Reliability and comparability of three dietary assessment methods for estimating fruit and vegetable intakes

Although fruits and vegetables have been evaluated in numerous epidemiologic studies, few validation studies have examined fruits and vegetables. We examined the reproducibility and comparability of fruit and vegetable intakes estimated by diet records, food frequency questionnaires, and modules (brief food frequency questionnaires) in 101 control participants of a 1-year diet intervention trial. For each method, mean intakes at baseline and 3 months were generally within 0.3 serving per day for juice, fruits, vegetables, and total fruits and vegetables. In addition, Pearson correlations for the two time periods generally exceeded 0.55 for these four groups for each method. We evaluated comparability of intakes for 15 days of diet records, 1-year food frequency questionnaires, and modules, respectively. Mean total fruit and vegetable intakes were 6.3, 6.5, and 3.8 servings per day for diet records, food frequency questionnaires, and modules. For each pair-wise combination of methods, Pearson correlations exceeded 0.45 for juice, fruits, and total fruits and vegetables; correlations were lower for vegetables. Exact agreement in quintile assignment was less than 45%, however. These results indicate that estimates of fruit and vegetable intakes and disease associations may differ depending on the method used to assess fruit and vegetable intake.