Objective: To evaluate the reproducibility of a modified Burke-type diet history within the context of a long-term, randomized, 29-center clinical diabetes study.
Design: Diet histories were collected by trained interviewers at the end of years 1 and 2 after subjects were randomly assigned to the intensive treatment group or the conventional treatment group. Mean daily intakes of energy, protein, carbohydrate, total fat, saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, cholesterol, and dietary fiber were calculated for each treatment group at each time period.
Subjects: The study population consisted of 139 subjects in the intensive treatment group and 128 subjects in the conventional treatment group. Ages ranged from 13 to 39 years; groups included men and women. Distribution by age, sex, race, proportion of smokers, weight reported as percent ideal body weight, and duration of IDDM were similar in both groups.
Statistical methods: Differences in nutrient intake between the conventional and intensive treatment groups at each time period were tested for significance using the Wilcoxon rank-sum test. The Wilcoxon paired differences test was used to assess changes between time periods within treatment groups. Linear agreement between repeated administrations of the diet history was evaluated using Pearson's correlation coefficient, and the extent of within-subject reproducibility was assessed by intraclass correlation.
Results: No statistically significant differences in energy and nutrient intakes were observed between the two groups at either year 1 or year 2. Within each treatment group, energy and nutrient intake differences between times were not statistically significant. Correlation coefficients between years 1 and 2 ranged from .51 for dietary fiber to .72 for dietary cholesterol; within-subject reproducibility was slightly higher.
Applications: These results demonstrate long-term reproducibility for the meal-based diet history in the DCCT population.