Estimating Individual Sample Residue Distributions from Composite Sample Residue Distributions. L. M. Barraj, K. D. Tucker, and J. R. Tomerlin, Novigen Sciences, Inc., Washington DC
Food and agricultural commodity samples collected by monitoring programs or field trials are generally analyzed as composite samples. Assessment of the potential acute exposure to a specific contaminant require information about the distribution of residues in individual units of food for those foods likely to be consumed as single-serving, e.g., raw apples, baked potato etc. The need for deriving realistic estimates of these distributions is accentuated by the fact that FQPA requires assessment of both the aggregate exposure to a given compound (i.e., exposure from all sources), as well as the cumulative exposure to compounds with similar toxicity. It thus becomes necessary either to analyze these foods as individual units or, alternatively, to use the information from the distribution residues in the composite samples to predict that in single units. Three methods for estimating the distribution of residues in individual units are discussed. All three methods assume that the residue levels in individual units are lognormal, but differ in the number of underlying distributions they assume, in the treatment of samples with non-detectable residues and in the method used to estimate the parameters of the distribution(s). The methods are compared using actual data from a national monitoring program as well as simulated data. The impact of the type of data, the amount of variability between and within the parent distributions, and the number of single servings per composite sample on the methods performance was assessed. The differences between the three methods were most accentuated at the upper tails of the estimated residue distributions and associated exposure distributions.
Go to . . .
1999 SRA Table of Contents
1999 SRA Author Index
Main Abstracts Menu Page
RiskWorld Home Page