The reliability and validity of short online questionnaires to measure fruit and vegetable intake in adults : the Fruit Test and Vegetable Test

The first aim of this study was to investigate the stability of the Fruit Test and Vegetable Test over time and whether the Fruit Test and Vegetable Test are capable of measuring fruit and vegetable intake with consistency. Second, the study aimed to examine criterion (concurrent) validity of the Fruit Test and Vegetable Test by testing their agreement with 7-day food diary-derived measures of fruit and vegetable intake. In total 58 adults (31% male, mean age = 30.0±12.09y) completed the Flemish Fruit and Vegetable test by indicating the frequency of days that they ate fruit and vegetables and the number of portions during the past week. Validity was tested by using a 7-day food diary as a golden standard. Adults were asked to register their fruit and vegetable intake daily in a diary during one week. Spearman correlations were measured to compare total intake reported in the Fruit and Vegetable Test and in the 7-day diary. Agreement plots were used to illustrate absolute agreement. Test-retest reliability was evaluated by having participants completing the Fruit Test and Vegetable Test twice. The Fruit Test (ICC = 0.81) and Vegetable Test (ICC = 0.78) showed excellent and substantial reliability. The Fruit Test (ρ = 0.73) and Vegetable Test showed good validity. Agreement plots showed modest variability in differences between vegetable and fruit intake as measured by the Vegetable and Fruit Test and the 7-day food diary. Also a small underestimation of fruit intake in the Fruit test and vegetable intake in the Vegetable test against the 7-day food diary was shown. Based on the results, it is suggested to include portion size pictures and consumption of mixed vegetables to prevent underestimation. To prevent overestimation, it is concluded to add a moderate number of representative fruit and vegetable items, questions on portion size, household sizes with sufficient detail and food items highly tailored to the dietary behaviors and local food items of the population surveyed. The questionnaires can easily be adapted for the use in other diets (e.g. Asian diet), but reliability and validity should then be examined again. Also, validity remains to be tested in other population groups (i.e. low socio economic status groups, other age groups)