Title: Comparison of probabilistic approaches to estimate the initial bacterial levels as a start in exposure assessment: Escherichia coli O157:H7 in beef as an example

Authors: E. Hoornstra, J. Telman

Addresses: TNO Quality of Life, Department of Microbiology, Utrechtseweg 48, 3704HE Zeist, The Netherlands. ' TNO Science and Industry, Department of Imaging Systems, Stieltjesweg 1, Delft, 2600 AD, The Netherlands

Abstract: The initial contamination of food products with pathogens is an important risk factor in microbial risk assessments. This factor contains variability and uncertainty that can be calculated from various available monitoring results. When performing dedicated quantitative risk assessments, different methods can be chosen to transform the data into probability distribution functions. This paper describes methods for the quantification of data for the initial level of Escherichia coli O157:H7 in beef. The initial contamination is estimated by expert opinion, using a slaughterhouse model and directly from available monitoring data. Two fundamentally different statistical methods are described and applied to the data: an uncertainty approach and a variability approach. This study compares and discusses these methods with the aim of giving insight to the different choices that can be made in order to estimate the initial level as a starting point in exposure assessments.

Keywords: beef; Escherichia coli; E. coli; exposure assessment; probabilistic approach; initial bacterial levels; food contamination; microbial risk assessment; quantitative risk assessment; uncertainty; variability .

DOI: 10.1504/IJRAM.2008.018210

International Journal of Risk Assessment and Management, 2008 Vol.8 No.3, pp.243- 265

Published online: 12 May 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article