Exposure analysis consumption data

Probabilistic approaches take advantage of current computational capabilities to combine all of the data in a pesticide residue distribution (rather than a single expected value) with food consumption data to develop a distribution of daily exposure. This approach is called a Monte Carlo simulation, although there are many ways to conduct this type of analysis. 

This paper has treated the food consumption data from the UK almost as representative of EU food consumption data in general. Doing so, however, is mainly out of convenience at this time. The UK food consumption database is well documented, among the most readily available in electronic format, and suitable for detailed distributional analysis. The 97.5 percentile consumption estimates from the UK database have formed the basis of the NESTI acute dietary exposure estimates discussed previously [8]. In addition to the UK data, a Dutch dietary exposure model using Dutch data is under development, but is not readily available at this time [19]. The German data have also been used to estimate chronic dietary exposure [20]. 

Driver JH, Ginevan ME, and Whitmyre GK (1996) Estimation of dietary exposure to chemicals A case study illustrating methods of distributional analyses for food consumption data. Risk Analysis 16 763-771. 

In the case of a single pesticide found on a single commodity, a Monte Carlo analysis would randomly select a residue data point and a food consumption level value and multiply them together to yield an exposure level. By repeating this process, often thousands or tens of thousands of times, it is possible to develop a distribution of daily exposures that would allow a determination of which levels represent, for example, the 50th, 99th, and 99.9th percentiles of consumption. 

Diary studies are used to determine in detail the consumption of a particular part of a diet. A population consuming above-average amounts of food that provides the main source of exposure to a contaminant can be identified using questionnaires. A record of the type and weight of food eaten, and the source, is then kept in a purpose-made diary by participants in the study. Representative samples of foods eaten are then analysed and the data combined. An extension of this approach is exemplified by the duplicate diet study, in which as exact a replicate as possible of all food consumed is collected for analysis. 

Identifying the key sources of uncertainties in food consumption rates and dietary residue data during a dietary exposure and risk analysis, and subsequently examining their respective influence on model predictions. 

Tools should be developed to support the identification of mixture exposure situations that may cause unexpectedly high risks compared to the standard null models of concentration addition and response addition, for example, based on an analysis of food consumption and behavioral patterns, and the occurrence of common mixture combinations that cause synergistic effects. Criteria should be developed for the inclusion of interaction data in mixture assessments. 

Concern about the possible harmful effects of caffeine on the outcome of pregnancy has evolved mainly from studies of animals which have shown a reduction in intrauterine fetal growth. However, the implications of these data for men are unclear, because of the differences in mode of exposure to caffeine, the amounts consumed, and caffeine metabolism. The possible effects of caffeine intake on the human fetus have been reviewed (SEDA-7, 8) the conclusion was that the scientific data currently available could not answer the question. In an analysis of interview and medical record data in 12 205 non-asthmatic women to evaluate the relation between coffee consumption and adverse outcomes in pregnancy, the findings were negative. 

TTP information. As with other NYS Angler Cohort studies, limitations include the reliance on self-reported exposure data and outcome data, and the lack of information on potential confounding factors such as occupational exposures, alcohol and caffeine consumption, and current smoking status. In addition, women with unplanned pregnancies were necessarily ruled out from the analysis Buck et al. (1997) noted that this may be a potential bias inherent in the study. 

---