LEARNING OUTCOME: To describe two methods used to estimate Food Guide Pyramid servings

Nutrition researchers are increasingly using the Food Guide Pyramid (FGP) as a standard to evaluate dietary intake data. Various methods have been described for evaluating FGP servings but little information is available on the comparability of methods within the same group of individuals. Therefore, we estimated the FGP servings in a single group of women (n = 156) using both a food frequency questionnaire (FFQ) and three 24-hour dietary recalls. Women (age >60 years) were recruited from a medical center clinic in rural Pennsylvania. FFQ data were collected by telephone using Health History and Habits Questionnaire Diet Analysis System software (National Cancer Institute). Following the FFQ interviews, three 24-hour diet recalls were collected by telephone on randomly assigned days using the Nutrition Data System (Nutrition Coordinating Center, University of Minnesota). FFQ data were categorized into food groups by converting frequency of consumption to daily intakes using a medium serving as the standard for FGP servings. Recall data were categorized into FGP servings by editing nutrient and food summary data generated within NDS; foods listed as ingredients were aggregated into whole foods and assigned to FGP servings with the appropriate serving sizes. The mean number of FGP servings from each of six groups: breads, vegetables, fruits, dairy, meats, and fats, oils and sweets were calculated from both methods. Paired t-tests were used to assess differences between methods. The difference between methods were statistically significant for all food groups (p < 0.05) except the vegetable group. Food group analysis using FFQ data provide higher estimates of FGP servings from the bread (1.1 servings), the meat group (0.5 servings) and fats, oils, and sweets (1.9 servings). Analyses from dietary recall data provided higher estimates for fruits (2.5 servings). These results suggest that estimates of the number of FGP servings differs depending on the method used to assess dietary intake. More research is needed to determine possible reasons for these differences