Multiple lines of evidence suggest that diet during childhood and adolescence may be related to the risk of cancer decades later. High caloric intake during childhood has been found to increase the risk of multiple non-smoking related malignancies in small cohort studies. Similarly, high fruit consumption has been linked to a decreased mortality from all cancers. Height has been positively associated with the risk of several cancers. More recently, using retrospective assessments of diet, some authors have found adolescent consumption of some foods and nutrients, such as vegetable fats and vitamin E, to be related to the risk of developing breast cancer. However, the usefulness of these findings relies on the validity of recall of childhood and adolescent diet decades later. The major constraint in evaluating the validity of dietary recall is the absence of records of the original diet as reference for validation studies. The Fels Longitudinal Study provides a unique opportunity to validate instruments of dietary recall. Seven-day diet records from birth until age 18 years are available for the study participants and most of the participants and their parents remain in contact with the investigators. We will ask the mother to recall the member's childhood diet and the participants of the Fels study to recall their adolescence diet using food frequency questionnaires (FFQs) and compare the reports with the original records of their diet kept decades earlier. Thus we can evaluate the validity of instruments assessing diet decades earlier and their usefulness for epidemiologic studies. We will calculate de-attenuated correlation coefficients between the diet records and the FFQs as a measure of agreement between the instruments. The method of triads will be used to evaluate the validity for selected nutrients. We will also evaluate the ability of the FFQs to correctly rank individuals in terms of their food and nutrient intake. Lastly we will determine whether the validity of recall depends on characteristics such as time between diet record and FFQ, number of children, gender of the subject, and nutritional status.
| Chavarro, Jorge E; Rosner, Bernard A; Sampson, Laura et al. (2009) Validity of adolescent diet recall 48 years later. Am J Epidemiol 170:1563-70 |
| Chavarro, Jorge E; Michels, Karin B; Isaq, Sheherazadh et al. (2009) Validity of maternal recall of preschool diet after 43 years. Am J Epidemiol 169:1148-57 |
