Risk Ranking for Foodborne Microbial Hazards in New Zealand: Burden of Disease Estimates

Priority setting for food safety management at a national level requires risks to be ranked according to defined criteria. In this study, two approaches (disability‐adjusted life years (DALYs) and cost of illness (COI)) were used to generate estimates of the burden of disease for certain potentially foodborne diseases (campylobacteriosis, salmonellosis, listeriosis (invasive, perinatal, and nonperinatal), infection with Shiga toxin‐producing Escherichia coli (STEC), yersiniosis, and norovirus infection) and their sequelae in New Zealand. A modified Delphi approach was used to estimate the food‐attributable proportion for these diseases. The two approaches gave a similar ranking for the selected diseases, with campylobacteriosis and its sequelae accounting for the greatest proportion of the overall burden of disease by far.     

Combining QMRA and Epidemiology to Estimate Campylobacteriosis Incidence

The disease burden of pathogens as estimated by QMRA (quantitative microbial risk assessment) and EA (epidemiological analysis) often differs considerably. This is an unsatisfactory situation for policymakers and scientists. We explored methods to obtain a unified estimate using campylobacteriosis in the Netherlands as an example, where previous work resulted in estimates of 4.9 million (QMRA) and 90,600 (EA) cases per year. Using the maximum likelihood approach and considering EA the gold standard, the QMRA model could produce the original EA estimate by adjusting mainly the dose‐infection relationship. Considering QMRA the gold standard, the EA model could produce the original QMRA estimate by adjusting mainly the probability that a gastroenteritis case is caused by Campylobacter. A joint analysis of QMRA and EA data and models assuming identical outcomes, using a frequentist or Bayesian approach (using vague priors), resulted in estimates of 102,000 or 123,000 campylobacteriosis cases per year, respectively. These were close to the original EA estimate, and this will be related to the dissimilarity in data availability. The Bayesian approach further showed that attenuating the condition of equal outcomes immediately resulted in very different estimates of the number of campylobacteriosis cases per year and that using more informative priors had little effect on the results. In conclusion, EA was dominant in estimating the burden of campylobacteriosis in the Netherlands. However, it must be noted that only statistical uncertainties were taken into account here. Taking all, usually difficult to quantify, uncertainties into account might lead to a different conclusion.     

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.     

Human Dose–Response Data for Francisella tularensis and a Dose‐ and Time‐Dependent Mathematical Model of Early‐Phase Fever Associated with Tularemia After Inhalation Exposure

Military health risk assessors, medical planners, operational planners, and defense system developers require knowledge of human responses to doses of biothreat agents to support force health protection and chemical, biological, radiological, nuclear (CBRN) defense missions. This article reviews extensive data from 118 human volunteers administered aerosols of the bacterial agent Francisella tularensis , strain Schu S4, which causes tularemia. The data set includes incidence of early‐phase febrile illness following administration of well‐characterized inhaled doses of F. tularensis . Supplemental data on human body temperature profiles over time available from de‐identified case reports is also presented. A unified, logically consistent model of early‐phase febrile illness is described as a lognormal dose–response function for febrile illness linked with a stochastic time profile of fever. Three parameters are estimated from the human data to describe the time profile: incubation period or onset time for fever; rise time of fever; and near‐maximum body temperature. Inhaled dose‐dependence and variability are characterized for each of the three parameters. These parameters enable a stochastic model for the response of an exposed population through incorporation of individual‐by‐individual variability by drawing random samples from the statistical distributions of these three parameters for each individual. This model provides risk assessors and medical decisionmakers reliable representations of the predicted health impacts of early‐phase febrile illness for as long as one week after aerosol exposures of human populations to F. tularensis .     

A Multinomial‐Dirichlet Model for Analysis of Competing Hypotheses

Analysis of competing hypothesis, a method for evaluating explanations of observed evidence, is used in numerous fields, including counterterrorism, psychology, and intelligence analysis. We propose a Bayesian extension of the methodology, posing the problem in terms of a multinomial‐Dirichlet hierarchical model. The yet‐to‐be observed true hypothesis is regarded as a multinomial random variable and the evaluation of the evidence is treated as a structured elicitation of a prior distribution on the probabilities of the hypotheses. This model provides the user with measures of uncertainty for the probabilities of the hypotheses. We discuss inference, such as point and interval estimates of hypothesis probabilities, ratios of hypothesis probabilities, and Bayes factors. A simple example involving the stadium relocation of the San Diego Chargers is used to illustrate the method. We also present several extensions of the model that enable it to handle special types of evidence, including evidence that is irrelevant to one or more hypotheses, evidence against hypotheses, and evidence that is subject to deception.     

Source Attribution of Food-Borne Zoonoses in New Zealand: A Modified Hald Model

A Bayesian approach was developed by Hald et al .^{( 1 )} to estimate the contribution of different food sources to the burden of human salmonellosis in Denmark. This article describes the development of several modifications that can be used to adapt the model to different countries and pathogens. Our modified Hald model has several advantages over the original approach, which include the introduction of uncertainty in the estimates of source prevalence and an improved strategy for identifiability. We have applied our modified model to the two major food-borne zoonoses in New Zealand, namely, campylobacteriosis and salmonellosis. Major challenges were the data quality for salmonellosis and the inclusion of environmental sources of campylobacteriosis. We conclude that by modifying the Hald model we have improved its identifiability, made it more applicable to countries with less intensive surveillance, and feasible for other pathogens, in particular with respect to the inclusion of nonfood sources. The wider application and better understanding of this approach is of particular importance due to the value of the model for decision making and risk management.     

A Bayes Linear Bayes Method for Estimation of Correlated Event Rates

Typically, full Bayesian estimation of correlated event rates can be computationally challenging since estimators are intractable. When estimation of event rates represents one activity within a larger modeling process, there is an incentive to develop more efficient inference than provided by a full Bayesian model. We develop a new subjective inference method for correlated event rates based on a Bayes linear Bayes model under the assumption that events are generated from a homogeneous Poisson process. To reduce the elicitation burden we introduce homogenization factors to the model and, as an alternative to a subjective prior, an empirical method using the method of moments is developed. Inference under the new method is compared against estimates obtained under a full Bayesian model, which takes a multivariate gamma prior, where the predictive and posterior distributions are derived in terms of well‐known functions. The mathematical properties of both models are presented. A simulation study shows that the Bayes linear Bayes inference method and the full Bayesian model provide equally reliable estimates. An illustrative example, motivated by a problem of estimating correlated event rates across different users in a simple supply chain, shows how ignoring the correlation leads to biased estimation of event rates.     

Inferring an Augmented Bayesian Network to Confront a Complex Quantitative Microbial Risk Assessment Model with Durability Studies: Application to Bacillus Cereus on a Courgette Purée Production Chain

The Monte Carlo (MC) simulation approach is traditionally used in food safety risk assessment to study quantitative microbial risk assessment (QMRA) models. When experimental data are available, performing Bayesian inference is a good alternative approach that allows backward calculation in a stochastic QMRA model to update the experts’ knowledge about the microbial dynamics of a given food‐borne pathogen. In this article, we propose a complex example where Bayesian inference is applied to a high‐dimensional second‐order QMRA model. The case study is a farm‐to‐fork QMRA model considering genetic diversity of Bacillus cereus in a cooked, pasteurized, and chilled courgette purée. Experimental data are Bacillus cereus concentrations measured in packages of courgette purées stored at different time‐temperature profiles after pasteurization. To perform a Bayesian inference, we first built an augmented Bayesian network by linking a second‐order QMRA model to the available contamination data. We then ran a Markov chain Monte Carlo (MCMC) algorithm to update all the unknown concentrations and unknown quantities of the augmented model. About 25% of the prior beliefs are strongly updated, leading to a reduction in uncertainty. Some updates interestingly question the QMRA model.     

Use of Markov Chain Monte Carlo Analysis with a Physiologically-Based Pharmacokinetic Model of Methylmercury to Estimate Exposures in U.S. Women of Childbearing Age

A Bayesian approach, implemented using Markov Chain Monte Carlo (MCMC) analysis, was applied with a physiologically‐based pharmacokinetic (PBPK) model of methylmercury (MeHg) to evaluate the variability of MeHg exposure in women of childbearing age in the U.S. population. The analysis made use of the newly available National Health and Nutrition Survey (NHANES) blood and hair mercury concentration data for women of age 16–49 years (sample size, 1,582). Bayesian analysis was performed to estimate the population variability in MeHg exposure (daily ingestion rate) implied by the variation in blood and hair concentrations of mercury in the NHANES database. The measured variability in the NHANES blood and hair data represents the result of a process that includes interindividual variation in exposure to MeHg and interindividual variation in the pharmacokinetics (distribution, clearance) of MeHg. The PBPK model includes a number of pharmacokinetic parameters (e.g., tissue volumes, partition coefficients, rate constants for metabolism and elimination) that can vary from individual to individual within the subpopulation of interest. Using MCMC analysis, it was possible to combine prior distributions of the PBPK model parameters with the NHANES blood and hair data, as well as with kinetic data from controlled human exposures to MeHg, to derive posterior distributions that refine the estimates of both the population exposure distribution and the pharmacokinetic parameters. In general, based on the populations surveyed by NHANES, the results of the MCMC analysis indicate that a small fraction, less than 1%, of the U.S. population of women of childbearing age may have mercury exposures greater than the EPA RfD for MeHg of 0.1 μg/kgg/day, and that there are few, if any, exposures greater than the ATSDR MRL of 0.3 μgg/kgg/day. The analysis also indicates that typical exposures may be greater than previously estimated from food consumption surveys, but that the variability in exposure within the population of U.S. women of childbearing age may be less than previously assumed.     

A Bayesian Analysis of Abundance,Trend, and Population Viability for Harbor Seals in Iliamna Lake,Alaska

Harbor seals in Iliamna Lake, Alaska, are a small, isolated population, and one of only two freshwater populations of harbor seals in the world, yet little is known about their abundance or risk for extinction. Bayesian hierarchical models were used to estimate abundance and trend of this population. Observational models were developed from aerial survey and harvest data, and they included effects for time of year and time of day on survey counts. Underlying models of abundance and trend were based on a Leslie matrix model that used prior information on vital rates from the literature. We developed three scenarios for variability in the priors and used them as part of a sensitivity analysis. The models were fitted using Markov chain Monte Carlo methods. The population production rate implied by the vital rate estimates was about 5% per year, very similar to the average annual harvest rate. After a period of growth in the 1980s, the population appears to be relatively stable at around 400 individuals. A population viability analysis assessing the risk of quasi‐extinction, defined as any reduction to 50 animals or below in the next 100 years, ranged from 1% to 3%, depending on the prior scenario. Although this is moderately low risk, it does not include genetic or catastrophic environmental events, which may have occurred to the population in the past, so our results should be applied cautiously.     

Bayesian Estimation of Dynamic Discrete Choice Models

We propose a new methodology for structural estimation of infinite horizon dynamic discrete choice models. We combine the dynamic programming (DP) solution algorithm with the Bayesian Markov chain Monte Carlo algorithm into a single algorithm that solves the DP problem and estimates the parameters simultaneously. As a result, the computational burden of estimating a dynamic model becomes comparable to that of a static model. Another feature of our algorithm is that even though the number of grid points on the state variable is small per solution‐estimation iteration, the number of effective grid points increases with the number of estimation iterations. This is how we help ease the “curse of dimensionality.” We simulate and estimate several versions of a simple model of entry and exit to illustrate our methodology. We also prove that under standard conditions, the parameters converge in probability to the true posterior distribution, regardless of the starting values.     

Modeling Risk‐Related Knowledge in Tunneling Projects

Knowledge on failure events and their associated factors, gained from past construction projects, is regarded as potentially extremely useful in risk management. However, a number of circumstances are constraining its wider use. Such knowledge is usually scarce, seldom documented, and even unavailable when it is required. Further, there exists a lack of proven methods to integrate and analyze it in a cost‐effective way. This article addresses possible options to overcome these difficulties. Focusing on limited but critical potential failure events, the article demonstrates how knowledge on a number of important potential failure events in tunnel works can be integrated. The problem of unavailable or incomplete information was addressed by gathering judgments from a group of experts. The elicited expert knowledge consisted of failure scenarios and associated probabilistic information. This information was integrated using Bayesian belief‐networks‐based models that were first customized in order to deal with the expected divergence in judgments caused by epistemic uncertainty of risks. The work described in the article shows that the developed models that integrate risk‐related knowledge provide guidance as to the use of specific remedial measures.     

Capturing and Integrating Knowledge for Managing Risks in Tunnel Works

Risk‐related knowledge gained from past construction projects is regarded as potentially extremely useful in risk management. This article describes a proposed approach to capture and integrate risk‐related knowledge to support decision making in construction projects. To ameliorate the problem related to the scarcity of risks information often encountered in construction projects, Bayesian Belief Networks are used and expert judgment is elicited to augment available information. Particularly, the article provides an overview of judgment‐based biases that can appear in the elicitation of judgments for constructing Bayesian Networks and the provisos that can be made in this respect to minimize these types of bias. The proposed approach is successfully applied to develop six models for top risks in tunnel works. More than 30 tunneling experts in the Netherlands and Germany were involved in the investigation to provide information on identifying relevant scenarios than can lead to failure events associated with tunneling risks. The article has provided an illustration of the applicability of the developed approach for the case of “face instability in soft soils using slurry shields.”     

Quantitative Assessment of Human MRSA Risks from Swine

The public health community, news media, and members of the general public have expressed significant concern that methicillin‐resistant Staphylococcus aureus (MRSA) transmitted from pigs to humans may harm human health. Studies of the prevalence and dynamics of swine‐associated (ST398) MRSA have sampled MRSA at discrete points in the presumed causative chain leading from swine to human patients, including sampling bacteria from live pigs, retail meats, farm workers, and hospital patients. Nonzero prevalence is generally interpreted as indicating a potential human health hazard from MRSA infections, but quantitative assessments of resulting risks are not usually provided. This article integrates available data from several sources to construct a conservative (plausible upper bound) probability estimate for the actual human health harm (MRSA infections and fatalities) arising from ST398‐MRSA from pigs. The model provides plausible upper bounds of approximately one excess human infection per year among all U.S. pig farm workers, and one human infection per 31 years among the remaining total population of the United States. These results assume the possibility of transmission events not yet observed, so additional data collection may reduce these estimates further.     

A gradient Markov chain Monte Carlo algorithm for computing multivariate maximum likelihood estimates and posterior distributions: mixture dose-response assessment

Multivariate probability distributions, such as may be used for mixture dose‐response assessment, are typically highly parameterized and difficult to fit to available data. However, such distributions may be useful in analyzing the large electronic data sets becoming available, such as dose‐response biomarker and genetic information. In this article, a new two‐stage computational approach is introduced for estimating multivariate distributions and addressing parameter uncertainty. The proposed first stage comprises a gradient Markov chain Monte Carlo (GMCMC) technique to find Bayesian posterior mode estimates (PMEs) of parameters, equivalent to maximum likelihood estimates (MLEs) in the absence of subjective information. In the second stage, these estimates are used to initialize a Markov chain Monte Carlo (MCMC) simulation, replacing the conventional burn‐in period to allow convergent simulation of the full joint Bayesian posterior distribution and the corresponding unconditional multivariate distribution (not conditional on uncertain parameter values). When the distribution of parameter uncertainty is such a Bayesian posterior, the unconditional distribution is termed predictive. The method is demonstrated by finding conditional and unconditional versions of the recently proposed emergent dose‐response function (DRF). Results are shown for the five‐parameter common‐mode and seven‐parameter dissimilar‐mode models, based on published data for eight benzene–toluene dose pairs. The common mode conditional DRF is obtained with a 21‐fold reduction in data requirement versus MCMC. Example common‐mode unconditional DRFs are then found using synthetic data, showing a 71% reduction in required data. The approach is further demonstrated for a PCB 126‐PCB 153 mixture. Applicability is analyzed and discussed. Matlab^® computer programs are provided.     

Sensitivity Analysis and Estimation of Extreme Tail Behavior in Two‐Dimensional Monte Carlo Simulation

Two‐dimensional Monte Carlo simulation is frequently used to implement probabilistic risk models, as it allows for uncertainty and variability to be quantified separately. In many cases, we are interested in the proportion of individuals from a variable population exceeding a critical threshold, together with uncertainty about this proportion. In this article we introduce a new method that can accurately estimate these quantities much more efficiently than conventional algorithms. We also show how those model parameters having the greatest impact on the probabilities of rare events can be quickly identified via this method. The algorithm combines elements from well‐established statistical techniques in extreme value theory and Bayesian analysis of computer models. We demonstrate the practical application of these methods with a simple example, in which the true distributions are known exactly, and also with a more realistic model of microbial contamination of milk with seven parameters. For the latter, sensitivity analysis (SA) is shown to identify the two inputs explaining the majority of variation in distribution tail behavior. In the subsequent prediction of probabilities of large contamination events, similar results are obtained using the new approach taking 43 seconds or the conventional simulation that requires more than 3 days.     

Predictive Bayesian Microbial Dose-Response Assessment Based on Suggested Self-Organization in Primary Illness Response: Cryptosporidium parvum

The probability of illness caused by very low doses of pathogens cannot generally be tested due to the numbers of subjects that would be needed, though such assessments of illness dose response are needed to evaluate drinking water standards. A predictive Bayesian dose-response assessment method was proposed previously to assess the unconditional probability of illness from available information and avoid the inconsistencies of confidence-based approaches. However, the method uses knowledge of the conditional dose-response form, and this form is not well established for the illness endpoint. A conditional parametric dose-response function for gastroenteric illness is proposed here based on simple numerical models of self-organized host-pathogen systems and probabilistic arguments. In the models, illnesses terminate when the host evolves by processes of natural selection to a self-organized critical value of wellness. A generalized beta-Poisson illness dose-response form emerges for the population as a whole. Use of this form is demonstrated in a predictive Bayesian dose-response assessment for cryptosporidiosis. Results suggest that a maximum allowable dose of 5.0 x 10(-7) oocysts/exposure (e.g., 2.5 x 10(-7) oocysts/L water) would correspond with the original goals of the U.S. Environmental Protection Agency Surface Water Treatment Rule, considering only primary illnesses resulting from Poisson-distributed pathogen counts. This estimate should be revised to account for non-Poisson distributions of Cryptosporidium parvum in drinking water and total response, considering secondary illness propagation in the population.     

Diminishing Willingness to Pay per Quality‐Adjusted Life Year: Valuing Acute Foodborne Illness

We design and conduct a stated‐preference survey to estimate willingness to pay (WTP) to reduce foodborne risk of acute illness and to test whether WTP is proportional to the corresponding gain in expected quality‐adjusted life years (QALYs). If QALYs measure utility for health, then economic theory requires WTP to be nearly proportional to changes in both health quality and duration of illness and WTP could be estimated by multiplying the expected change in QALYs by an appropriate monetary value. WTP is elicited using double‐bounded, dichotomous‐choice questions in which respondents (randomly selected from the U.S. general adult population, n = 2,858) decide whether to purchase a more expensive food to reduce the risk of foodborne illness. Health risks vary by baseline probability of illness, reduction in probability, duration and severity of illness, and conditional probability of mortality. The expected gain in QALYs is calculated using respondent‐assessed decrements in health‐related quality of life if ill combined with the duration of illness and reduction in probability specified in the survey. We find sharply diminishing marginal WTP for severity and duration of illness prevented. Our results suggest that individuals do not have a constant rate of WTP per QALY, which implies that WTP cannot be accurately estimated by multiplying the change in QALYs by an appropriate monetary value.     

Bayesian Inference for Hospital Quality in a Selection Model

This paper develops new econometric methods to infer hospital quality in a model with discrete dependent variables and nonrandom selection. Mortality rates in patient discharge records are widely used to infer hospital quality. However, hospital admission is not random and some hospitals may attract patients with greater unobserved severity of illness than others. In this situation the assumption of random admission leads to spurious inference about hospital quality. This study controls for hospital selection using a model in which distance between the patient's residence and alternative hospitals are key exogenous variables. Bayesian inference in this model is feasible using a Markov chain Monte Carlo posterior simulator, and attaches posterior probabilities to quality comparisons between individual hospitals and groups of hospitals. The study uses data on 74,848 Medicare patients admitted to 114 hospitals in Los Angeles County from 1989 through 1992 with a diagnosis of pneumonia. It finds the smallest and largest hospitals to be of the highest quality. There is strong evidence of dependence between the unobserved severity of illness and the assignment of patients to hospitals, whereby patients with a high unobserved severity of illness are disproportionately admitted to high quality hospitals. Consequently a conventional probit model leads to inferences about quality that are markedly different from those in this study's selection model.