Studies in Preschool Age Populations
Table 4.2 includes summary data on 12 surveys and studies on preschool populations. Table 4.3 contains more detail on each study. In the 1999-2000 NHANES and the 1994-96 CSFII, 24HR (24-Hour Recall) interviews were conducted with the preschool age child's parent or a designated proxy using the same portion size estimating aides used in interviews with adults. Parents were asked to estimate serving size based on the adult aides. If intake of a food occurred when the respondent was not present, data retrieval was conducted with the caretaker, day care center, or preschool. Table 2.4 presents the diet-related question in each of these surveys as well as in the integrated What We Eat in America-NHANES survey currently in the field. The rationale for the instrument selection was recently reviewed (64). In the integrated What We Eat in America-NHANES, preschool age children's intake is assessed through two 24HR interviews (one in-person and one by telephone), and a 100-item propensity questionnaire. Supplement use questions are included in another questionnaire (Table 2.4).
Three-day weighed FRs (Food Records) were used in both the German DONALD longitudinal survey (154) and the Adelaide Nutrition Cohort in South Australia (117). The DONALD survey applied the Goldberg ratio (EI:BMR or Energy Intake:Basal Metabolic Rate) of less than 1.06 to identify implausible intakes. Protein intakes were validated with nitrogen excretion from a 24-hour urine collection (154).
The UK ALSPAC study (112), the Framingham Nutrition Study (156), and the Iowa Fluoride Study (98;114) each employed a 3-day estimated FR in this age group. A longitudinal survey of infants and preschoolers in upper-income Tennessee households used both an annual 24HR interview and a 2-day estimated FR in 3 to 5 year olds (157). In the Bogalusa Heart Study, 24HR interviews were conducted annually in a subset of cohort members aged 3 to 4 years (118). Investigators studying a cohort of Hispanic preschoolers in New York City used multiple 24HR interviews and the Harvard FFQ paired with serum lipid levels in a series of validation studies and surveys (70;135;136).
As apparent from Table 4.2 and Table 4.3, a standardized approach to collecting supplement information was not found. Often, studies did not report whether and how supplement information was collected. Reported approaches included collecting information in the diet assessment instrument or using a separate questionnaire.
Last Modified: 11 Apr 2014