Repeatability of estimates of nutrient and energy intake: The quantitative food frequency approach

One aspect of the quality of a measurement is its repeatability. This study addresses the repeatability of estimates of nutrient and energy intake derived from a quantitative food frequency questionnaire. One hundred individuals who, as controls in case-control studies of breast disease, had completed food questionnaires relating to usual intake, completed a second such questionnaire on re-contact approximately one year later. While, in general, the estimates of nutrient and energy intake derived from the repeat questionnaires were lower than the earlier (original) estimates, the two estimates were moderately correlated, indicating some stability over a one year period in the ranking of individuals within a population with respect to dietary intake. When the repeat estimates, categorized into three levels, were compared with the original estimates, the minority of individuals who were not classified similarly on the two occasions were as likely to have repeat estimate higher as they were to have one lower than their original estimate, indicative of non-systematic misclassification. The results indicate that the food frequency questionnaire represents a convenient technique for the measurement of dietary intake in epidemiological studies.