Development and validation of a computerized web-based quantitative food frequency questionnaire

Epidemiological research is progressing towards digital data collection. This study aimed to evaluate the validity and reproducibility of our new computerized, and easy-to-use Food Frequency Questionnaire (FFQ).Participants' dietary intake was assessed using 24-h Dietary Recalls (DRs) and our FFQ, consisting of 133 food items and beverages. The software allows users to choose between three visualized portion sizes, categorizes the selections into food groups with three degrees of food processing levels, and produces a visualized output of the results. The reproducibility of the FFQ was evaluated based on two user submissions, and its validity was measured by comparing its calculated caloric intake and macro and micro-nutrient consumption to the equivalent mean values from three 24-h DRs. Thirty-nine women of fertility age [18-45] were recruited to the study, of whom twenty-six qualified for reproducibility testing and thirty-one qualified for validity testing.For most nutrient intakes, the FFQs yielded higher scores than the 24-h DRs, resulting in a less satisfactory agreement between them due to FFQs overestimation. The Intra Class Correlation (ICC) coefficient between the two FFQs ranged from moderate for calcium (0.55) to high for magnesium (0.83) (p 0.05), indicating good reproducibility. Evaluation of food groups and processed food reproducibility scores yielded ICC coefficients ranging from moderate (0.53; super-processed foods) to high (0.83; non-processed foods) (p 0.05). Spearman's correlation coefficient showed a moderate (sugar-sweetened beverages) to strong (non-processed foods) correlation (p 0.05).The FFQ we developed and validated in this study showed moderate to high reproducibility and acceptable validity in a group of thirty-nine women of fertility age. Moreover, it is highly adjustable and easy to use, and its digital-based delivery enables large-scale, multilingual nutritional research.