Better Methods for Assessing Dietary Intake

Healthy food is thought to be important in preventing many diseases of adult life, including cancer and heart disease. However, few definite links have been established due to the difficulties of measuring usual diet so that this can be linked to the development or protection from such diseases in later life. We propose to model underlying methods for an easy to use but accurate system of measuring diet which can be used in studies assessing the relationship between the food we eat and susceptibility to these diseases.

Large prospective studies are needed to test hypotheses arising from established and novel lines of investigation involving gene-diet interactions in affecting risk of chronic disease such as heart disease, cancer, digestive disorders and diabetes. However, real time dietary intake methods are associated with less measurement error than retrospective questionnaires usually used in such large scale studies. Utlising the largest set of real time data worldwide, from dietary 7 day diaries kept by 20,000 participants in the EPIC Norfolk study, we propose to model methods which can be used to develop straightforward, easy to use direct-to-data electronic diet assessment platforms. Such platforms can be used in new cohort studies, such as Biobank, research and clinical studies. In existing cohorts such a development allows the potential to repeatedly update information at very little extra cost/time to the research investigators and only a minor burden to the study participant.