View Report Details
Detecting and modelling MIS-reporting of food intake in adults
Project Code: N08001
Background and objectives
The recent development of new approaches to objectively measure total energy expenditure (EE) has revealed that reporting of energy intake (EI) in study groups usually falls short of actual EE. This implies that subjects are in a negative energy balance (EB) or that they are mis-reporting their EI. Based on assumptions of energy balance it is now generally accepted that so called under-reporting or mis-reporting (MR) is widespread.
Initial attempts were made to establish energy intake to basal metabolic rate (EI:BMR) ratio at which intakes can be deemed implausibly low. These cut-offs have been called the Goldberg cut-offs. Their use has been widespread in attempting to identify and exclude mis-reporters from analysis of diet surveys.
The prevalence of mis-reporting of food intake seems so high that to either include or exclude mis-reporters from a data-set will distort data regarding dietary intakes. Because mis-reporting of dietary intakes has never been directly quantified, its true nature and extent is largely unknown. We have therefore attempted to detect and if possible, model mis-reporting under carefully controlled laboratory conditions, which was then related to diet survey conditions.
This is the first project to directly measure the nature and extent of mis-reporting. The original objectives were to examine current means of detecting mis-reporting and to conduct 3 separate and ambitious studies. We used the diet survey literature, the specialised facilities of the Rowett and the statistical capabilities of Biomathematics and Statistics, Scotland to (i) examine current limitations to the use of the Goldberg cut-off to identify mis-reporters (ii) to precisely quantify the nature and extent of MR in a laboratory-simulation of the domestic feeding environment and relate this to mis-reporting in free-living subjects (iii) use the information from (i) and (ii) to derive a model, which predicts nature and extent of mis-reporting (iv) test this model in a larger scale diet-survey study (v) evaluate the robusticity of the model for use in formulating correction factors for mis-reporting in large scale diet surveys.
Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: