Background: This study assesses the relative validity of a diet history interview in 56 free-living individuals with Type II diabetes mellitus.
Methods: The diet history interview was compared to a 3-day food record in a 1-year dietary intervention trial. The plausibility of energy intake data was examined throughout the trial. Paired data were compared for differences and the presence of systematic error was determined by Bland Altman analysis. Changes in accuracy and responsiveness were assessed over time.
Results: The proportion of underreporters was larger in the diet history at baseline. Underrecording with the food record was more common in subjects with BMIs > 30 kg/m(2). There was no difference between paired dietary data from the two methods; however, data on fatty acids failed to correlate. These correlations improved when outliers were removed. There was no evidence of a relationship between bias and mean intake of dietary variables. Accuracy of diet history measurement did not change during the trial for energy or macronutrients, but data on protein and monounsaturated fat were both affected by BMI. The diet history was more responsive than the food record to changes in monounsaturated fatty acid intake after 3 months, but this changed at the end of the trial.
Conclusion: The diet history provided good estimates of energy and macronutrient intakes in a sample group with Type II diabetes mellitus. However, energy intake data revealed a high prevalence of underreporting especially in people with higher BMIs.