*Geek Box: Semi-Quantitative FFQ

*Geek Box: Semi-Quantitative FFQ

The conceptual basis of food-frequency questionnaires [FFQ], as the primary dietary assessment method in nutritional epidemiology, is dietary intake over a number of years. Assessing frequency of intake over the previous year allows for responses to be independent of seasonal variations in intake, which may occur in individuals.

The frequency aspect of an FFQ provides a scale of potential answers which are designed to minimise the variation in actual frequency of intake which could result if a narrower range of options were provided. This allows for more options in the higher frequency categories, to capture foods which make more regular [frequent] contributions to nutritional intake.

However, this only reflects data on frequency, with no additional information on portion sizes: this is known as a ’simple FFQ’. A semi-quantitative FFQ, on the other hand, also specifies a portion size with the question on frequency, e.g., instead of “yogurt” it would say “yogurt [1 cup]” (where ‘cup’ is US measures, i.e., 250ml). Because the specification of portion size is one option, and it is not a measured food diary, it is ’semi-quantitative’.

The quantitative aspect of the FFQ is influenced by the food itself, as certain foods have more natural portion sizes that are intuitive, e.g., “slice of bread”, “1 egg”, or “1 banana”. Conversely, other foods do not have natural portion sizes, for example pasta or meat. Typical portion sizes are often used for these foods, but a “bowl of pasta” for me may be substantially different to yours. The challenge for the semi-quantitative aspect of FFQ for foods without natural portion sizes is that research demonstrates humans are quite poor at accurately describing a portion size, even immediately following a meal. To attempt to navigate this issue, average portion sizes are used, e.g., 6-8oz for meats, 2 slices of bacon, or 1 hamburger, or 1/2cup rice or pasta: this is set within the question with more frequency categories.

The aim is that the frequency categories provide sufficient options to capture diet, as research shows that portion sizes vary less than frequency of intake, and the variation in intake of any food is more explained by frequency. Thus, research on dietary assessment methods indicates that separate questions on portion sizes are not superior to specifying a portion size – either a natural portion size or typical portion size – as part of the question on frequency of intake of the particular food.