There is currently no harmonised way of matching food consumption derived from UK food composition tables to food groups needed to answer a specific dietary study’s research question. This study aims to utilise dietary data from three established studies (Bristol PREDIMED, STAMP-2, Early ACTID) in order to (i) harmonise food codes across the three studies to the NDNS food codes, (ii) create a universal database of food items and (iii) develop codes/syntaxes to organise universal food items into food groups. This database will initially be tested by developing Mediterranean food groups (to reflect the questions of two Mediterranean adherence questionnaires) from the three studies. In the longer term, this project will aid us to merge and compare dietary datasets and allow a uniform way of assessing intake of food groups in dietary studies. The syntaxes developed will be adaptable for other projects using different food groups.