Understanding the Purpose and Limitations of FFQs
A food frequency questionnaire (FFQ) is a retrospective tool that assesses habitual dietary intake over a specific period, such as the past month or year. FFQs are commonly used in large-scale epidemiological studies due to their relatively low cost and minimal respondent burden compared to other methods like weighed food records or 24-hour recalls.
Types of Food Frequency Questionnaires
FFQs can be broadly categorized based on whether they collect portion size information:
- Non-quantitative FFQs: These questionnaires only collect information on the frequency of consumption (e.g., times per day, week, or month) without asking for portion sizes. They are useful for ranking individuals by intake but are not ideal for estimating absolute nutrient amounts.
- Semi-quantitative FFQs: This type includes both consumption frequency and portion size data. Portion sizes are typically standardized or offered as a range of choices. Most prominent FFQs fall into this category.
Inherent Limitations of FFQs
Despite their benefits, FFQs have limitations that must be considered:
- Recall Bias: Respondents may struggle to accurately remember consumption patterns over a long period, which can be influenced by recent diets or social desirability (over-reporting "healthy" foods, under-reporting "unhealthy" ones).
- Population Specificity: FFQs contain a fixed list of foods and may not be suitable for diverse or different populations unless adapted and re-validated. Dietary habits vary significantly by culture, ethnicity, and geography.
- Difficulty with Portion Sizes: The ability of participants to accurately estimate portion sizes is often limited, even with visual aids.
- Systematic Error: FFQs are prone to systematic error, which can lead to over- or under-estimation of nutrient intake compared to objective methods or food records.
Comparison of Prominent Food Frequency Questionnaires
Different FFQs have been developed for various purposes and populations. Here is a comparison of some well-known examples:
| Feature | NCI Diet History Questionnaire (DHQ) | Harvard FFQ (Willett FFQ) | Culturally-Specific FFQs |
|---|---|---|---|
| Developer | National Cancer Institute (NCI) | Harvard University (Walter Willett) | Custom-developed for specific populations |
| Availability | Web-based version (DHQ) is free for research | Analysis services available via Harvard | Requires specific development and validation |
| Design | Semi-quantitative, uses embedded portion size questions | Semi-quantitative, portion size included with food item | Varies based on dietary patterns of target population |
| Strengths | Lower cost (DHQ), web-based format is convenient, well-validated for nutrients | Extensively validated, particularly strong for long-term diet-disease studies | Highly relevant and accurate for a specific population's diet |
| Weaknesses | Can have higher recall burden for certain items | May not capture culturally diverse diets without modification | High initial development and validation cost |
| Ideal Use | Large epidemiological studies needing low cost and ease of administration | Long-term prospective studies, assessing nutrients like simple sugars | Research focusing on unique dietary patterns, low-literacy populations |
Choosing the Right FFQ: Key Considerations
To determine the best food frequency questionnaire for your needs, consider the following factors:
- Your Research Question: If you are interested in habitual intake and ranking individuals for a large epidemiological study, a validated FFQ like the DHQ or Harvard is appropriate. If you need precise absolute intake data, consider a different method or use the FFQ with a calibration study.
- Target Population: The FFQ's food list must be relevant to the population you are studying. An FFQ developed for a Western population will be invalid for an Asian or ethnic minority group. Adapting or creating a new FFQ may be necessary.
- Available Resources: FFQs are generally inexpensive to administer, especially web-based versions. However, developing a new FFQ or using a service like Harvard's involves costs.
- Study Design: FFQs are ideal for prospective cohort or large case-control studies. They are less suitable for short-term dietary tracking or intervention studies where participants might alter their eating habits.
- Acceptable Error Level: FFQs provide an estimate of usual diet and have known systematic errors, often underestimating overall energy intake. The decision to use an FFQ should align with the acceptable level of measurement error for your study's objectives. When high precision is crucial, combining an FFQ with a more accurate reference method (e.g., 24-hour recalls) for a calibration sub-study is recommended.
Conclusion: No Single "Best" FFQ Exists
In conclusion, there is no single "best" food frequency questionnaire. The most suitable tool is the one that is carefully selected to align with the specific research question, target population, and available resources. For large-scale studies where ranking individuals is the primary goal, validated FFQs like the web-based DHQ or the Harvard FFQ are excellent, low-cost options. However, researchers must be aware of inherent limitations, such as recall and reporting biases, and consider using or adapting culturally specific FFQs for diverse populations. Ultimately, the quality of the data is a result of a thoughtful and deliberate choice, not simply selecting a pre-existing tool. Researchers are encouraged to review existing literature and validation studies for FFQs relevant to their target populations before making a final decision.
Visit the NCI's Dietary Assessment Primer for more information on dietary assessment methods.
How FFQs are Validated
FFQs are validated by comparing their results against a more objective or accurate dietary assessment method. Common reference methods include:
- Repeated 24-hour Dietary Recalls: Multiple 24HRs are used to capture the variation in a person's diet over several non-consecutive days.
- Weighed Food Records: Considered a gold standard, this method requires participants to weigh and record all food and drink consumed over a specific period, but it is highly burdensome for respondents.
- Biomarkers: Objective measures like doubly-labeled water for energy or urinary nitrogen for protein can be used to validate reported intake, as they are not subject to the same recall biases as FFQs.
