Food Data Quality in Nutritional Surveys: which Issues are to be Tackled?

Food data (food intake or food consumption) quality is a central issue to be tackled in order to ensure a correct interpretation of the results from population studies in the nutritional field. In fact, several sources of errors can be identified in this type of study because of inherent characteristics and specific possible procedural inaccuracies. Particularly for the quantitative aspect, inherent imprecision and procedural errors (non-probabilistic) should be clearly identified in order to adopt preventive measures. Therefore, necessary quantitative data, their numerical expression and the calculation used to obtain net intake must be defined together with the detail of food section and the related food coding system in open-ended forms and the classification criteria in the closed-ended ones. Furthermore, the precision of food components intake depends upon both the food data quality and either the food composition data or the algorithms to assess the equivalent for other substances contained in the food products. Comprehensively, the data quality can be evaluated by considering the quantitative aspect (coherence of consumed quantities, percentage of under/over-reporting subjects, percentage of missing data for nutrients) and the detail of food items (percentage of “unspecified” items). Subsidiary tools for people collecting data will be useful.