Dependence of dietary intake estimates on the time frame of assessment

Abstract

Food chemical risk management needs, among other things, assessment of exposure. For dietary intake food consumption surveys are the data source to be used. One complicating factor in the usage of these data is the dependence of dietary intake estimates on the time frame of assessment. Central to this time dependence is the within-subject variation regarding the usage of food products and, as a consequence, the intake of chemicals. Within-subject variation is mostly as large as or larger than between-subject variation. Expressed per kilogram body weight, average (total) variation in intake variables depends on the age group, with variation usually being greater at younger age, most likely as a result of the higher intake levels at that age. Combination of age groups results in an increase in between-subject variation, and correction based on the figures for the total population will be too small. Ideally, exposure data for all days of one's life should be available to assess lifetime exposure. Since information on all these days is not an attainable and practical option, and not an option to strive for either, the most recent available data should be used that can be extended with simulation studies to anticipate future developments. The present food consumption surveys available in European countries are based on data that vary from 1 day (24-h recall and dietary record) to habitual intake (dietary history and food frequency). The data of a survey based on 1 day refer to 0.004% of an average lifetime of 70 years. Based on the demographic picture of the population, a reasonable approximation of lifetime intake can be obtained. The proportion of users and the consumption level among users depend on the time frame of assessment, especially for irregularly consumed products. Usage of the concept of "users only" overestimates lifetime exposure of the population, the extent of overestimation depending on the duration of the survey. The likelihood that all consumers have been exposed to a chemical once during a lifetime period is realistic in the sense of the best approximation of reality. As a result of this assumption all exposure assessments will have a similar point of departure and the dependence of the results on the food consumption method will be reduced.