition Coordinating Center 's Food and Nutrient Database.
Calculation of food groups consumed
A related use of FFQs is to determine how much people consume broad groups of foods. Examples of such food groups are fruits, vegetables, and added sugars. A useful database for this purpose is the USDA's Food Patterns Equivalents Database.Comparison data
In the United States, data from NHANES is used to provide nationwide comparison data.Population-level analyses
FFQs can be used for a variety of population-level analyses:- Estimation of total consumption of specific foods, food groups, and nutrients in a population.
- Estimation of the correlation between patterns of food consumption and other non-diet variables, such as health status or specific medical conditions.
- Estimation of the effectiveness of interventions to change diet. In this experimental design, both the experiment subjects and the control group are asked to fill in the FFQ before and after the intervention, and changes in food consumption patterns for the experimental group are measured. However, because of concern about potential differential response bias due to the tendency of the intervention group to misreport their diets to a greater extent than the control group, use of an FFQ as the sole evaluation method should be carefully considered.
Criticism
Alan R. Kristal, Ulrike Peters and John D. Potter wrote in their article "Is It Time to Abandon the Food Frequency Questionnaire?" that there was not much scope to learn more from FFQs, and that a more curious and exploratory approach was needed to uncover new insights on diet and its correlation with health status.
The National Cancer Institute has argued that through careful bias correction, some of the limitations of FFQs can be overcome, and they can be very useful.