Assessment of usual dietary intake in population studies of gene-diet interaction.

AIMS: Dietary intake is a critical environmental exposure when considering the effect of many genetic factors on disease risk. However, dietary intake is a complex and changing measure that requires particular care in assessment. DATA SYNTHESIS: Although weighed diet records can theoretically provide the most accurate assessment of intake, they are usually not realistic in large population studies due to heavy respondent burden, likelihood of poor compliance, and the cost of data entry. Multiple 24-h dietary recalls can provide excellent detail, allowing for diverse dietary practices, but they are costly and require multiple contacts with participants. Food frequency questionnaires are the most cost-effective tool for assessing usual intake, particularly for micronutrients with high day-to-day variability. However, they have limitations for diverse populations and recent studies have questioned their ability to measure macronutrient intakes for assessing diet and disease relationships.
CONCLUSION: At the present time, food frequencies remain the most cost-effective tool for large population studies. However, their limitations must be fully appreciated and demonstration of validity for nutrients of concern in the populations under study is essential. When macronutrients are of key interest, consideration should be given to the use of multiple recalls. Records may be used only in educated and compliant populations. Continued efforts to improve dietary assessment methodology must be investigated.