Full length article
Comparison of three modelling approaches to predict the risk of campylobacteriosis in New Zealand

https://doi.org/10.1016/j.mran.2019.06.001Get rights and content

Highlights

  • New Zealand has a higher rate of reported campylobacteriosis cases.

  • Predicting campylobacteriosis cases becomes a matter of considerable concern.

  • Three modelling approaches were tested for predicting the cases in this study.

  • ARIMA intervention, the Holt-Winters and Bayesian were considered the best models.

  • The Bayesian model provides insight into the food chain and it is more informative.

Abstract

New Zealand has the highest rate of reported campylobacteriosis in the developed world. Due to the large economic and health consequences of campylobacteriosis, prediction models for disease are required to be designed to predict accurately the number of campylobacteriosis cases. The Bayesian approach has gained increased interest in recent years for calculating the outcomes of quantitative microbial risk assessment (QMRA). A classical time series and Monte Carlo (MC) modelling were also explored as appropriate techniques to predict campylobacteriosis. A simplified model representing the entire food chain from the farm to the fork with all the variables of interest was used with the Bayesian method. The Auto Regressive Integrated Moving-Average intervention models (ARIMA additive and multiplicative), Holt-Winters (HW multiplicative) and Bayesian methods were considered the best models for predicting the actual 7333 notified campylobacteriosis cases with 7990, 8442, 8666 and 9250 cases, respectively. It is also noteworthy that the notification rate has more or less stabilised since 2008 until 2017. MC modelling provided the least realistic prediction (846,451 cases). The HW method is simple to use and reliable method for time series predictions. However, the Bayesian method provides a prior assessment of any possible intervention in the food chain and provides satisfactory prediction accuracy in spite of the complexity involved in constructing and assigning probabilities from expert knowledge or prior information, linking the nodes and complex software. This study highlighted the importance of the Bayesian model to assess all the factors which may contribute to the campylobacteriosis risk and confirmed that it can provide better conclusions for QMRA than the MC technique because of its interactive link between the data and the parameter (backward inference).

Introduction

Risk assessment (RA) in food is a scientifically based process of identifying and characterising the hazards and risk factors associated with a given food system. Thus, microbial risk assessment (MRA) is the scientific evaluation of known or potential adverse health effects resulting from human exposure to biological hazards. This approach employs mathematical models that are used to predict: (i) the introduction of pathogens into food, (ii) the replication of pathogens in food over time, (iii) the destruction of pathogens by heat treatment or other techniques, (iv) the consumption of pathogens in food and (v) the subsequent probability of illness. New Zealand has a high rate of reported campylobacteriosis cases, especially when compared with other developed countries (Baker et al., 2006). Therefore, predicting campylobacteriosis cases has become a matter of considerable concern in New Zealand. Moreover, the intervention strategies selected by policymakers can be evaluated by the cost and benefit in terms of the number of campylobacteriosis cases which are expected to be prevented. Therefore, there is a need to develop a model to accurately predict accurately the number of campylobacteriosis cases in New Zealand. Time series forecasting methods have been successfully applied in the fields of engineering, science, sociology and economics (Brockwell and Davis, 2002). The classical time series modelling approach is considered as an appropriate technique to predict campylobacteriosis in New Zealand (Weisent et al., 2010). The objective of this study is to investigate which modelling technique can provide a suitable model to accurately predict the number of campylobacteriosis cases in New Zealand, to list the pros and cons of each modelling approach and to explore possible risk factors for the disease. Previously, Parsons et al. (2005) compared three modelling approaches for the QMRA, and others (Smid et al., 2010) compared two approaches for the QMRA or others introduced the fuzzy model (Davidson et al., 2006) but none of these studies has included the time series method for predicting risk.

Section snippets

2.1. QMRA approach description

The concentration of the pathogen is the variable which each quantitative microbial risk assessment (QMRA) will quantify in order to follow the dynamics of the pathogen through the food chain. The concentration of the pathogen is usually dependent on the variable describing the concentration at a previous stage in the food chain and other process parameters, such as temperature, time, pH, water activity etc. Generally, the QMRA divides the food chain into a specific number of modules which can

3.1. Bayesian techniques

The statistical output of the prior distribution and posterior distribution were summarized in Table 3. The Monte Carlo error (MC error) for each parameter is also shown in the table. The MC error assesses the accuracy of the posterior estimates. It represents the difference between the mean of the sampled values, which are used as the estimate of the posterior mean for each parameter, and the true posterior mean. The statistics table also reports the sample standard deviation (sd) with the

Conclusion

The complete food chain has been modelled by a Bayesian hierarchal model, which provides better insight than the time series model into the food chain and is more informative as it incorporates all the factors that impact the final risk estimation. Thus, it can easily determine the impact of any intervention in the food chain. Therefore, the effect of a new planned intervention such as a consumer education plan was clear for the policy makers, risk managers and health professionals before the

Acknowledgement

The author would like to thank Professor Geoff Jones from the Institute of Fundamental Science, Massey University for his statistical support. The author greatly acknowledges the provision of data by The Poultry Industry Association of New Zealand (PIANZ) and the author appreciates the support of Michael Brooks, executive director, and Kerry Mulqueen from The Poultry Industry Association of New Zealand. Dr. Roy Biggs, Q.A Manager at Tegel Foods Ltd, and Don Thomas, Manager of Poultry Research

References (88)

  • H.D. Kusumaningrum et al.

    A quantitative analysis of cross-contamination of Salmonella and Campylobacter spp. via domestic kitchen surfaces

    J. Food Prot.

    (2004)
  • R. Lindqvist et al.

    Quantitative risk assessment of thermophilic Campylobacter spp. and cross-contamination during handling of raw broiler chickens evaluating strategies at the producer level to reduce human campylobacteriosis in Sweden

    Int. J. Food Microbiol.

    (2008)
  • P. Luber

    Cross-contamination versus undercooking of poultry meat or eggs — which risks need to be managed first?

    Int. J. Food Microbiol.

    (2009)
  • R. Maijala et al.

    A quantitative risk assessment of the public health impact of the Finnish Salmonella control program for broilers

    Int. J. Food Microbiol.

    (2005)
  • M. Mataragas et al.

    Quantitative microbiological risk assessment as a tool to obtain useful information for risk managers - Specific application to Listeria monocytogenes and ready-to-eat meat products

    Int. J. Food Microbiol.

    (2010)
  • M. Nauta et al.

    A comparison of risk assessments on Campylobacter in broiler meat

    Int. J. Food Microbiol.

    (2009)
  • D.J. Parsons et al.

    A comparison of three modelling approaches for quantitative risk assessment using the case study of Salmonella spp. in poultry meat.

    Int. J. Food Microbiol.

    (2005)
  • M.R. Powell et al.

    Bayesian synthesis of a pathogen growth model: listeria monocytogenes under competition

    Int. J. Food Microbiol.

    (2006)
  • G. Purnell et al.

    The use of `hot wash' treatments to reduce the number of pathogenic and spoilage bacteria on raw retail poultry

    J. Food Eng.

    (2004)
  • J.H. Smid et al.

    Strengths and weaknesses of Monte Carlo simulation models and Bayesian belief networks in microbial risk assessment

    Int. J. Food Microbiol.

    (2010)
  • A. Spor et al.

    Hierarchical Bayesian Modelling for Saccharomyces cerevisiae population dynamics

    Int. J. Food Microbiol.

    (2010)
  • N. Toft et al.

    Assessing the convergence of Markov Chain Monte Carlo methods: an example from evaluation of diagnostic tests in absence of a gold standard

    Prev. Vet. Med.

    (2007)
  • P. Tuominen et al.

    Salmonella risk in imported fresh beef, beef preparations, and beef products

    J. Food Prot.

    (2006)
  • L. Uusitalo

    Advantages and challenges of Bayesian networks in environmental modelling

    Ecol. Model.

    (2007)
  • P. Whyte et al.

    An assessment of steam pasteurization and hot water immersion treatments for the microbiological decontamination of broiler carcasses

    Food Microbiol.

    (2003)
  • A. Al-Sakkaf et al.

    Comparison of time-series models for predicting campylobacteriosis risk in New Zealand

    Zoonoses Public Health

    (2013)
  • I. Albert et al.

    Quantitative risk assessment from farm to fork and beyond: a global Bayesian approach concerning food-borne diseases

    Risk Anal.

    (2008)
  • S. Altizer et al.

    Seasonality and the dynamics of infectious diseases

    Ecol. Lett.

    (2006)
  • M. Baker et al.

    Is the major increase in notified campylobacteriosis in New Zealand real?

    Epidemiol. Infect

    (2006)
  • J. Benschop et al.

    Temporal and longitudinal analysis of Danish swine salmonellosis control programme data: implications for surveillance

    Epidemiol. Infect.

    (2008)
  • R.E. Black et al.

    Experimental Campylobacter jejuni infection in humans

    J. Infect. Dis.

    (1988)
  • C.S. Bonham et al.

    Intervention analysis with co-integrated time series: the case of the Hawaii hotel room tax

    Appl. Econ.

    (1996)
  • G.E.P. Box et al.

    Time series analysis: forecasting and control

    (1976)
  • G. Box et al.

    Intervention analysis with applications to economic and environmental problems

    J. Am. Stat. Assoc.

    (1975)
  • N. Boxall
  • P.J. Brockwell et al.

    Introduction to Time Series and Forecasting

    (2002)
  • S.P. Brooks et al.

    General methods for monitoring convergence of iterative simulations

    J. Comput. Gr. Stat.

    (1998)
  • M. Brown et al.

    Microbiological Risk Assessment in Food Processing

    (2002)
  • S.M. Cahill

    Risk assessment and campylobacteriosis

  • C. Chatfield

    The Analysis of Time Series: An Introduction

    (2004)
  • S.C. De Greeff et al.

    Seasonal patterns in time series of pertussis

    Epidemiol. Infect.

    (2009)
  • S.A. DeLurgio

    Forecasting Principles and Applications

    (1998)
  • T.J. Doyle et al.

    Completeness of notifiable infectious disease reporting in the United States: an analytical literature review

    Am. J. Epidemiol.

    (2002)
  • N. French

    Enhancing surveillance of potentially foodborne enteric diseases in New Zealand: Human campylobacteriosis in the Manawatu

    (2008)
  • 1

    Present address: Lincoln University, Faculty of Agriculture and Life Sciences PO Box 85084, Lincoln 7647, Christchurch, New Zealand.

    View full text