A comparison of two methods to ascertain dietary intake: the CARDIA Study

Data on dietary intake were collected in the Coronary Artery Risk Development in Young Adults (CARDIA) Study at the baseline examination in 1985-86 and again at the second examination 2 years later. At baseline, a diet history questionnaire developed for the CARDIA study was used; at the second exam the NCI (Block) food frequency questionnaire was used. The purpose of the present report is to compare the estimated nutrient intakes obtained with the two instruments; to compare correlations of nutrient intakes obtained at the two exams with those observed for other lifestyle and physiological variables also measured 2 years apart; and to assess ability to test hypotheses relating 2-year changes in risk factors to between-exam differences in reported nutrient intakes. Mean levels of reported intake were generally greater for both blacks and whites on the CARDIA diet history than on the Block food frequency. Rank order correlations of reported nutrient intakes between the two questionnaires indicated greater consistency between instruments for whites (r's ranging between 0.35 and 0.52) than for blacks (r's ranging between 0.29 and 0.45). Correlations over time for nutrients were smaller than those observed for body size measures and lipid levels but were similar in magnitude to those for blood pressure, physical activity, and life events. At both exams, total caloric intake was positively associated with physical activity (range of r's for CARDIA were 0.07 for white women to 0.23 for black men, the range of r's for Block were 0.06 for women to 0.11 for white men). Using data from the two examinations, 2-year changes in total plasma cholesterol were significantly related to 2 year changes in Keys scores. The results of this comparison are useful in that they show similarities and differences between two instruments developed to gather dietary intake data. The study also illustrates the need to monitor young adults during a time when rapid changes occur in many lifestyle and physiologic factors.