Background The complexity of dietary assessment is increasing. Approximately 70,000 foods are available in UK supermarkets. The UK is the largest mixed ingredient/ready meal consumer in Europe; many of these foods are high in saturated fat, sugar or salt. The level of detail of existing food composition tables is not sufficient to capture this wide variety. The standard UK food tables have 3423 generic food items; even the expanded version used for the National Diet and Nutrition Survey only contains 7000 items. The growing diversity of the UK diet has not been matched by an expansion of the standard food composition tables.

Methods We have developed a new UK food composition table by linking the standard tables to food industry, supermarket and fast food outlet branded nutrient information, including date stamping allowing for future reformulation. We have utilised back of pack (BOP) nutrient data and mapped these to generic items from the standard food tables creating a database with full micronutrient data. Mapping was based on kcal, fat, protein, carbohydrate. The percentage difference between BOP and generic items was calculated for each nutrient and then added. Mapping of foods was also guided by the food description. Ingredients lists were used with some ready meals being mapped to multiple generic ingredients. BOP information is controlled by EU directives. Natural, production and storage variations mean that foods may not contain the exact nutrient levels labelled. However, the nutrient content of foods should not deviate substantially from labelled values.

Results 38,417 branded foods have had BOP nutrients mapped to combinations of generic items. Over 50% of items were mapped to within 10% agreement with the generic food item with regard to energy. 80% of foods were mapped to a single generic item with 20% to multiple items; for example, one korma was mapped to 23 ingredients. The largest food groups mapped included cakes, biscuits, chocolates and other snacks (6918 items); alcoholic drinks (5692 items); sauces and condiments (3635 items); dairy and eggs (3596 items); ready meals including pizza (3315 items).

Conclusion These detailed data have the potential to greatly increase the precision with which we assess dietary intake since the generic food tables are limited in scope and variety of foods available. This new food database is a core element of a new online dietary assessment tool ‘myfood24’.

Acknowledgement This project is funded by the MRC award G1100235; A UK on-line 24h dietary recall tool for population studies: development, validation and practical application.