A procedure for grouping food consumption data for use in food allergen risk assessment

• Published in: Journal of food composition and analysis Vol. 59; pp. 111 - 123
• Main Authors: Birot, Sophie, Madsen, Charlotte B., Kruizinga, Astrid G., Christensen, Tue, Crépet, Amélie, Brockhoff, Per B.
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01-06-2017; Elsevier
• Subjects: Allergen contamination; Automatic procedure; Food allergy; Food analysis; Food and Nutrition; Food composition; Food groups; Life Sciences; National Food Consumption Surveys; Probabilistic risk assessment; Allergen contamination; Food allergy; Food groups; Probabilistic risk assessment; Food composition; National Food Consumption Surveys; Food analysis; Automatic procedure; Allergens; Probabilistic Models; Food Consumption
• ISSN: 0889-1575; 1096-0481
• DOI: 10.1016/j.jfca.2017.01.008

•Customized clustering method followed by allergen risk assessment per food group.•Food groups automatically created for France, Netherlands and Denmark.•Danish consumption data available per food group created with the procedure.
Food allergic subjects need to avoid the allergenic food that triggers their allergy. However, foods can also contain unintended allergens. Food manufacturers or authorities need to perform a risk assessment to be able to decide if unintended allergen presence constitutes a risk to food allergic consumers. One of the input parameters in risk assessment is the amount of a given food consumed in a meal. There has been little emphasis on how food consumption data can be used in food allergen risk assessment. The aim of the study was to organize the complex datasets from National Food Consumption Surveys from different countries (France, Netherlands and Denmark) to be manageable in food allergen risk assessment. To do this, a two-step method was developed. First, based on initial groups of similar food items, the homogeneity of consumption was evaluated using a customized clustering method. Then, the risk was calculated for each initial food group and its subgroups to verify if it also represents a relevant difference in risk. Forty-eight food groups were designated in Denmark (53 in the Netherlands, 54 in France). Finally, summary statistics and names for each food group for the Danish data illustrate the results when applying the procedure.