Cronan McNamara explains how Data Analytics in the Food Industry is a Team Sport at the Food & Drink Business Europe IT Summit in Birmingham.
EFSA Publishes Creme Global Report on EU Consumption Database
EFSA has recently published an external scientific report, written by Creme Global (authors: Cian O’Mahony and Giulia Vilone). The report details a statistical model that was used to generate a database of food consumption at a pan-European level, as opposed to the existing data available at individual European country level.
The full report can be viewed and downloaded here (pdf).
The model generated a database of 28-day food diaries for over 40,000 simulated European consumers. The project involved a high level of mathematical and computational analysis to produce this unique database.
One of the main incentives for the project was that food intake data in the EU lacks any kind of consistency between countries. Creme Global carries out a significant number of projects in the United States, working with industry and regulators on risk assessment models for foods consumed by the American population. Key to this work is food intake data, and the main source of this in the US is the National Health and Nutrition Examination Survey (NHANES). This has a harmonised survey methodology and is carried out regularly to assess the health and nutritional status of the US population. My American colleagues often ask the question “How do they do it in Europe?”, to which I answer “Every way you could possibly imagine!”
Currently, dietary surveys in Europe exist in a huge variety of shapes and sizes because each country in the EU has a unique way of doing things. Surveys vary in lengths (1-7 days), and the numbers of consumers (anything between a few hundred to thousands of consumers). Some use consecutive days of consumption while others use non-consecutive days. These differences mean that it is very difficult to develop a picture of “European” consumption because it is not statistically valid to combine these surveys.
The idea of developing a “Compiled European Food Consumption Database” was to combine all the data using a statistical model that would overcome all these differences and could be used to estimate pan-European chronic exposure, i.e. over a 28-day period. By adapting the model to the variation and shortcomings in the data, we were able to generate a very realistic picture of Europe wide consumption.
If you are interested in the report and have any questions please contact us. Many thanks to the Creme Global Team who helped out on this work and report and to Fanny Heraud from EFSA who worked closely with us on developing the model.