Is a Single 24-hour Dietary Recall per Person Sufficient to Estimate the Population Distribution of Usual Dietary Intake?

Self-reporting dietary assessment instruments are prone to various types of measurement errors that should be taken into account when analyzing nutritional data (1). The National Cancer Institute (NCI) Dietary Assessment Primer provides a good overview of the advantages and disadvantages of the various dietary assessment instruments currently in use (2). Statistical modeling can only partially address potential measurement errors. Instruments for dietary assessment need to be appropriate to the study design and research question. In the following, we will discuss issues encountered when undertaking statistical modeling of population-based estimates of underlying distributions of dietary intake variables from foods and nutrients consumed on a nearly daily basis. It should be noted that such modeling differs from modeling of individual dietary intakes in health-outcome models or the evaluation of dietary interventions.
Source: Journal of Nutrition - Category: Nutrition Source Type: research