Data Capture

Mode of Administration

• Can be self- or interviewer-administered. If interviewer administered, can be done in person or over the telephone.
• Can be administered by paper or electronically (Learn More about Technology in Dietary Assessment).

Population Considerations

• Cultural adaptation may be required to ensure that the food list captures foods commonly consumed in the particular study population, the food list terms used are commonly understood, and the portion sizes used reflect the amounts typically consumed.
• Not appropriate for populations that are cognitively or numerically challenged.
• Interviewer-administration is appropriate for low-literacy groups; self-administration is not.
• Some groups cannot answer for themselves (e.g., children or some people with disabilities). Proxy reporting is possible, but data collected may be limited by knowledge of the [glossary term:] surrogate reporter (Learn More about Surrogate Reporting).

Resource Requirements

• Data collection and data entry system is required, either paper or electronic.
• Hardcopy versions require optically scannable forms, ability to scan, and software (Learn More about Software for Food Frequency Questionnaires and Screeners) to convert data to nutrient or food group intake estimates.
• Electronic versions require that respondents have computer and/or Internet access and that the researchers have software to convert data to nutrient or food group intake estimates.

Study Design Considerations

For more details on the following issues when considering whether to use an FFQ to answer a particular research question, see Choosing an Approach for Dietary Assessment.

General Considerations

• For long-term [glossary term:] prospective studies, multiple administrations over time may be beneficial, so that reported intake spans the entire study period of interest.
• Choice of time frame (e.g., past year or past month), should be guided by the study objective and also may be informed by factors related to seasonality (i.e., how often and how much food is consumed in season and not in season) or cognitive abilities. Some FFQs, such as the Diet History Questionnaire (DHQ), incorporate seasonality into the instrument, possibly easing the task of accurately reporting intake and reducing [glossary term:] measurement error [14].
• The population's diet can differ systematically over the course of a year, and the possibility of [glossary term:] season effect, should be considered in study design or analysis (Learn More about Season Effect). The effect of season of administration on reporting [glossary term:] accuracy has not been studied extensively. One study found a small difference in reporting frequency over the past year depending on season of administration [14].
• Even though intakes are known to vary from day to day (e.g., from weekdays to weekend days; Learn More about Day-of-Week Effect), this [glossary term:] day-of-week effect is not a factor for FFQs. An FFQ can be administered on any day of the week when the time frame queried is longer than a week.
• [glossary term:] Within-person random error in FFQs is due to individuals reporting their diet differently at two time periods. Although this is not much of an issue with FFQs, a second administration some months after the first allows measurement of and statistical correction for within-person random error, assuming no change in diet in that time interval.
• The collection of [glossary term:] recovery biomarkers in a subsample of participants should be considered; these can be used as [glossary term:] reference instruments to reduce [glossary term:] bias for some nutrients.

Guidance for Specific Research Objectives

• If your research objective is to estimate the [glossary term:] mean intakes of a group, a single administration of an FFQ may be sufficient, although the absolute means are likely to be biased. An [glossary term:] internal calibration sub-study in which multiple non-consecutive 24HR or food records are administered in conjunction with the FFQ on at least a portion of the sample allows statistical adjustment for bias in data from FFQs. Alternatively, data from an external source (called an [glossary term:] external calibration study) can be used (Learn More about Calibration).
• If your research objective is to estimate [glossary term:] usual dietary intake distributions for a group (e.g., to examine percentiles or estimate the proportion above or below some threshold), the use of an FFQ without [glossary term:] calibration is not recommended. Distributions estimated from an FFQ (and a screener) are narrower than true distributions. Thus, prevalence estimates in the tails of the distribution are biased (Learn More about Usual Dietary Intakes).
• If your research objective is to analyze the [glossary term:] association between diet as an independent variable and another variable (e.g., diet at baseline and onset of cancer), a single FFQ can be used, but multiple FFQs will allow correction for any within-person random error that is present. More importantly, conducting an internal calibration sub-study allows for some statistical correction of bias using [glossary term:] regression calibration (Learn More about Regression Calibration). Alternatively, data from an external calibration study can be used.
• If your research objective is to analyze the association between an [glossary term:] independent variable (e.g., socioeconomic status) and diet as a [glossary term:] dependent variable, variables known to affect quality of report, such as body mass index, should be assessed and later included as [glossary term:] covariates in analyses.

A study in which multiple non-consecutive 24HRs, food records, or recovery biomarkers are administered in conjunction with an FFQ, on at least a portion of the sample, allows for some statistical correction of bias.

• If your research objective is to analyze change in diet as a result of an intervention (e.g., to evaluate the effectiveness of an educational program to encourage fruit and vegetable intake), the sole use of an FFQ is not recommended because of overall bias and potential [glossary term:] differential response bias.

To avoid the effects of these potential biases, an objective measure, such as measurement of serum carotenoids as a marker for fruit and vegetable intake, could be considered. In addition, collecting and adjusting for other non-dietary self-report information (e.g., social desirability scale) may help characterize individuals who may be more prone to [glossary term:] response bias (Learn More about Social Desirability).

If an FFQ is used as the [glossary term:] main dietary assessment instrument, an internal calibration sub-study with the collection of objective measures is recommended to assess and adjust for overall bias. If objective measures are not feasible, multiple non-consecutive 24HR, collected at each relevant time period and for each group (i.e., treatment and control) may allow adjustment of biases in estimates of the means. However, the ability of 24HRs to accomplish such adjustment is questionable because recalls themselves may be subject to differential response bias in an intervention setting.

An FFQ used should be sufficiently sensitive to detect changes of interest.