Food Standards Agency
Thursday 2 July 2009
Safer food better business banner
AZ-Directory
What's NewRSS
What is RSS?Listen
Listen to this siteFriday 10 September 2004
This research project aims to produce a statistical model which can estimate variability in dietary intake, between-person and within-person, whilst accounting for day-to-day variation.
Study Duration : August 1999 to March 2002
Contractor : Cancer Research UK
Previous models for estimating average consumption of chemicals contained in food and the levels of variability of the amount consumed on a person to person, day-to-day basis have incorporated major simplifications and assumptions regarding consumption patterns. For instance, one person may have a higher day-to-day variability of consumption of a particular product than another person, and differences in the consumption of particular foods between persons may also be influenced by more systematic factors such as age, sex or location of residence.
When there is a risk of chemical contaminants being found in food, e.g. mycotoxins or pesticide residues, it is important to know how frequently and in what amounts a person might be consuming a potentially hazardous food and the concentration of the contaminant within the food. There is a need for a model that can account for between-person variation, and within-person variation in consumption on a day-to-day basis, that can be applied in risk assessment of food chemicals.
The model aimed to have flexibility to account for real-life influences on between-person and within-person variation in consumption, and be a sophisticated improvement on current models whilst remaining simple enough to be practicable to those involved in risk assessment.
Data for modelling was taken from the Dietary and Nutritional Survey of British Adults; a study of 2197 adults who each kept a detailed diary of all food and drinks consumed over a seven-day period. As details on pesticides and mycotoxins were originally unavailable, the average calorific and retinol intake information was applied to the construction of a formal mathematical model that took account of day-to-day variability differing from person to person. The model was gradually extended to include fixed personal traits such as age and sex, and limited published data on pesticide residues and mycotoxins were eventually introduced.
Having such a complex variance structure, it was necessary to fit the model using the Bayesian technique of Markov chain Monte Carlo modelling. The accuracy of prediction of numbers of days when particular intake levels were exceeded was used as a measure of the model reliability with this data set.
The project aimed to produce a more refined statistical model than current simpler models, for estimating numbers of individuals exceeding permitted maximum levels of chemicals. This project complements T01013.
Analysis of the Dietary and Nutritional Survey of British Adults exposed significant differences in the intake of various nutrients and contaminants between individuals on a day-to-day basis. Such differences required inclusion into the model for accurate prediction of the level of consumption of specific foods and their contaminants.
Model reliability tests showed it was possible to estimate aspects of dietary consumption whilst allowing for complex patterns of variation between and within individuals. This statistical tool allowed for personal traits such as age to be taken into account for an individual, and also allows for an individual's day-to-day variation in consumption. The complex model that accounted for these subject variations out-performed more traditional models that contained simplifying assumptions regarding consumption patterns.
These methods have the potential to improve estimation of upper levels of consumption of contaminants and provide improved estimates of the numbers of individuals or periods for which stated hazardous levels are exceeded. An alternative, and potentially more robust, estimation of the likely effects of changes in food composition or consumption habits may be available in the future. The FSA is considering options for incorporating such techniques into intake calculations.
Final report is available from the FSA Library and Information centre.
To obtain a copy, please contact the Enquiry Desk, Dr Elsie Widdowson Library and Information Services, Food Standards Agency tel: 020 7276 8181/8182 or email:
library&info@foodstandards.gsi.gov.uk
The results of this project have been published:
Myles JP, Price GM, Hunter N, Day M, Duffy SW.
A potentially useful distribution model for dietary intake data.
Public Health Nutrition. 2003 Aug;6(5):513-9.
Contact
: For any enquiries concerning this research project, please contact the relevant Programme contact or email:
science@foodstandards.gsi.gov.uk
Find out what our other sites have to offer