(PTFS17-01-26) Agreement of Food Composition Values Generated via Aggregation of Branded Data Versus Chemical Analysis- Informing the New Era of Food Composition Databases

Objectives: In a world of increasing packaged food consumption and budget cuts in central government, the use of label data to update national food composition tables presents a relevant and cost-effective approach. Aggregation of branded products into meaningful, generic food items is feasible but it remains unclear if the composition values produced are comparable to those derived via chemical analysis.

Methods: The HelTH branded food composition database, which includes data for 8,915 foods as sold in the Greek market from July to October 2023, was used to derive aggregated food composition values for n=506 commonly consumed foods. Composition data were collected from n=126 foods analysed in certified laboratories at the same period. After aggregation, a total of n=45 generic foods were identified with values derived from both label data aggregation and chemical analyses. These corresponded to n=551 unique branded foods and n=149 food samples from chemical analysis. Mann-Whitney U test was used to compare the compositional differences between the two methods of value derivation with significance set at p< 0.05.

Results: Mean absolute content in fats, protein, carbohydrates, sugars and energy derived either by label data aggregation or chemical analyses exhibited < 10% differences. Total carbohydrates and total sugars were more likely to exhibit higher differences compared to other nutrients. However, none of the differences detected reached statistical significance (p >0.05).

Conclusions: Value aggregation from label data can yield nutrient composition data for the major macronutrients comparable to those obtained by traditional chemical analysis. The use of label data allows for better representation of the variability within the proposed generic foods, providing benefits in risk assessment and composition modelling activities.

Funding Sources: This research received funding under the EUOpenFood EFSA project and the HFRI PhD Scholarships.