A comparison of risk assessments on Campylobacter in broiler meat

Can stochastic consumer phase models in QMRA be simplified to a single factor?

“…Scenario 3 simulates, for example, implementation of protocols where heavily contaminated meat products are diverted from the fresh broiler meat production chain to reduce Campylobacter, which results in a decrease of the standard deviation (Nauta et al, 2009;. Scenario 4 could, for example, represent a situation of protocols where less control samples are taken, which results in more variation in the concentration of the pathogen in food, and increases the standard deviation.…”

“…This step is different from all the other steps of the food chain, because it is associated with high variability in human behaviour and because consumer food safety practices cannot be enforced by legislation (Nauta et al, 2009). Next, representative quantitative data on consumer food handling practices are scarce.…”

“…This is due to the difficulty of obtaining unbiased, representative data on human behaviour in the domestic setting in general (Redmond & Griffith, 2003), which restrains research in the consumer phase. Also, there is a high variability in food handling practices, and cultural and social differences between the considered population groups in relation to food preparation practices need to be accounted for (Nauta et al, 2009). Despite these challenges, a consumer phase model (CPM) always needs to be included in a risk assessment: to allow an evaluation of the effectiveness of intervention measures in food production and processing in terms of human health risk, the step from contaminated food product to exposure has to be taken and always includes a consumer that handles the food product.…”

“…The study from Nauta et al (2009) suggests that "there is no alternative but for a probabilistic approach to risk assessment models of the consumer phase". This would imply that one should always use a stochastic CPM to obtain accurate risk estimates in QMRA.…”

“…However, it is stated in the study that the assumption of an identical effect of consumer food handling for all serving is unlikely to be correct (Duarte et al, 2016). This is based on the fact that for example studies for Campylobacter on broiler meat have shown that there is variation present in transfer and survival during the consumer phase, and that it is essential to consider this in a risk assessment (Duarte et al, 2016;Nauta and Christensen, 2011;Nauta et al, 2009). However, the performance of this surrogate model in comparison to a stochastic CPM has never been evaluated.…”