Bayesian Temporal Source Attribution of Foodborne Zoonoses: Campylobacter in Finland and Norway

Statistical source attribution approaches of food‐related zoonoses can generally be based on reported diagnosed human cases and surveillance results from different food sources or reservoirs of bacteria. The attribution model, or probabilistic classifier, can thus be based on the (sub)typing information enabling comparison between human infections and samples derived from source surveillance. Having time series of both data allows analyzing temporal patterns over time providing a repeated natural experiment. A Bayesian approach combining both sources of information over a long time series is presented in the case of Campylobacter in Finland and Norway. The full model is transparently presented and derived from the Bayes theorem. Previous statistical source attribution approaches are here advanced (1) by explicit modeling of the cases not associated with any of the sources under surveillance over time, (2) by modeling uncertain prevalence in a food source by bacteria type over time, and (3) by implementing formal model fit assessment using posterior predictive discrepancy functions. Large proportion of all campylobacteriosis can be attributed to broiler, but considerable uncertainty remains over time. The source attribution is inherently incomplete if only the sources under surveillance are included in the model. All statistical source attribution approaches should include a model fit assessment for judgment of model performance with respect to relevant quantities of interest. It is especially relevant when the model aims at a synthesis of several incomplete information sources under significant uncertainty of explanatory variables.     

Cross-Contamination During Food Preparation: A Mechanistic Model Applied to Chicken-Borne Campylobacter

We develop a model for bacterial cross-contamination during food preparation in the domestic kitchen and apply this to the case of Campylobacter-contaminated chicken breast. Building blocks of the model are the routines performed during food preparation, with their associated probabilities of bacterial transfer between food items and kitchen utensils. The model is used in a quantitative microbiological risk assessment (QMRA) of Campylobacter in the Netherlands. Using parameter values from the literature and performing elementary sensitivity analyses, we show that cross-contamination can contribute significantly to the risk of Campylobacter infection and find that cleaning frequency of kitchen utensils and thoroughness of rinsing of raw food items after preparation has more impact on cross-contamination than previously emphasized. Furthermore, we argue that especially more behavioral data on hygiene during food preparation is needed for a comprehensive Campylobacter risk assessment.     

A Comparative Exposure Assessment of Campylobacter in Ontario,Canada

To inform source attribution efforts, a comparative exposure assessment was developed to estimate the relative exposure to Campylobacter, the leading bacterial gastrointestinal disease in Canada, for 13 different transmission routes within Ontario, Canada, during the summer. Exposure was quantified with stochastic models at the population level, which incorporated measures of frequency, quantity ingested, prevalence, and concentration, using data from FoodNet Canada surveillance, the peer‐reviewed and gray literature, other Ontario data, and data that were specifically collected for this study. Models were run with @Risk software using Monte Carlo simulations. The mean number of cells of Campylobacter ingested per Ontarian per day during the summer, ranked from highest to lowest is as follows: household pets, chicken, living on a farm, raw milk, visiting a farm, recreational water, beef, drinking water, pork, vegetables, seafood, petting zoos, and fruits. The study results identify knowledge gaps for some transmission routes, and indicate that some transmission routes for Campylobacter are underestimated in the current literature, such as household pets and raw milk. Many data gaps were identified for future data collection consideration, especially for the concentration of Campylobacter in all transmission routes.     

Comparison of Six Dose-Response Models for Use with Food-Borne Pathogens

Food-related illness in the United States is estimated to affect over six million people per year and cost the economy several billion dollars. These illnesses and costs could be reduced if minimum infectious doses were established and used as the basis of regulations and monitoring. However, standard methodologies for dose-response assessment are not yet formulated for microbial risk assessment. The objective of this study was to compare dose-response models for food-borne pathogens and determine which models were most appropriate for a range of pathogens. The statistical models proposed in the literature and chosen for comparison purposes were log-normal, log-logistic, exponential, -Poisson and Weibull-Gamma. These were fit to four data sets also taken from published literature, Shigella flexneri, Shigella dysenteriae,Campylobacter jejuni, and Salmonella typhosa, using the method of maximum likelihood. The Weibull-gamma, the only model with three parameters, was also the only model capable of fitting all the data sets examined using the maximum likelihood estimation for comparisons. Infectious doses were also calculated using each model. Within any given data set, the infectious dose estimated to affect one percent of the population ranged from one order of magnitude to as much as nine orders of magnitude, illustrating the differences in extrapolation of the dose response models. More data are needed to compare models and examine extrapolation from high to low doses for food-borne pathogens.     

A Modular Bayesian Salmonella Source Attribution Model for Sparse Data

Several statistical models for salmonella source attribution have been presented in the literature. However, these models have often been found to be sensitive to the model parameterization, as well as the specifics of the data set used. The Bayesian salmonella source attribution model presented here was developed to be generally applicable with small and sparse annual data sets obtained over several years. The full Bayesian model was modularized into three parts (an exposure model, a subtype distribution model, and an epidemiological model) in order to separately estimate unknown parameters in each module. The proposed model takes advantage of the consumption and overall salmonella prevalence of the studied sources, as well as bacteria typing results from adjacent years. The latter were used for a smoothed estimation of the annual relative proportions of different salmonella subtypes in each of the sources. The source‐specific effects and the salmonella subtype‐specific effects were included in the epidemiological model to describe the differences between sources and between subtypes in their ability to infect humans. The estimation of these parameters was based on data from multiple years. Finally, the model combines the total evidence from different modules to proportion human salmonellosis cases according to their sources. The model was applied to allocate reported human salmonellosis cases from the years 2008 to 2015 to eight food sources.     

Quantitative Risk Assessment from Farm to Fork and Beyond: A Global Bayesian Approach Concerning Food-Borne Diseases

A novel approach to the quantitative assessment of food-borne risks is proposed. The basic idea is to use Bayesian techniques in two distinct steps: first by constructing a stochastic core model via a Bayesian network based on expert knowledge, and second, using the data available to improve this knowledge. Unlike the Monte Carlo simulation approach as commonly used in quantitative assessment of food-borne risks where data sets are used independently in each module, our consistent procedure incorporates information conveyed by data throughout the chain. It allows "back-calculation" in the food chain model, together with the use of data obtained "downstream" in the food chain. Moreover, the expert knowledge is introduced more simply and consistently than with classical statistical methods. Other advantages of this approach include the clear framework of an iterative learning process, considerable flexibility enabling the use of heterogeneous data, and a justified method to explore the effects of variability and uncertainty. As an illustration, we present an estimation of the probability of contracting a campylobacteriosis as a result of broiler contamination, from the standpoint of quantitative risk assessment. Although the model thus constructed is oversimplified, it clarifies the principles and properties of the method proposed, which demonstrates its ability to deal with quite complex situations and provides a useful basis for further discussions with different experts in the food chain.     

A Poultry-Processing Model for Quantitative Microbiological Risk Assessment

A poultry-processing model for a quantitative microbiological risk assessment (QMRA) of campylobacter is presented, which can also be applied to other QMRAs involving poultry processing. The same basic model is applied in each consecutive stage of industrial processing. It describes the effects of inactivation and removal of the bacteria, and the dynamics of cross-contamination in terms of the transfer of campylobacter from the intestines to the carcass surface and the environment, from the carcasses to the environment, and from the environment to the carcasses. From the model it can be derived that, in general, the effect of inactivation and removal is dominant for those carcasses with high initial bacterial loads, and cross-contamination is dominant for those with low initial levels. In other QMRA poultry-processing models, the input-output relationship between the numbers of bacteria on the carcasses is usually assumed to be linear on a logarithmic scale. By including some basic mechanistics, it is shown that this may not be realistic. As nonlinear behavior may affect the predicted effects of risk mitigations; this finding is relevant for risk management. Good knowledge of the variability of bacterial loads on poultry entering the process is important. The common practice in microbiology to only present geometric mean of bacterial counts is insufficient: arithmetic mean are more suitable, in particular, to describe the effect of cross-contamination. The effects of logistic slaughter (scheduled processing) as a risk mitigation strategy are predicted to be small. Some additional complications in applying microbiological data obtained in processing plants are discussed.     

Harnessing the Theoretical Foundations of the Exponential and Beta‐Poisson Dose‐Response Models to Quantify Parameter Uncertainty Using Markov Chain Monte Carlo

Dose‐response models are the essential link between exposure assessment and computed risk values in quantitative microbial risk assessment, yet the uncertainty that is inherent to computed risks because the dose‐response model parameters are estimated using limited epidemiological data is rarely quantified. Second‐order risk characterization approaches incorporating uncertainty in dose‐response model parameters can provide more complete information to decisionmakers by separating variability and uncertainty to quantify the uncertainty in computed risks. Therefore, the objective of this work is to develop procedures to sample from posterior distributions describing uncertainty in the parameters of exponential and beta‐Poisson dose‐response models using Bayes's theorem and Markov Chain Monte Carlo (in OpenBUGS). The theoretical origins of the beta‐Poisson dose‐response model are used to identify a decomposed version of the model that enables Bayesian analysis without the need to evaluate Kummer confluent hypergeometric functions. Herein, it is also established that the beta distribution in the beta‐Poisson dose‐response model cannot address variation among individual pathogens, criteria to validate use of the conventional approximation to the beta‐Poisson model are proposed, and simple algorithms to evaluate actual beta‐Poisson probabilities of infection are investigated. The developed MCMC procedures are applied to analysis of a case study data set, and it is demonstrated that an important region of the posterior distribution of the beta‐Poisson dose‐response model parameters is attributable to the absence of low‐dose data. This region includes beta‐Poisson models for which the conventional approximation is especially invalid and in which many beta distributions have an extreme shape with questionable plausibility.     

A Quantitative Microbial Risk Assessment Model for Legionnaires'' Disease: Animal Model Selection and Dose-Response Modeling

Legionnaires' disease (LD), first reported in 1976, is an atypical pneumonia caused by bacteria of the genus Legionella, and most frequently by L. pneumophila (Lp). Subsequent research on exposure to the organism employed various animal models, and with quantitative microbial risk assessment (QMRA) techniques, the animal model data may provide insights on human dose-response for LD. This article focuses on the rationale for selection of the guinea pig model, comparison of the dose-response model results, comparison of projected low-dose responses for guinea pigs, and risk estimates for humans. Based on both in vivo and in vitro comparisons, the guinea pig (Cavia porcellus) dose-response data were selected for modeling human risk. We completed dose-response modeling for the beta-Poisson (approximate and exact), exponential, probit, logistic, and Weibull models for Lp inhalation, mortality, and infection (end point elevated body temperature) in guinea pigs. For mechanistic reasons, including low-dose exposure probability, further work on human risk estimates for LD employed the exponential and beta-Poisson models. With an exposure of 10 colony-forming units (CFU) (retained dose), the QMRA model predicted a mild infection risk of 0.4 (as evaluated by seroprevalence) and a clinical severity LD case (e.g., hospitalization and supportive care) risk of 0.0009. The calculated rates based on estimated human exposures for outbreaks used for the QMRA model validation are within an order of magnitude of the reported LD rates. These validation results suggest the LD QMRA animal model selection, dose-response modeling, and extension to human risk projections were appropriate.     

A Practical Framework for the Construction of a Biotracing Model: Application to Salmonella in the Pork Slaughter Chain

A novel purpose of the use of mathematical models in quantitative microbial risk assessment (QMRA) is to identify the sources of microbial contamination in a food chain (i.e., biotracing). In this article we propose a framework for the construction of a biotracing model, eventually to be used in industrial food production chains where discrete numbers of products are processed that may be contaminated by a multitude of sources. The framework consists of steps in which a Monte Carlo model, simulating sequential events in the chain following a modular process risk modeling (MPRM) approach, is converted to a Bayesian belief network (BBN). The resulting model provides a probabilistic quantification of concentrations of a pathogen throughout a production chain. A BBN allows for updating the parameters of the model based on observational data, and global parameter sensitivity analysis is readily performed in a BBN. Moreover, a BBN enables “backward reasoning” when downstream data are available and is therefore a natural framework for answering biotracing questions. The proposed framework is illustrated with a biotracing model of Salmonella in the pork slaughter chain, based on a recently published Monte Carlo simulation model. This model, implemented as a BBN, describes the dynamics of Salmonella in a Dutch slaughterhouse and enables finding the source of contamination of specific carcasses at the end of the chain.     

A Risk Assessment Model for Campylobacter in Broiler Meat

A quantitative microbiological risk assessment model describes the transmission of Campylobacter through the broiler meat production chain and at home, from entering the processing plant until consumption of a chicken breast fillet meal. The exposure model is linked to a dose-response model to allow estimation of the incidence of human campylobacteriosis. The ultimate objective of the model is to serve as a tool to assess the effects of interventions to reduce campylobacteriosis in the Netherlands. The model describes some basic mechanistics of processing, including the nonlinear effects of cross-contamination between carcasses and their leaking feces. Model input is based on the output of an accompanying farm model and Dutch count data of Campylobacters on the birds' exterior and in the feces. When processing data are lacking, expert judgment is used for model parameter estimation. The model shows that to accurately assess of the effects of interventions, numbers of Campylobacter have to be explicitly incorporated in the model in addition to the prevalence of contamination. Also, as count data usually vary by several orders of magnitude, variability in numbers within and especially between flocks has to be accounted for. Flocks with high concentrations of Campylobacter in the feces that leak from the carcasses during industrial processing seem to have a dominant impact on the human incidence. The uncertainty in the final risk estimate is large, due to a large uncertainty at several stages of the chain. Among others, more quantitative count data at several stages of the production chain are needed to decrease this uncertainty. However, this uncertainty is smaller when relative risks of interventions are calculated with the model. Hence, the model can be effectively used by risk management in deciding on strategies to reduce human campylobacteriosis.     

Application of Whole-Genome Sequences and Machine Learning in Source Attribution of Salmonella Typhimurium

Prevention of the emergence and spread of foodborne diseases is an important prerequisite for the improvement of public health. Source attribution models link sporadic human cases of a specific illness to food sources and animal reservoirs. With the next generation sequencing technology, it is possible to develop novel source attribution models. We investigated the potential of machine learning to predict the animal reservoir from which a bacterial strain isolated from a human salmonellosis case originated based on whole-genome sequencing. Machine learning methods recognize patterns in large and complex data sets and use this knowledge to build models. The model learns patterns associated with genetic variations in bacteria isolated from the different animal reservoirs. We selected different machine learning algorithms to predict sources of human salmonellosis cases and trained the model with Danish Salmonella Typhimurium isolates sampled from broilers (n = 34), cattle (n = 2), ducks (n = 11), layers (n = 4), and pigs (n = 159). Using cgMLST as input features, the model yielded an average accuracy of 0.783 (95% CI: 0.77–0.80) in the source prediction for the random forest and 0.933 (95% CI: 0.92–0.94) for the logit boost algorithm. Logit boost algorithm was most accurate (valid accuracy: 92%, CI: 0.8706–0.9579) and predicted the origin of 81% of the domestic sporadic human salmonellosis cases. The most important source was Danish produced pigs (53%) followed by imported pigs (16%), imported broilers (6%), imported ducks (2%), Danish produced layers (2%), Danish produced cattle and imported cattle (<1%) while 18% was not predicted. Machine learning has potential for improving source attribution modeling based on sequence data. Results of such models can inform risk managers to identify and prioritize food safety interventions.     

A Dynamic Model to Assess Microbial Health Risks Associated with Beneficial Uses of Biosolids

There is increasing interest in the development of a microbial risk assessment methodology for regulatory and operational decision making. This document presents a methodology for assessing risks to human health from pathogen exposure using a population-based model that explicitly accounts for properties unique to an infectious disease process, specifically secondary transmission and immunity. To demonstrate the applicability of this risk-based method, numerical simulations were carried out for a case study example in which the route of exposure was direct consumption of biosolids-amended soil and the pathogen present in the soil was enterovirus. The output from the case study yielded a decision tree that differentiates between conditions in which the relative risk from biosolids exposure is high and those conditions in which the relative risk from biosolids is low. This decision tree illustrates the interaction among the important factors in quantifying risk. For the case study example, these factors include biosolids treatment processes, the pathogen shedding rate of infectious individuals, secondary transmission, and immunity. Further refinement in methods for determining biosolids exposures under field conditions would certainly increase the utility of these approaches.     

Cooking Practices in the Kitchen—Observed Versus Predicted Behavior

Cross-contamination and undercooking are major factors responsible for campylobacteriosis and as such should be incorporated in microbiological risk assessment. A previous paper by van Asselt et al. ^{( 1 )} quantified cross-contamination routes from chicken breast fillet via hand, cutting board, and knife ending up in a prepared chicken-curry salad in the domestic kitchen. The aim of the current article was to validate the obtained transfer rates with consumer data obtained by video observations and microbial analyses of a home prepared chicken-curry salad. Results showed a wide range of microbial contamination levels in the final salad, caused by various cross-contamination practices and heating times varying from 2'44" to 41'30". Model predictions indicated that cooking times should be at least 8 minutes and cutting boards need to be changed after cutting raw chicken in order to obtain safe bacterial levels in the final salad. The model predicted around 75% of the variance in cross-contamination behavior. Accuracy of the model can further be improved by including other cross-contamination routes besides hands, cutting boards, and knives. The model proved to be fail-safe, which implies it can be used as a worst-case estimate to assess the importance of cross-contamination in the home.     

A Probabilistic Transmission Model of Salmonella in the Primary Broiler Production Chain

Annual data from the Finnish National Salmonella Control Programme were used to build up a probabilistic transmission model of salmonella in the primary broiler production chain. The data set consisted of information on grandparent, parent, and broiler flock populations. A probabilistic model was developed to describe the unknown true prevalences, vertical and horizontal transmissions, as well as the dynamical model of infections. By combining these with the observed data, the posterior probability distributions of the unknown parameters and variables could be derived. Predictive distributions were derived for the true number of infected broiler flocks under the adopted intervention scheme and these were compared with the predictions under no intervention. With the model, the effect of the intervention used in the programme, i.e., eliminating salmonella positive breeding flocks, could be quantitatively assessed. The 95% probability interval of the posterior predictive distribution for (broiler) flock prevalence under current (1999) situation was [1.3%-17.4%] (no intervention), and [0.9%-5.8%] (with intervention). In the scenario of one infected grandparent flock, these were [2.8%-43.1%] and [1.0%-5.9%], respectively. Computations were performed using WinBUGS and Matlab softwares.     

A Linear Model for Managing the Risk of Antimicrobial Resistance Originating in Food Animals

A linear population risk model used by the U.S. Food and Drug Administration (FDA) Center for Veterinary Medicine (CVM) estimates the risk of human cases of campylobacteriosis caused by fluoroquinolone-resistant Campylobacter. Among the cases of campylobacteriosis attributed to domestically produced chicken, the fluoroquinolone resistance is assumed to result from the use of fluoroquinolones in poultry in the United States. Properties of the linear population risk model are contrasted with those of a farm-to-fork model commonly used for microbial risk assessments. The utility of the linear population model for the purpose for which it was used by CVM is discussed.     

An Integrated Model of Infection Risk in a Health-Care Environment

Certain respiratory tract infections can be transmitted by hand-to-mucous-membrane contact, inhalation, and/or direct respiratory droplet spray. In a room occupied by a patient with such a transmissible infection, pathogens present on textile and nontextile surfaces, and pathogens present in the air, provide sources of exposure for an attending health-care worker (HCW); in addition, close contact with the patient when the latter coughs allows for droplet spray exposure. We present an integrated model of pertinent source-environment-receptor pathways, and represent physical elements in these pathways as "states" in a discrete-time Markov chain model. We estimate the rates of transfer at various steps in the pathways, and their relationship to the probability that a pathogen in one state has moved to another state by the end of a specified time interval. Given initial pathogen loads on textile and nontextile surfaces and in room air, we use the model to estimate the expected pathogen dose to a HCW's mucous membranes and respiratory tract. In turn, using a nonthreshold infectious dose model, we relate the expected dose to infection risk. The system is illustrated with a hypothetical but plausible scenario involving a viral pathogen emitted via coughing. We also use the model to show that a biocidal finish on textile surfaces has the potential to substantially reduce infection risk via the hand-to-mucous-membrane exposure pathway.     

Food Safety in the Domestic Environment: The Effect of Consumer Risk Information on Human Disease Risks

The improvement of food safety in the domestic environment requires a transdisciplinary approach, involving interaction between both the social and natural sciences. This approach is applied in a study on risks associated with Campylobacter on broiler meat. First, some web-based information interventions were designed and tested on participant motivation and intentions to cook more safely. Based on these self-reported measures, the intervention supported by the emotion "disgust" was selected as the most promising information intervention. Its effect on microbial cross-contamination was tested by recruiting a set of participants who prepared a salad with chicken breast fillet carrying a known amount of tracer bacteria. The amount of tracer that could be recovered from the salad revealed the transfer and survival of Campylobacter and was used as a measure of hygiene. This was introduced into an existing risk model on Campylobacter in the Netherlands to assess the effect of the information intervention both at the level of exposure and the level of human disease risk. We showed that the information intervention supported by the emotion "disgust" alone had no measurable effect on the health risk. However, when a behavioral cue was embedded within the instruction for the salad preparation, the risk decreased sharply. It is shown that a transdisciplinary approach, involving research on risk perception, microbiology, and risk assessment, is successful in evaluating the efficacy of an information intervention in terms of human health risks. The approach offers a novel tool for science-based risk management in the area of food safety.     

An Ecological Model for Quantitative Risk Assessment for Schistosomiasis: The Case of a Patchy Environment in the Coastal Tropical Area of Northeastern Brazil

We developed a stochastic model for quantitative risk assessment for the Schistosoma mansoni (SM) parasite, which causes an endemic disease of public concern. The model provides answers in a useful format for public health decisions, uses data and expert opinion, and can be applied to any landscape where the snail Biomphalaria glabrata is the main intermediate host (South and Central America, the Caribbean, and Africa). It incorporates several realistic and case‐specific features: stage‐structured parasite populations, periodic praziquantel (PZQ) drug treatment for humans, density dependence, extreme events (prolonged rainfall), site‐specific sanitation quality, environmental stochasticity, monthly rainfall variation, uncertainty in parameters, and spatial dynamics. We parameterize the model through a real‐world application in the district of Porto de Galinhas (PG), one of the main touristic destinations in Brazil, where previous studies identified four parasite populations within the metapopulation. The results provide a good approximation of the dynamics of the system and are in agreement with our field observations, i.e., the lack of basic infrastructure (sanitation level and health programs) makes PG a suitable habitat for the persistence and growth of a parasite metapopulation. We quantify the risk of SM metapopulation explosion and quasi‐extinction and the time to metapopulation explosion and quasi‐extinction. We evaluate the sensitivity of the results under varying scenarios of future periodic PZQ treatment (based on the Brazilian Ministry of Health's plan) and sanitation quality. We conclude that the plan might be useful to slow SM metapopulation growth but not to control it. Additional investments in better sanitation are necessary.