Attribution of pathogen-specific costs of foodborne illness to food commodity groups - combining a costing model with expert judgement

Attribution of pathogen-specific costs of foodborne illness to food commodity groups - combining a costing model with expert judgement | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Attribution of pathogen-specific costs of foodborne illness to food commodity groups - combining a costing model with expert judgement Anca Hanea, Angus McLure, Ben Daughtry, Snezana Smiljanic, Kathryn Glass This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8110329/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Foodborne disease and its sequelae exert a significant cost to Australia through healthcare costs, lost productivity, and occasional fatal illness. While estimating the cost of illness for all foodborne pathogens or for specific pathogens has value in quantifying this disease burden, it is also informative to estimate costs by food commodity and to identify priority areas for improving food safety. We combined a cost of illness model for foodborne illness in Australia with an expert elicitation of the food commodities associated with illness for key pathogens. The total cost of the six modelled pathogens was 721 million (June 2023 AUD), with campylobacteriosis having the greatest overall cost (AUD 420 million). Across food categories, AUD 328 million was attributed to poultry, AUD 107 million to vegetables, while dairy, beef, and pork each had costs over AUD 55 million. Strong associations were found between Campylobacter and poultry (69% of campylobacteriosis cases attributed to poultry) and Yersinia and pork (54% of yersiniosis cases attributed to pork). This study highlights poultry as a key cause of foodborne illness in Australia, responsible for almost half of the total costs due to Campylobacter , non-typhoidal Salmonella , Yersinia enterocolitica , Listeria monocytogenes , and STEC. Health Economics & Outcomes Research Epidemiology disease burden pathogen-specific costs of foodborne illness attribution of foodborne illnesses to food commodities structured expert judgement IDEA Classical Model Figures Figure 1 Figure 2 1. Introduction Circa 2019, foodborne illness was estimated to cost Australia 2.44 billion Australian dollars (AUD) annually, with AUD 365 million attributed to Campylobacter and AUD 140 million to Salmonella (Glass and others 2023). While pathogen-specific costs highlight high-impact pathogens, attributing costs to specific foods helps target interventions. Combined with cost–benefit analyses, such estimates guide industry, government, and consumers toward policies that most effectively reduce the foodborne disease burden (Butler and others 2015; Sapp and others 2022; Scharff 2012 ). Several approaches have been used to attribute foodborne illness to food groups or animal reservoirs (Batz and others 2012; Kumagai and others 2020). Mathematical source attribution models combine typed isolates from humans with typed isolates from foods and food-producing animals to estimate the proportion of cases attributable to each reservoir (Hald and others 2004; Mughini Gras and others 2012; Mullner and others 2009; Pires and others 2009). Such approaches require large datasets and generally identify animal reservoirs rather than human infection pathways. An alternative approach to source attribution is expert elicitation (Butler and others 2015). Experts have been used to determine the proportion of foodborne illness cases attributable to different transmission routes (Beshearse and others 2021; Havelaar and others 2008; Vally and others 2014). Evidence of transmission routes contributes to burden of disease studies (Daniel and others 2020; Havelaar and others 2015; Lagerweij and others 2020; Scallan and others 2011). Expert elicitation gathers estimates of unknown quantities from experts. Structured expert judgement (SEJ) elicitation protocols are a collection of steps effective for working with experts' uncertainty estimates. While SEJ lacks an official definition, Cooke provides a working one (Cooke 1991 ) highlighting that “structured” expert judgement should be treated as scientific data, and formulates four necessary principles for SEJ as a scientific method: Accountability : All data, analyses, and software are peer-reviewed, ensuring reproducibility for reviewers while keeping expert identities confidential. Empirical control : Experts' performance as uncertainty assessors is measured based on calibration variables (variables for which true values exist). Neutrality : The evaluation and aggregation of expert assessments encourage experts to state their true opinion. Fairness : Experts are not pre-judged prior to empirical control. An expert elicitation process using Cooke’s Classical Model (CM) method was undertaken in 2020 to estimate the attribution of foodborne illness to specific foods for five pathogens: non-typhoidal Salmonella , Shiga-toxin producing Escherichia coli (STEC), Listeria monocytogenes , hepatitis A virus and norovirus (Food Standards Australia New Zealand 2022 ). The current study extends this elicitation to eight pathogens: non-typhoidal Salmonella , Campylobacter , STEC, L . monocytogenes , Y . enterocolitica , Vibrio spp. and B . cereus and fourteen food commodities, and combines these findings with an existing cost of illness model for six of the eight pathogens (excluding Vibrio spp. and B . cereus ) to attribute costs of foodborne illness to food commodity groups in Australia. 2. Methods 2.1. SEJ protocol SEJ protocols include the IDEA protocol (Hanea and others 2017) and the CM, used here in conjunction. SEJ protocols use some structured steps, like independent estimation, anonymisation, controlled group discussion, to mitigate cognitive biases (anchoring, overconfidence) and group biases (dominance, conformity) that can distort judgments [e.g., Cooke (1991); Hemming and others (2018); O'Hagan and others (2006)]. Briefly, IDEA involves four stages: · Investigate : Experts work individually to provide point and interval (low–high) estimates of unknown quantities, following a group meeting to clarify terms, resolve ambiguities, and confirm the process. Experts’ estimates are de-identified and summarised by facilitators for the discussion stage. · Discuss : Summaries are shared and discussed in a workshop. Experts discuss the spread of results, disagreements, and reasoning. · Estimate : Experts may then privately revise their estimates, though updates are optional and need not be disclosed. · Aggregate : The revised estimates are converted into probability distributions per expert, then mathematically combined (via linear pooling with equal or weighted contributions). Independent initial estimates prevent anchoring on others' estimates (likely in unstructured group discussions), the discussion fosters counterfactual thinking and addresses the availability bias, and the second anonymous round of estimates limits the dominance effect and groupthink. Because the aggregation stage is done mathematically, CM’s mathematical apparatus can be added to an IDEA elicitation, by eliciting calibration questions, allowing for the calculation of performance-based weights to be used in the linear pooling of experts' distributions. Answers to the calibration questions are used to gauge the “quality” of expert judgements, by measuring performance. The performance of the experts on the calibration questions is taken as indicative of their performance on the target questions. Therefore, the calibration questions must capture the type of knowledge and reasoning needed to answer the target questions. The more calibration variables the better, but ten has proven to be sufficient (Colson and Cooke 2017). Performance-based weights reflect how calibrated and informative expert judgements are. A good probability assessor is one whose assessments capture the true values consistently in the long run (well calibrated), with distributions that are as narrow as possible (informative). We used the calibration and informativeness measures defined in (Cooke 1991). Calibration (ranging from 0 to 1) was measured as the p-value at which the hypothesis that the expert is well calibrated would be falsely rejected. Higher scores are better, since low values mean that it is very unlikely that the discrepancy between an expert's judgements and the observed outcomes arose by chance. Informativeness is measured using the Kullback-Leibler divergence (Kullback and Leibler 1951) with respect to the uniform distribution (which is what one would assume in absence of experimental data). Informativeness scores are non-negative, with higher scores being better (Hanea and Nane 2021). Calibration and informativeness scores were combined (multiplied and normalised) to form weights and performance-based aggregations. 2.1.1. Details of the present elicitation The elicitation scope, target and calibration questions were developed with the expert input of external experts (please see Acknowledgements). Target questions covered the attribution of foodborne illness to specific foods for eight pathogens (listed in Introduction). The food groups were developed from published sources (Hoffmann 2017, Painter 2009) and existing knowledge, and they are: beef, lamb, pork, poultry, eggs, dairy (milk and cream, fresh uncured cheese, brined cheese, soft-ripened cheese, firm-ripened cheese), finfish, crustaceans, molluscs, fruit, grains and seeds, nuts, vegetables (fungi, leafy vegetables and herbs, root vegetables, sprouts, vine-stalk), and ‘other’. 2.1.2. Process Thirteen experts (listed in Table (S1) from Supplementary Material, SM) were identified and approached, and eleven participated. Their expertise covered health science, environmental and food microbiology, biosecurity, seafood safety, zoonotic disease epidemiology, food allergen risk management, and environmental health management. Experts signed Consent forms, and were briefed on the expectations via a Plain Language Statement. During an inception meeting the scope and steps of the elicitation were reiterated. Experts reviewed and provided feedback on the food commodity/pathogen matrix, a practice elicitation question and the calibration questions. The first round estimates for calibration questions were elicited during this meeting, in a controlled environment. Expert feedback on the commodity/pathogen matrix was incorporated, and target questions were finalised and sent post meeting. Experts had three weeks to answer the target questions and were instructed to not research the (already answered) calibration questions. Their Judgements were collected via email, de-identified, and collated. The facilitated workshop (the Discuss stage) comprised two recorded online sessions with ten experts each session, and eleven experts attending at least once. All questions were discussed and refined as needed. The de-identified data and the recordings were sent to two experts who missed both sessions. Following the workshop, experts reconsidered their estimates. 2.1.3. Questions To acknowledge uncertainty, we elicited point estimates together with upper and lower bounds, eliciting the lower bound first, followed by the upper bound, and finally the best estimate. This question order aims to reduce anchoring and overconfidence (narrow intervals), though empirical evidence for its effectiveness in this exact context is limited. We asked 12 calibration questions (listed in SM) and 72 target questions corresponding to pairs of pathogens and food commodities. An example target question set is shown in Figure (S1). 2.1.4. Aggregated estimates We described the probability distribution per question, per expert, using the revised estimates, with best estimates set to be the medians of distributions and the upper and lower bounds set to represent 5% and 95% percentiles. Non-parametric distributions were then fitted on the three percentiles (5%, median, 95%). For technical details we refer to the supplementary material SM3. An aggregated distribution per question was obtained by combining distributions across experts linearly with equal or differential weights. We used the calibration questions to calculate calibration and informativeness scores. These scores were used to calculate differential weights, sometimes called global weights (proportional to the product of the two scores) which were further used to calculate a performance based weighted distribution that represents the group’s judgements. A variant of the globally weighted aggregation is the optimised globally weighted aggregation. The distinction between the two methods is detailed in SM3. Briefly, the optimisation sequentially makes small weights zero until the aggregation of the remaining distributions (with non-zero weights) reaches the best combined score. 2.2. Costing model The total burden and cost estimation for the six pathogens is identical to previous methods detailed elsewhere (Glass and others 2023), though here we report inflation-adjusted costs from December 2019 to June 2023 (+15.1%). Briefly, we estimated burden in terms of the annual number of cases, hospitalisations, and deaths due the pathogens and four sequelae arising from them: Guillain-Barré syndrome (GBS), Haemolytic uraemic syndrome (HUS), irritable bowel syndrome (IBS), and Reactive arthritis (ReA). Disease burden estimates followed prior approaches (Ford and others 2014; Kirk and others 2014), including the selection of sequelae. For each pathogen and sequelae, we estimated costs from premature mortality, lost productivity, healthcare, and non-financial costs of pain and suffering. For the five notifiable pathogens ( Campylobacter, non-typhoidal Salmonella , Yersinia enterocolitica, Listeria monocytogenes , and STEC) the number of cases was estimated as the number of notifications in 2019 multiplied by underreporting multipliers (Hall and others 2008). For Toxoplasma gondii, which is not notifiable, we estimated annual number of cases from seroprevalence curves by age (Molan and others 2020), assuming 15% of exposures that lead to seroconversion result in symptomatic cases, consistent with prior work (Glass and others 2023; Kirk and others 2014). We estimated the number of sequelae by multiplying initial cases by the risk of sequelae determined previously (Ford and others 2014). All cases of listeriosis, GBS, and HUS were assumed to be hospitalised. For all other conditions, we estimated hospitalisations using publicly available administrative healthcare data on hospitalisation by principal diagnosis (Australian Government Institute of Health and Welfare 2020), with adjustments for underdiagnosis and the fraction of admissions where the condition was a secondary diagnosis (Kirk and others 2014). We estimated annual deaths by averaging reported data for 2001-2010 from ABS (Australian Bureau of Statistics 2010), with adjustment for underdiagnosis (Kirk and others 2014) and additional listeriosis deaths in neonates from OzFoodNet annual reports (OzFoodNet Working Group 2015; 2018; 2021a; 2021b). All burden estimates were multiplied by the proportion of cases that are foodborne as estimated in a previous expert elicitation (Vally and others 2014) and the number of incident cases were multiplied by the proportion that were domestically acquired (Kirk and others 2014). Premature mortality was costed using the Value of Statistical Life (Australian Government Department of Prime Minister and Cabinet: Office of Best Practice Regulation) applied to each death regardless of age. We sourced healthcare costs from the Medicare Benefits Schedule, the Pharmaceutical Benefits Scheme, and the Australian Refined Diagnosis Related Groups. We costed lost productivity for cases and their carers using published workforce participation data (Australian Bureau of Statistics), estimates of days of work missed from the National Gastroenteritis survey (Kirk 2008) for non-hospitalised cases, and the average length of stay for hospitalised cases (Australian Government Institute of Health and Welfare). Non-financial costs were costed using a previous discrete choice experiment which quantified willingness to pay to avoid pain and suffering (Manipis and others 2023). To estimate burden/costs attributable to specific food categories for each pathogen, we multiplied our previously published estimates for each pathogen (Glass and others 2023) by the proportion of cases of that pathogen attributable to each food category, i.e. we assumed that outcomes (hospitalisations, sequelae, deaths) and costs did not vary by the food category that led to the infection. All estimates were stratified by age group (<5, 5-64, 65+). Uncertainty in attribution proportions, multipliers, and other inputs was incorporated into final estimates using Monte Carlo simulations, and is reported as 90% credible intervals. 3. Results 3.1. Expert elicitation Ten experts completed the revised estimates, which were aggregated using three aggregation schemes: equal weighting, global weighting without optimisations and with optimisation (see the Excel spreadsheets uploaded as SM for details ). The optimisation procedure selected one expert to be the optimum “aggregation”. Although the calibration score of this expert is larger than the calibration of the other two aggregations, the differences in the aggregated distributions are minimal (see Hanea and Nane (2021) for examples of apparent large differences in calibration scores that generated insignificant changes in distributions). The large informativeness score may be indicative of overconfidence, so we choose not to use the optimised version of the globally-weighted aggregation. Although the calibration scores of the equally and globally-weighted aggregations were equal, we used the globally-weighted aggregation for two reasons: 1) the large differences between experts’ calibration scores, and 2) the larger informativeness score of the latter. Figure (1) shows the calibration and informativeness scores of all experts and aggregations. The vertical line marks the 0.05 calibration level (often considered the minimal "good" calibration score). For the exact scores see Table (S2) from SM1. Table (1) provides the distributions obtained using the chosen aggregation, including examples of best estimates and bounds for the food/commodity with the largest contribution per pathogen. Particularly strong associations were found between Campylobacter and poultry (median 69%) and Yersinia and pork (54%). The complete aggregated data are provided in Tables (S3) and (S4). 3.2. Attributed burden and costs Across the six pathogens considered, the overall annual burden was 357,000 cases (Figure (2), Table (S5)) and total cost was AUD 721 million (inflation-adjusted to June 2023, Table 2). Campylobacteriosis was the cause of the most cases (264000; Figure (2), Table (S5)) and had the highest overall cost (AUD 420 million; Figure (2), Table (S6)). Campylobacter was the leading cause of hospitalisation, with 5,640 hospitalisations due to initial disease and a further 598 due to ReA, 2,300 due to IBS and 98 due to GBS (Figure (2)). Listeriosis, although causing far fewer cases and hospitalisations (101), was the leading cause of deaths (fifteen; Figure (2)). Extensive additional details on burden and costs can be found in our previous publication on this model (Glass and others 2023). Poultry had the highest foodborne disease burden across cases, hospitalisations, deaths, and cost (Figure (2)). Of the AUD 328 million cost attributed to poultry, AUD 279 million was from Campylobacter, 35.5 million from non-typhoidal Salmonella , and 2.9 million from L. monocytogenes. Of the 191 thousand cases attributed to poultry, 174 thousand were attributed to Campylobacter and 13.2 thousand to non-typhoidal Salmonella . Vegetables was the second highest category associated with cases, hospitalisations, deaths, and costs (AUD 107 million), with non-typhoidal Salmonella the leading pathogen (AUD 41.7 million). Dairy, beef, and pork had the next highest costs, each category costing over AUD 55 million. Full details of estimated cases, deaths, hospitalisations, and costs can be found in Tables (S12-S18). Table (2) provides a breakdown of costs by age group. As the elicitation was not stratified by age group, age-specific differences reflect disease burden rather than varying food-pathogen associations. Total costs across all food commodities were highest in the 5–64-year-olds (404 million) and lowest in children under 5 (78 million), reflecting the relative population in each group. However, for crustaceans, finfish, molluscs, and fruit, attributed costs were highest in people over 65, as listeriosis accounted for most costs associated with these commodities (Figure 2) and the cost of listeriosis was highest in this age group (Table (S7)). Summaries of cases for each food category and detailed breakdowns of costs by food category, pathogen, and age group are presented in the supplementary materials, Tables (S5-S11). 4. Discussion The six foodborne pathogens included in the model were estimated to cost AUD 721 million annually, with almost half of this cost (AUD 328 million) attributed to poultry. Other food categories with high costs included vegetables (AUD 107 million), dairy (AUD 61 million), beef (AUD 56 million) and pork (AUD 56 million) with AUD 56 million attributed to additional ‘other’ categories. High costs, case numbers, and hospitalisations attributed to poultry were largely from the strong association between poultry and Campylobacter. Food categories with high numbers of attributed deaths included poultry, vegetables and dairy, from associations with Campylobacter, Listeria monocytogenes , and non-typhoidal Salmonella . Experts identified relatively strong associations between vegetables and both non-typhoidal Salmonella and Listeria monocytogenes , with vegetables having the highest attribution proportion for both pathogens (median of 26% for Salmonella and 24% for Listeria monocytogenes ). Salmonellosis and listeriosis are often associated with outbreaks of fresh produce, commonly vegetables (Reddy and others 2016; Smith and others 2018), although a review of foodborne outbreaks in Australia over 2001–2016 found stronger associations between salmonellosis outbreaks and eggs (Ford and others 2018). Internationally, non-typhoidal Salmonella is often associated with poultry or eggs (Batz and others 2012; Davidson and others 2011; Kumagai and others 2020; Sapp and others 2022; Scharff 2020 ), although associations have been found with vegetables (Rose and others 2025). While poultry and eggs each had a lower attribution proportion than vegetables in this elicitation, they combined to a higher total proportion. Listeria monocytogenes has been associated with both luncheon meats (Batz and others 2012; Davidson and others 2011) and dairy (Mangen and others 2015). While we have not included luncheon meat as a category, dairy had the second highest association with listeriosis after vegetables. Many of our findings are consistent with pathogen-food associations identified elsewhere. The strong association between Campylobacter and poultry has been found in the United States (Batz and others 2012; Scharff 2020 ), Canada (Davidson and others 2011), Japan (Kumagai and others 2020), in an expert elicitation in three African countries (Sapp and others 2022), and in a global attribution study (Hoffmann and others 2017). Strong associations have been found between E. coli and beef (Batz and others 2012; Davidson and others 2011; Kumagai and others 2020; Mangen and others 2015), as has Yersinia and pork (Batz and others 2012; Davidson and others 2011). Our finding of a relatively low cost associated with Toxoplasma gondii and associations with beef and lamb differ from international studies. A study in the United States identified pork and T. gondii as the second most important pathogen-food combination in most food rankings (Batz and others 2012), while pork was identified as contributing around half the cost from T. gondii in a study in the Netherlands (Mangen and others 2015). Our model of T. gondii results in relatively low costs per case (Glass and others 2023), with poor data quality limiting our ability to refine these estimates. A limitation of our work is that only six of ten pathogens (available in the costing model) were included in the elicitation due to time constraints, and that the elicitation did not consider differences in pathogen-food associations by age. A priority for further work is conducting an elicitation for norovirus, which was estimated to cost Australia AUD 128 million circa 2019 (Glass and others 2023). Owing to high levels of person-to-person transmission for norovirus (Vally and others 2014), this elicitation may need to repeat an assessment of transmission pathways to ensure pathogen-food associations are appropriately contextualised. Although Cryptosporidium was not prioritised for inclusion in our costing model, it is commonly included in international studies and could be considered in future modelling. A key strength of this work is the combination of a detailed costing model with SEJ, which enables costing of food contributions to foodborne disease. While SEJ can improve group judgements, it cannot guarantee perfect estimates. Using calibration questions helps, but only under the assumption that the calibration questions are representative for the target questions. Another untested aspect of using SEJ protocols with facilitated discussion is the effect of a remote discussion (as opposed to a face-to-face setting) on the performance scores. Nevertheless, SEJ provides a valuable approach to identifying food categories associated with foodborne disease pathogens and, when combined with costing data, enables cost estimations of disease burden. Our analysis for Australia indicated poultry as a key cause of foodborne illness, responsible for almost half of the total costs due to Campylobacter , Non-typhoidal Salmonella , Yersinia enterocolitica , Listeria monocytogenes , and STEC. The identified costs and food-pathogen links will guide priorities for improving food safety in Australian production systems. Declarations Acknowledgements We would like to acknowledge Dr Sandra Hoffmann (Economic Research Service, United States Department of Agriculture), Dr Sara Monteiro Pires (National Food Institute, Technical University of Denmark) and the Food Safety & Microbiology team at Food Standards Australia New Zealand (FSANZ) for helping to identify the substantive knowledge needed for formulating the calibration questions. Thirteen experts completed all rounds of the expert elicitation process: Kate Astridge, Mark Chan, Karen Ferres, Stacy Kane, Thea King, Allison McNamara, Stewart Quinn, Tom Ross, Robin Sherlock, Henry Tan, Alison Turnbull, Mark Turner, and Helen Withers. This work was funded by FSANZ under contract 2022-23/37. Authorship contribution statement Anca Hanea, Angus McLure, Ben Daughtry, Snezana Smiljanic, Kathryn Glass Conceptualization, Methodology, Writing- Reviewing and Editing , Anca Hanea, Angus McLure, Kathryn Glass: Data curation, Formal analysis, Writing- Original draft preparation Conflict of interest The Authors declare no conflicts of interest. Funding Statement This work was funded by Food Standards Australia New Zealand (FSANZ) under contract 2022-23/37 References Australian Bureau of Statistics. Customised Report: Causes of Death 2001-2010. 2010. Australian Bureau of Statistics. Labour Force, Australia, Detailed: Detailed monthly and quarterly Labour Force Survey data, including hours, regions, families, job search, job duration, casual, industry and occupation, Sept 2021. 2021. Australian Government Department of Prime Minister and Cabinet: Office of Best Practice Regulation. Best Practice Regulation Guidance Note: Value of Statistical Life; 2019. 2019. Australian Government Institute of Health and Welfare. Principal Diagnosis data cubes 2020. 2020. Batz M.B., Hoffmann S. and Morris Jr J.G. Ranking the disease burden of 14 pathogens in food sources in the United States using attribution data from outbreak investigations and expert elicitation. Journal of food protection 2012; 75:1278-1291. Beshearse E., Bruce B.B., Nane G.F., et al. . Attribution of Illnesses Transmitted by Food and Water to Comprehensive Transmission Pathways Using Structured Expert Judgment, United States. Emerging Infectious Diseases 2021; 27:182-195. Butler A.J., Thomas M.K. and Pintar K.D.M. Systematic Review of Expert Elicitation Methods as a Tool for Source Attribution of Enteric Illness. Foodborne Pathogens and Disease 2015; 12:367-382. Colson A.R. and Cooke R.M. Cross validation for the classical model of structured expert judgment. Reliability Engineering & System Safety 2017; 163:109-120. Cooke R.M. Experts in uncertainty: Opinion and subjective probability in science . New York: Oxford University Press, 1991. Cooke R.M. and Goossens L.L.H.J. TU Delft expert judgment data base. Reliability Engineering & System Safety 2008; 93:657-674. Daniel N., Casadevall N., Sun P., Sugden D. and Aldin V. The Burden of Foodborne Disease in the UK, 2018. Food Standards Agency, 2020. Davidson V.J., Ravel A., Nguyen T.N., Fazil A. and Ruzante J.M. Food-Specific Attribution of Selected Gastrointestinal Illnesses: Estimates from a Canadian Expert Elicitation Survey. Foodborne Pathogens and Disease 2011; 8:983-995. Food Standards Australia New Zealand. Primary Production and Processing (PPP) Requirements for Horticulture (Berries, Leafy Vegetables and Melons): Supporting document 2 to the Decision Regulation Impact Statement: Cost-benefit analysis. Volume P1052, 2022. Ford L., Kirk M., Glass K. and Hall G. Sequelae of Foodborne Illness Caused by 5 Pathogens, Australia, Circa 2010. Emerging Infectious Diseases 2014; 20:1865-1871. Ford L., Moffatt C.R.M., Fearnley E., et al. . The Epidemiology of Salmonella enterica Outbreaks in Australia, 2001–2016. Frontiers in Sustainable Food Systems 2018; 2. Glass K., McLure A., Bourke S., et al. . The Cost of Foodborne Illness and Its Sequelae in Australia Circa 2019. Foodborne Pathogens and Disease 2023; 20:419-426. Hald T., Vose D., Wegener H.C. and Koupeev T. A Bayesian Approach to Quantify the Contribution of Animal‐Food Sources to Human Salmonellosis. Risk Analysis 2004; 24:255-269. Hall G., Yohannes K., Raupach J., Becker N. and Kirk M. Estimating Community Incidence of Salmonella , Campylobacter , and Shiga Toxin–producing Escherichia coli Infections, Australia. Emerging Infectious Diseases 2008; 14:1601-1609. Hanea A., McBride M., Burgman M., et al. . I nvestigate D iscuss E stimate A ggregate for structured expert judgement. International Journal of Forecasting 2017; 33:267-279. Hanea A.M. and Nane G.F. An In-Depth Perspective on the Classical Model. In Expert Judgement in Risk and Decision Analysis . Hanea AM, Nane GF, Bedford T, and French S, (eds.). Cham: Springer, 2021, pp 225-256. Havelaar A.H., Galindo Á.V., Kurowicka D. and Cooke R.M. Attribution of Foodborne Pathogens Using Structured Expert Elicitation. Foodborne Pathogens and Disease 2008; 5:649-659. Havelaar A.H., Kirk M.D., Torgerson P.R., et al. . World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLOS Medicine 2015; 12:e1001923. Hoffmann S., Devleesschauwer B., Aspinall W., et al. . Attribution of global foodborne disease to specific foods: Findings from a World Health Organization structured expert elicitation. PLOS ONE 2017; 12:e0183641. Kirk M. National Gastroenteritis Survey II (NGSII). 2008. Kirk M., Ford L., Glass K. and Hall G. Foodborne Illness, Australia, Circa 2000 and Circa 2010. Emerging Infectious Diseases 2014; 20:1857-1864. Kullback S. and Leibler R.A. On information and sufficiency. The annals of mathematical statistics 1951; 22:79-86. Kumagai Y., Pires S.M., Kubota K. and Asakura H. Attributing human foodborne diseases to food sources and water in Japan using analysis of outbreak surveillance data. Journal of food protection 2020; 83:2087-2094. Lagerweij G.R., Pijnacker R., Friesema I.H.M., Mughini Gras L. and Franz E. Disease burden of food-related pathogens in the Netherlands, 2019. Bilthoven: National Institute for Public Health and the Environment, RIVM, 2020. Mangen M.-J.J., Bouwknegt M., Friesema I.H.M., et al. . Cost-of-illness and disease burden of food-related pathogens in the Netherlands, 2011. International Journal of Food Microbiology 2015; 196:84-93. Manipis K., Mulhern B., Haywood P., Viney R. and Goodall S. Estimating the willingness-to-pay to avoid the consequences of foodborne illnesses: a discrete choice experiment. The European Journal of Health Economics 2023; 24:831-852. Molan A., Nosaka K., Hunter M. and Wang W. Seroprevalence and associated risk factors of Toxoplasma gondii infection in a representative Australian human population: The Busselton health study. Clinical Epidemiology and Global Health 2020; 8:808-814. Mughini Gras L., Smid J.H., Wagenaar J.A., et al. . Risk Factors for Campylobacteriosis of Chicken, Ruminant, and Environmental Origin: A Combined Case-Control and Source Attribution Analysis. PLoS ONE 2012; 7:e42599. Mullner P., Jones G., Noble A., Spencer S.E.F., Hathaway S. and French N.P. Source Attribution of Food-Borne Zoonoses in New Zealand: A Modified Hald Model. Risk Analysis 2009; 29:970-984. O'Hagan A., Buck C.E., Daneshkhah A., et al. . Uncertain judgements: eliciting experts' probabilities . Chichester: Wiley, 2006. OzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2011. Communicable Disease Intelligence 2015; 39:E236-64. OzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2012. Communicable Disease Intelligence 2018; 42. OzFoodNet Working Group. Monitoring the incidence and causes of disease potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2016. Communicable Disease Intelligence 2021a; 45. OzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2013-2015. Communicable Disease Intelligence 2021b; 45. Pires S.M., Evers E.G., Van Pelt W., et al. . Attributing the Human Disease Burden of Foodborne Infections to Specific Sources. Foodborne Pathogens and Disease 2009; 6:417-424. Reddy S.P., Wang H., Adams J.K. and Feng P.C. Prevalence and characteristics of Salmonella serotypes isolated from fresh produce marketed in the United States. Journal of Food Protection 2016; 79:6-16. Rose E.B., Steele M.K., Tolar B., et al. . Attribution of Salmonella enterica to Food Sources by Using Whole-Genome Sequencing Data. Emerging Infectious Diseases 2025; 31. Sapp A.C., Amaya M.P., Havelaar A.H. and Nane G.F. Attribution of country level foodborne disease to food group and food types in three African countries: Conclusions from a structured expert judgment study. PLOS Neglected Tropical Diseases 2022; 16:e0010663. Scallan E., Hoekstra R.M., Angulo F.J., et al. . Foodborne illness acquired in the United States—major pathogens. Emerging infectious diseases 2011; 17:7. Scharff R.L. Economic burden from health losses due to foodborne illness in the United States. Journal of food protection 2012; 75:123-131. Scharff R.L. Food Attribution and Economic Cost Estimates for Meat- and Poultry-Related Illnesses. Journal of Food Protection 2020; 83:959-967. Smith A., Moorhouse E., Monaghan J., Taylor C. and Singleton I. Sources and survival of Listeria monocytogenes on fresh, leafy produce. Journal of Applied Microbiology 2018; 125:930-942. Vally H., Glass K., Ford L., et al. . Proportion of illness acquired by foodborne transmission for nine enteric pathogens in Australia: an expert elicitation. Foodborne pathogens and disease 2014; 11:727-733. Tables Table 1 : Food or commodity with the strongest association with each pathogen included in the expert elicitation with median, 5% and 95% percentile of the distributions. Two foods are provided for Toxoplasma gondii due to very similar median estimates. Hazard Food or commodity Lower bound (5th percentile) Best estimate (50th percentile) Upper bound (95th percentile) Non-typhoidal Salmonella spp. Vegetables 3.7 26.23 48.56 Campylobacter spp. Poultry 40.49 69.18 89.49 Listeria monocytogenes Vegetables 8.964 23.54 39.26 Toxoplasma gondii Beef 5.031 22.05 57.86 Toxoplasma gondii Lamb 1.973 22.46 59.28 STEC Beef 7.315 36.3 62.79 Yersinia spp. Pork 11.83 54.5 81.84 Vibrio spp. Molluscs 51.94 76.65 94.71 Bacillus cereus Grains and seeds 8.104 35.01 75.46 Table 2 : Estimated annual costs circa 2019 for six prioritised pathogens by age group and food commodity (90% uncertainty intervals). Costs are thousands of AUD, inflation adjusted to June 2023 <5 5–64 65+ All Ages † Beef 6,050 (2,770 - 21,300) 32,400 (14,500 - 99,900) 17,200 (7,280 - 62,400) 55,900 (25,500 - 182,000) Crustaceans 1,200 (118 - 10,800) 2,440 (222 - 21,900) 3,170 (316 - 28,200) 7,080 (688 - 64,400) Dairy 7,550 (2,990 - 22,700) 30,800 (11,300 - 94,400) 21,700 (8,390 - 65,700) 60,700 (23,900 - 181,000) Eggs 4,860 (495 - 15,400) 16,300 (1,680 - 50,600) 12,500 (1,280 - 37,100) 34,000 (3,540 - 101,000) Finfish 2,740 (539 - 13,000) 5,250 (1,250 - 25,300) 7,080 (1,360 - 33,400) 15,400 (3,330 - 73,500) Fruit 3,700 (1,100 - 12,700) 9,100 (3,170 - 29,000) 9,530 (2,750 - 33,500) 22,800 (7,390 - 77,100) Grains and seeds 1,070 (81.6 - 6,250) 3,830 (403 - 20,900) 2,550 (161 - 14,200) 7,530 (690 - 39,600) Lamb 4,020 (1,600 - 19,100) 22,800 (8,280 - 86,000) 11,600 (4,100 - 55,900) 38,700 (14,600 - 160,000) Molluscs 1,000 (192 - 10,400) 2,170 (429 - 21,500) 2,630 (508 - 27,400) 5,990 (1,200 - 63,200) Nuts 547 (41.9 - 3,660) 1,940 (202 - 13,400) 1,310 (81.7 - 8,010) 3,840 (350 - 23,900) Pork 6,510 (2,810 - 21,600) 30,900 (13,200 - 96,000) 17,800 (6,910 - 62,400) 55,700 (24,100 - 179,000) Poultry 27,800 (14,200 - 47,800) 198,000 (98,000 - 341,000) 99,700 (52,900 - 161,000) 328,000 (170,000 - 537,000) Vegetables 13,900 (6,530 - 29,000) 54,200 (24,600 - 117,000) 38,100 (17,900 - 79,700) 107,000 (51,200 - 221,000) Other 4,560 (1,000 - 17,500) 34,900 (6,310 - 102,000) 15,600 (2,390 - 56,400) 55,600 (10,000 - 176,000) All Food 78,000 (59,300 - 104,000) 404,000 (301,000 - 566,000) 237,000 (190,000 - 304,000) 721,000 (574,000 - 946,000) † Totals reflect the median of model simulations, so may not equal the sum of individual columns or rows, and uncertainty intervals (UIs) of totals will typically be narrower than the sum of the component UIs. Cost for ‘All Food’ is 100% of cost of foodborne disease for the six pathogens and differs to sums across food commodities as sums of attribution proportions across food commodities do not sum to exactly 100% for all pathogens. Additional Declarations The authors declare no competing interests. Supplementary Files SM1and2CostingElicitationFSANZ.docx Supplementary Materials 1 SM3CMTechnicalDetails.pdf Supplementary Materials 2 and 3 GlobalWeightOptimisedAggregation.xlsx Global Weight Optimised Aggregations GlobalWeightAggregation.xlsx Global Weight Aggregations EqualWeightAggregation.xlsx Equal Weight Aggregations Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8110329","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":544841554,"identity":"5a8afe85-5aea-489b-80f2-5492a4d9ed77","order_by":0,"name":"Anca Hanea","email":"","orcid":"https://orcid.org/0000-0003-3870-5949","institution":"University of Melbourne","correspondingAuthor":false,"prefix":"","firstName":"Anca","middleName":"","lastName":"Hanea","suffix":""},{"id":544841627,"identity":"dc169d51-2ac3-4639-9853-7e1892b73084","order_by":1,"name":"Angus McLure","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIiWNgGAWjYDADPhDxAcI2IE4LGxAzziBZCzMPMVr4Z7dfk/zBsE2eTSL34WObmjuJDezN2yQYag7j1CJx50yZNA/DbcM2iXRj45xjzxIbeI6VSTAcw62F4UZOmjQDw+0ENok0NukctsOJDRI5ZhIMbLi1yAO1AB0G1sL+2+IfUIv8G6CWf7i1GNxIPybBA7WFmbENZAuPmQSQgVOL4Y0cZmseA6BfeJ4xS/b2HTZu40krtkjsS8epRe5G+sObPypuy/OzpzF++PHtsGw/++GNNz58s8btfQYeA9SIAEUQQwIeDQwM7A/wSo+CUTAKRsEoYAAA2VFNPzwE3kAAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-2890-2703","institution":"Australian National University","correspondingAuthor":true,"prefix":"","firstName":"Angus","middleName":"","lastName":"McLure","suffix":""},{"id":544841684,"identity":"00180277-8c5c-44bd-9c0c-57932fca1abe","order_by":2,"name":"Ben Daughtry","email":"","orcid":"","institution":"Food Standards Australian New Zealand","correspondingAuthor":false,"prefix":"","firstName":"Ben","middleName":"","lastName":"Daughtry","suffix":""},{"id":544841722,"identity":"96000492-8d42-48c7-8f6f-87628275f309","order_by":3,"name":"Snezana Smiljanic","email":"","orcid":"","institution":"Food Standards Australia New Zealand","correspondingAuthor":false,"prefix":"","firstName":"Snezana","middleName":"","lastName":"Smiljanic","suffix":""},{"id":544841847,"identity":"421cf632-3b1a-4d4c-825d-a35f9fe07bce","order_by":4,"name":"Kathryn Glass","email":"","orcid":"https://orcid.org/0000-0001-5905-1310","institution":"Australian National University","correspondingAuthor":false,"prefix":"","firstName":"Kathryn","middleName":"","lastName":"Glass","suffix":""}],"badges":[],"createdAt":"2025-11-14 03:40:18","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8110329/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8110329/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96062223,"identity":"c628f460-24aa-4b1a-a0d2-ef1ba4121643","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":118169,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptPreprintNoFigures1.docx","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/ce00970f277b18d12104f718.docx"},{"id":96248210,"identity":"27cd41bf-4b31-4214-9a73-eca33ba3feb1","added_by":"auto","created_at":"2025-11-19 07:28:10","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs8110329.json","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/86c18477aad9a90674891e28.json"},{"id":96062232,"identity":"7d2b8feb-b90e-4d37-9282-1128c487fed1","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102143,"visible":true,"origin":"","legend":"","description":"","filename":"rs81103290enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/1014c78bff3b64c94385010d.xml"},{"id":96062229,"identity":"32065972-bc77-41c3-948b-5eff1221e9dd","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"xml","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97227,"visible":true,"origin":"","legend":"","description":"","filename":"rs81103290structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/60cd25764ad52687b5f5b77a.xml"},{"id":96062230,"identity":"1c79ff46-9fd2-42c6-ae3a-a9103d68034c","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"html","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108521,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/c5b6227e21e7a66644bfd3a2.html"},{"id":96062221,"identity":"66d81952-6f69-450a-8b5a-4c3d3e571d98","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":22470,"visible":true,"origin":"","legend":"\u003cp\u003eThe calibration and informativeness scores for the 10 experts and the 3 aggregations. The vertical line marks the 0.05 calibration level (considered the minimal \"good\" calibration score), the horizontal line marks an arbitrary threshold for informativeness, which in this case differentiates the equally weighted aggregation from the performance-based aggregation.\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/d2d533f0061ea5dce45b24a3.png"},{"id":96062222,"identity":"55ba0216-4b4f-46dd-a587-280887bccef1","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":362077,"visible":true,"origin":"","legend":"\u003cp\u003eEstimated annual cases, hospitalisations, deaths, and costs (circa 2019) arising from foodborne cases for six pathogens, attributed to thirteen food categories and ‘other’. For clarity, the total of each outcome has been plotted on a separate horizontal scale. Costs have been inflation adjusted to June 2023. Cost for ‘All Food’ is 100% of cost of foodborne disease for the six pathogens and differs to sums across food commodities as sums of attribution proportions across food commodities do not sum to exactly 100% for all pathogens. Numerical results can be found in Tables S11-S17.\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/8cca51971ffe0a98d762df51.png"},{"id":96256009,"identity":"8a25952e-2be8-46e6-a11a-cce755e092d0","added_by":"auto","created_at":"2025-11-19 07:49:16","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1161765,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/4e46f3e1-004d-430c-8520-986387642e79.pdf"},{"id":96062226,"identity":"3c2f4234-1edc-42b9-8eec-1bea77f565dc","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":262998,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Materials 1\u003c/p\u003e","description":"","filename":"SM1and2CostingElicitationFSANZ.docx","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/c2c8ccd8996017ce55cef7ee.docx"},{"id":96062227,"identity":"9ea503cc-21f5-45a3-90ee-e208a2ecd982","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":186563,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Materials 2 and 3\u003c/p\u003e","description":"","filename":"SM3CMTechnicalDetails.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/ab940a545c454d3e8b8961ff.pdf"},{"id":96247265,"identity":"f06a9be7-010d-425d-81fd-10799dc469ea","added_by":"auto","created_at":"2025-11-19 07:27:19","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":68093,"visible":true,"origin":"","legend":"\u003cp\u003eGlobal Weight Optimised Aggregations\u003c/p\u003e","description":"","filename":"GlobalWeightOptimisedAggregation.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/50a576c1d93ea0e1f43c9735.xlsx"},{"id":96247580,"identity":"468d17c2-95ba-47a5-a61a-464e4b2167c7","added_by":"auto","created_at":"2025-11-19 07:27:34","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":53927,"visible":true,"origin":"","legend":"\u003cp\u003eGlobal Weight Aggregations\u003c/p\u003e","description":"","filename":"GlobalWeightAggregation.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/a31d174ab5dc0159a4fcba85.xlsx"},{"id":96062228,"identity":"f9c94426-ed5a-42a6-a4ee-f30c9a149780","added_by":"auto","created_at":"2025-11-17 08:44:37","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":60507,"visible":true,"origin":"","legend":"\u003cp\u003eEqual Weight Aggregations\u003c/p\u003e","description":"","filename":"EqualWeightAggregation.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8110329/v1/83efee00cc8d80c8836a02e0.xlsx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eAttribution of pathogen-specific costs of foodborne illness to food commodity groups - combining a costing model with expert judgement\u003c/p\u003e","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eCirca 2019, foodborne illness was estimated to cost Australia 2.44\u0026nbsp;billion Australian dollars (AUD) annually, with AUD 365\u0026nbsp;million attributed to \u003cem\u003eCampylobacter\u003c/em\u003e and AUD 140\u0026nbsp;million to \u003cem\u003eSalmonella\u003c/em\u003e (Glass and others 2023). While pathogen-specific costs highlight high-impact pathogens, attributing costs to specific foods helps target interventions. Combined with cost\u0026ndash;benefit analyses, such estimates guide industry, government, and consumers toward policies that most effectively reduce the foodborne disease burden (Butler and others 2015; Sapp and others 2022; Scharff \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eSeveral approaches have been used to attribute foodborne illness to food groups or animal reservoirs (Batz and others 2012; Kumagai and others 2020). Mathematical source attribution models combine typed isolates from humans with typed isolates from foods and food-producing animals to estimate the proportion of cases attributable to each reservoir (Hald and others 2004; Mughini Gras and others 2012; Mullner and others 2009; Pires and others 2009). Such approaches require large datasets and generally identify animal reservoirs rather than human infection pathways. An alternative approach to source attribution is expert elicitation (Butler and others 2015). Experts have been used to determine the proportion of foodborne illness cases attributable to different transmission routes (Beshearse and others 2021; Havelaar and others 2008; Vally and others 2014). Evidence of transmission routes contributes to burden of disease studies (Daniel and others 2020; Havelaar and others 2015; Lagerweij and others 2020; Scallan and others 2011).\u003c/p\u003e\u003cp\u003eExpert elicitation gathers estimates of unknown quantities from experts. Structured expert judgement (SEJ) elicitation protocols are a collection of steps effective for working with experts' uncertainty estimates. While SEJ lacks an official definition, Cooke provides a working one (Cooke \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e1991\u003c/span\u003e) highlighting that \u0026ldquo;structured\u0026rdquo; expert judgement should be treated as scientific data, and formulates four necessary principles for SEJ as a scientific method:\u003c/p\u003e\u003cp\u003e\u003cul\u003e\u003cli\u003e\u003cp\u003e\u003cem\u003eAccountability\u003c/em\u003e: All data, analyses, and software are peer-reviewed, ensuring reproducibility for reviewers while keeping expert identities confidential.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cem\u003eEmpirical control\u003c/em\u003e: Experts' performance as uncertainty assessors is measured based on calibration variables (variables for which true values exist).\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cem\u003eNeutrality\u003c/em\u003e: The evaluation and aggregation of expert assessments encourage experts to state their true opinion.\u003c/p\u003e\u003c/li\u003e\u003cli\u003e\u003cp\u003e\u003cem\u003eFairness\u003c/em\u003e: Experts are not pre-judged prior to empirical control.\u003c/p\u003e\u003c/li\u003e\u003c/ul\u003e\u003c/p\u003e\u003cp\u003eAn expert elicitation process using Cooke\u0026rsquo;s Classical Model (CM) method was undertaken in 2020 to estimate the attribution of foodborne illness to specific foods for five pathogens: non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e, Shiga-toxin producing \u003cem\u003eEscherichia coli\u003c/em\u003e (STEC), \u003cem\u003eListeria monocytogenes\u003c/em\u003e, hepatitis A virus and norovirus (Food Standards Australia New Zealand \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). The current study extends this elicitation to eight pathogens: non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e, \u003cem\u003eCampylobacter\u003c/em\u003e, STEC, \u003cem\u003eL\u003c/em\u003e. \u003cem\u003emonocytogenes\u003c/em\u003e, \u003cem\u003eY\u003c/em\u003e. \u003cem\u003eenterocolitica\u003c/em\u003e, \u003cem\u003eVibrio\u003c/em\u003e spp. and \u003cem\u003eB\u003c/em\u003e. \u003cem\u003ecereus\u003c/em\u003e and fourteen food commodities, and combines these findings with an existing cost of illness model for six of the eight pathogens (excluding \u003cem\u003eVibrio\u003c/em\u003e spp. and \u003cem\u003eB\u003c/em\u003e. \u003cem\u003ecereus\u003c/em\u003e) to attribute costs of foodborne illness to food commodity groups in Australia.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003ch3\u003e2.1.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;SEJ protocol\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eSEJ protocols include the IDEA protocol (Hanea and others 2017) and the CM, used here in conjunction. SEJ\u0026nbsp;protocols use some structured steps, like independent estimation, anonymisation, controlled group discussion, \u0026nbsp;to mitigate cognitive biases (anchoring, overconfidence) and group biases (dominance, conformity) that can distort judgments [e.g.,\u0026nbsp;Cooke (1991); Hemming and others (2018); O\u0026apos;Hagan and others (2006)].\u003c/p\u003e\n\u003cp\u003eBriefly, IDEA \u0026nbsp;involves four stages:\u003c/p\u003e\n\u003cp\u003e\u0026middot; \u003cem\u003eInvestigate\u003c/em\u003e: Experts work individually to provide point and interval (low\u0026ndash;high) estimates of unknown quantities, following a group meeting to clarify terms, resolve ambiguities, and confirm the process. Experts\u0026rsquo; estimates are de-identified and summarised by facilitators for the discussion stage.\u003c/p\u003e\n\u003cp\u003e\u0026middot; \u003cem\u003eDiscuss\u003c/em\u003e: \u0026nbsp;Summaries are shared and discussed in a workshop. Experts discuss the spread of results, disagreements, and reasoning.\u003c/p\u003e\n\u003cp\u003e\u0026middot; \u003cem\u003eEstimate\u003c/em\u003e: Experts may then privately revise their estimates, though updates are optional and need not be disclosed.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026middot; \u003cem\u003eAggregate\u003c/em\u003e: \u0026nbsp;The revised estimates are converted into probability distributions per expert, then mathematically combined (via linear pooling with equal or weighted contributions).\u003c/p\u003e\n\u003cp\u003eIndependent initial estimates prevent anchoring on others\u0026apos; estimates (likely in unstructured group discussions), the discussion fosters counterfactual thinking and addresses the availability bias, and the second anonymous round of estimates limits the dominance effect and groupthink.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBecause the aggregation stage is done mathematically, CM\u0026rsquo;s mathematical apparatus can be added to an IDEA elicitation, by eliciting calibration questions, allowing for the calculation of performance-based weights to be used in the linear pooling of experts\u0026apos; distributions. Answers to the calibration questions are used to gauge the \u0026ldquo;quality\u0026rdquo; of expert judgements, by measuring performance. The performance of the experts on the calibration questions is taken as indicative of their performance on the target questions. Therefore, the calibration questions must capture the type of knowledge and reasoning needed to answer the target questions. The more calibration variables the better, but ten has proven to be sufficient (Colson and Cooke 2017).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePerformance-based weights reflect how calibrated and informative expert judgements are. A good probability assessor is one whose assessments capture the true values consistently in the long run (well calibrated), with distributions that are as narrow as possible (informative). We used the calibration and informativeness measures defined in (Cooke 1991). Calibration (ranging from 0 to 1) was measured as the p-value at which the hypothesis that the expert is well calibrated would be falsely rejected. Higher scores are better, since low values mean that it is very unlikely that the discrepancy between an expert\u0026apos;s judgements and the observed outcomes arose by chance. Informativeness is measured using the Kullback-Leibler divergence (Kullback and Leibler 1951) with respect to the uniform distribution (which is what one would assume in absence of experimental data). Informativeness scores are non-negative, with higher scores being better (Hanea and Nane 2021).\u003c/p\u003e\n\u003cp\u003eCalibration and informativeness scores were combined (multiplied and normalised) to form weights and performance-based aggregations.\u003c/p\u003e\n\u003ch4\u003e2.1.1.\u0026nbsp; \u0026nbsp;\u0026nbsp;Details of the present elicitation\u0026nbsp;\u003c/h4\u003e\n\u003cp\u003eThe elicitation scope, target and calibration questions were developed with the expert input of external experts (please see Acknowledgements). Target questions covered the attribution of foodborne illness to specific foods for eight pathogens (listed in Introduction).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe food groups were developed from published sources (Hoffmann 2017, Painter 2009) and existing knowledge, and they are: beef, lamb, pork, poultry, eggs, dairy (milk and cream, fresh uncured cheese, brined cheese, soft-ripened cheese, firm-ripened cheese), finfish, crustaceans, molluscs, fruit, grains and seeds, nuts, vegetables (fungi, leafy vegetables and herbs, root vegetables, sprouts, vine-stalk), and \u0026lsquo;other\u0026rsquo;.\u003c/p\u003e\n\u003ch4\u003e2.1.2.\u0026nbsp; \u0026nbsp;\u0026nbsp;Process\u0026nbsp;\u003c/h4\u003e\n\u003cp\u003eThirteen experts (listed in Table (S1) from Supplementary Material, SM) were identified and approached, and eleven participated. Their expertise covered health science, environmental and food microbiology, biosecurity, seafood safety, zoonotic disease epidemiology, food allergen risk management, and environmental health management. Experts signed Consent forms, and were briefed on the expectations via a Plain Language Statement.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDuring an \u0026nbsp;inception meeting the scope and steps of the elicitation were reiterated. Experts reviewed and provided feedback on the food commodity/pathogen matrix, a practice elicitation question and the calibration questions. The first round estimates for calibration questions were elicited during this meeting, in a controlled environment. Expert feedback on the commodity/pathogen matrix was incorporated, and target questions were finalised and sent post meeting. Experts had three weeks to answer the target questions and were instructed to not research the (already answered) calibration questions. Their Judgements were collected via email, de-identified, and collated.\u003c/p\u003e\n\u003cp\u003eThe facilitated workshop (the \u003cem\u003eDiscuss\u003c/em\u003e stage) comprised two recorded online sessions with ten experts each session, and eleven experts attending at least once. All questions were discussed and refined as needed. The de-identified data and the recordings were sent to two experts who missed both sessions. Following the workshop, experts reconsidered their estimates.\u003c/p\u003e\n\u003ch4\u003e2.1.3.\u0026nbsp; \u0026nbsp;\u0026nbsp;Questions\u003c/h4\u003e\n\u003cp\u003eTo acknowledge uncertainty, we elicited point estimates together with upper and lower bounds, eliciting the lower bound first, followed by the upper bound, and finally the best estimate.\u0026nbsp;\u0026nbsp;This question order aims to reduce anchoring and overconfidence (narrow intervals), though empirical evidence for its effectiveness in this exact context is limited.\u0026nbsp;We asked 12 calibration questions (listed in SM) and 72 target questions corresponding to pairs of pathogens and food commodities. An example target question set is shown in Figure (S1).\u003c/p\u003e\n\u003ch4\u003e2.1.4.\u0026nbsp; \u0026nbsp;\u0026nbsp;Aggregated estimates\u003c/h4\u003e\n\u003cp\u003eWe described the probability distribution per question, per expert, using the revised estimates, with best estimates set to be the medians of distributions and the upper and lower bounds set to represent 5% and 95% percentiles.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNon-parametric distributions were then fitted on the three percentiles (5%, median, 95%). For technical details we refer to the supplementary material SM3.\u003c/p\u003e\n\u003cp\u003eAn aggregated distribution per question was obtained by combining distributions across experts linearly with \u003cem\u003eequal\u003c/em\u003e or \u003cem\u003edifferential\u003c/em\u003e weights. \u0026nbsp;We used the calibration questions to calculate calibration and informativeness scores. These scores were used to calculate differential weights, sometimes called global weights (proportional to the product of the two scores) which were further used to calculate a performance based weighted distribution that represents the group\u0026rsquo;s judgements. A variant of the globally weighted aggregation is the optimised globally weighted aggregation. The distinction between the two methods is detailed in SM3. Briefly, the optimisation sequentially makes small weights zero until the aggregation of the remaining distributions (with non-zero weights) reaches the best combined score.\u003c/p\u003e\n\u003ch3\u003e2.2.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Costing model\u003c/h3\u003e\n\u003cp\u003eThe total burden and cost estimation for the six pathogens is identical to previous methods detailed elsewhere (Glass and others 2023), though here we report inflation-adjusted costs from December 2019 to June 2023 (+15.1%). Briefly, we estimated burden in terms of the annual number of cases, hospitalisations, and deaths due the pathogens and four sequelae arising from them: Guillain-Barr\u0026eacute; syndrome (GBS), Haemolytic uraemic syndrome (HUS), irritable bowel syndrome (IBS), and Reactive arthritis (ReA). Disease burden estimates followed prior approaches (Ford and others 2014; Kirk and others 2014), including the selection of sequelae. For each pathogen and sequelae, we estimated costs from premature mortality, lost productivity, healthcare, and non-financial costs of pain and suffering.\u003c/p\u003e\n\u003cp\u003eFor the five notifiable pathogens (\u003cem\u003eCampylobacter,\u0026nbsp;\u003c/em\u003enon-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e,\u003cem\u003e\u0026nbsp;Yersinia enterocolitica, Listeria monocytogenes\u003c/em\u003e, and STEC) the number of cases was estimated as the number of notifications in 2019 multiplied by underreporting multipliers (Hall and others 2008).\u003cem\u003e\u0026nbsp;\u003c/em\u003eFor \u003cem\u003eToxoplasma gondii,\u003c/em\u003e which is not notifiable, we estimated annual number of cases from seroprevalence curves by age (Molan and others 2020), assuming 15% of exposures that lead to seroconversion result in symptomatic cases, consistent with prior work (Glass and others 2023; Kirk and others 2014). We estimated the number of sequelae by multiplying initial cases by the risk of sequelae determined previously (Ford and others 2014). All cases of listeriosis, GBS, and HUS were assumed to be hospitalised. For all other conditions, we estimated hospitalisations using publicly available administrative healthcare data on hospitalisation by principal diagnosis (Australian Government Institute of Health and Welfare 2020), with adjustments for underdiagnosis and the fraction of admissions where the condition was a secondary diagnosis (Kirk and others 2014). We estimated annual deaths by averaging reported data for 2001-2010 from ABS (Australian Bureau of Statistics 2010), with adjustment for underdiagnosis (Kirk and others 2014) and additional listeriosis deaths in neonates from OzFoodNet annual reports (OzFoodNet Working Group 2015; 2018; 2021a; 2021b). All burden estimates were multiplied by the proportion of cases that are foodborne as estimated in a previous expert elicitation (Vally and others 2014) and the number of incident cases were multiplied by the proportion that were domestically acquired (Kirk and others 2014).\u003c/p\u003e\n\u003cp\u003ePremature mortality was costed using the Value of Statistical Life (Australian Government Department of Prime Minister and Cabinet: Office of Best Practice Regulation) applied to each death regardless of age. We sourced healthcare costs from the Medicare Benefits Schedule, the Pharmaceutical Benefits Scheme, and the Australian Refined Diagnosis Related Groups. We costed lost productivity for cases and their carers using published workforce participation data (Australian Bureau of Statistics), estimates of days of work missed from the National Gastroenteritis survey (Kirk 2008) for non-hospitalised cases, and the average length of stay for hospitalised cases (Australian Government Institute of Health and Welfare). Non-financial costs were costed using a previous discrete choice experiment which quantified willingness to pay to avoid pain and suffering (Manipis and others 2023).\u003c/p\u003e\n\u003cp\u003eTo estimate burden/costs attributable to specific food categories for each pathogen, we multiplied our previously published estimates for each pathogen (Glass and others 2023) by the proportion of cases of that pathogen attributable to each food category, i.e. we assumed that outcomes (hospitalisations, sequelae, deaths) and costs did not vary by the food category that led to the infection. All estimates were stratified by age group (\u0026lt;5, 5-64, 65+). Uncertainty in attribution proportions, multipliers, and other inputs was incorporated into final estimates using Monte Carlo simulations, and is reported as 90% credible intervals.\u0026nbsp;\u003c/p\u003e"},{"header":"3. Results","content":"\u003ch3\u003e3.1.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;Expert elicitation\u003c/h3\u003e\n\u003cp\u003eTen experts completed the revised estimates, which were aggregated using three aggregation schemes: equal weighting, global weighting without optimisations and with optimisation (see the Excel spreadsheets uploaded as SM for details ). The optimisation procedure selected one expert to be the optimum \u0026ldquo;aggregation\u0026rdquo;. \u0026nbsp;Although the calibration score of this expert is larger than the calibration of the other two aggregations, the differences in the aggregated distributions are minimal (see Hanea and Nane (2021) for examples of apparent large differences in calibration scores that generated insignificant changes in distributions). The large informativeness score may be indicative of overconfidence, so we choose \u003cem\u003enot\u003c/em\u003e to use the optimised version of the globally-weighted aggregation. Although the calibration scores of the equally and globally-weighted aggregations were equal, we used the globally-weighted aggregation for two reasons: 1) the large differences between experts\u0026rsquo; calibration scores, and 2) the larger informativeness score of the latter. Figure (1) shows the calibration and informativeness scores of all experts and aggregations. The vertical line marks the 0.05 calibration level (often considered the minimal \u0026quot;good\u0026quot; calibration score). For the exact scores see Table (S2) from SM1.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable (1) provides the distributions obtained using the chosen aggregation, including examples of best estimates and bounds for the food/commodity with the largest contribution per pathogen. Particularly strong associations were found between \u003cem\u003eCampylobacter\u0026nbsp;\u003c/em\u003eand poultry (median 69%) and \u003cem\u003eYersinia\u0026nbsp;\u003c/em\u003eand pork (54%). The complete aggregated data are provided in Tables (S3) and (S4).\u0026nbsp;\u003c/p\u003e\n\u003ch3\u003e3.2. Attributed burden and costs\u0026nbsp;\u003c/h3\u003e\n\u003cp\u003eAcross the six pathogens considered, the overall annual burden was 357,000 cases (Figure (2), Table (S5)) and total cost was AUD 721 million (inflation-adjusted to June 2023, Table 2). Campylobacteriosis was the cause of the most cases (264000; Figure (2), Table (S5)) and had the highest overall cost (AUD 420 million; Figure (2), Table (S6)). \u003cem\u003eCampylobacter\u0026nbsp;\u003c/em\u003ewas the leading cause of hospitalisation, with 5,640 hospitalisations due to initial disease and a further 598 due to ReA, 2,300 due to IBS and 98 due to GBS (Figure (2)). Listeriosis, although causing far fewer cases and hospitalisations (101), was the leading cause of deaths (fifteen; Figure (2)). Extensive additional details on burden and costs can be found in our previous publication on this model (Glass and others 2023).\u003c/p\u003e\n\u003cp\u003ePoultry had the highest foodborne disease burden across cases, hospitalisations, deaths, and cost (Figure (2)). Of the AUD 328 million cost attributed to poultry, AUD 279 million was from \u003cem\u003eCampylobacter,\u0026nbsp;\u003c/em\u003e35.5 million from non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e, and 2.9 million from \u003cem\u003eL.\u0026nbsp;monocytogenes.\u003c/em\u003e Of the 191 thousand cases attributed to poultry, 174 thousand were attributed to \u003cem\u003eCampylobacter\u003c/em\u003e and 13.2 thousand to non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e. Vegetables was the second highest category associated with cases, hospitalisations, deaths, and costs (AUD 107 million), with non-typhoidal \u003cem\u003eSalmonella\u0026nbsp;\u003c/em\u003ethe leading pathogen (AUD 41.7 million). Dairy, beef, and pork had the next highest costs, each category costing over AUD 55 million. Full details of estimated cases, deaths, hospitalisations, and costs can be found in Tables (S12-S18).\u003c/p\u003e\n\u003cp\u003eTable (2) provides a breakdown of costs by age group. As the elicitation was not stratified by age group, age-specific differences reflect disease burden rather than varying food-pathogen associations.\u0026nbsp;Total costs across all food commodities were highest in the 5\u0026ndash;64-year-olds (404 million) and lowest in children under 5 (78 million), reflecting the relative population in each group. However, for crustaceans, finfish, molluscs, and fruit, attributed costs were highest in people over 65, as listeriosis accounted for most costs associated with these commodities (Figure 2) and the cost of listeriosis was highest in this age group (Table (S7)).\u003c/p\u003e\n\u003cp\u003eSummaries of cases for each food category and detailed breakdowns of costs by food category, pathogen, and age group are presented in the supplementary materials, Tables (S5-S11).\u0026nbsp;\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe six foodborne pathogens included in the model were estimated to cost AUD 721\u0026nbsp;million annually, with almost half of this cost (AUD 328\u0026nbsp;million) attributed to poultry. Other food categories with high costs included vegetables (AUD 107\u0026nbsp;million), dairy (AUD 61\u0026nbsp;million), beef (AUD 56\u0026nbsp;million) and pork (AUD 56\u0026nbsp;million) with AUD 56\u0026nbsp;million attributed to additional \u0026lsquo;other\u0026rsquo; categories. High costs, case numbers, and hospitalisations attributed to poultry were largely from the strong association between poultry and \u003cem\u003eCampylobacter.\u003c/em\u003e Food categories with high numbers of attributed deaths included poultry, vegetables and dairy, from associations with \u003cem\u003eCampylobacter, Listeria monocytogenes\u003c/em\u003e, and non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e.\u003c/p\u003e\u003cp\u003eExperts identified relatively strong associations between vegetables and both non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e and \u003cem\u003eListeria monocytogenes\u003c/em\u003e, with vegetables having the highest attribution proportion for both pathogens (median of 26% for \u003cem\u003eSalmonella\u003c/em\u003e and 24% for \u003cem\u003eListeria monocytogenes\u003c/em\u003e). Salmonellosis and listeriosis are often associated with outbreaks of fresh produce, commonly vegetables (Reddy and others 2016; Smith and others 2018), although a review of foodborne outbreaks in Australia over 2001\u0026ndash;2016 found stronger associations between salmonellosis outbreaks and eggs (Ford and others 2018). Internationally, non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e is often associated with poultry or eggs (Batz and others 2012; Davidson and others 2011; Kumagai and others 2020; Sapp and others 2022; Scharff \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), although associations have been found with vegetables (Rose and others 2025). While poultry and eggs each had a lower attribution proportion than vegetables in this elicitation, they combined to a higher total proportion. \u003cem\u003eListeria monocytogenes\u003c/em\u003e has been associated with both luncheon meats (Batz and others 2012; Davidson and others 2011) and dairy (Mangen and others 2015). While we have not included luncheon meat as a category, dairy had the second highest association with listeriosis after vegetables.\u003c/p\u003e\u003cp\u003eMany of our findings are consistent with pathogen-food associations identified elsewhere. The strong association between \u003cem\u003eCampylobacter\u003c/em\u003e and poultry has been found in the United States (Batz and others 2012; Scharff \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2020\u003c/span\u003e), Canada (Davidson and others 2011), Japan (Kumagai and others 2020), in an expert elicitation in three African countries (Sapp and others 2022), and in a global attribution study (Hoffmann and others 2017). Strong associations have been found between \u003cem\u003eE. coli\u003c/em\u003e and beef (Batz and others 2012; Davidson and others 2011; Kumagai and others 2020; Mangen and others 2015), as has \u003cem\u003eYersinia\u003c/em\u003e and pork (Batz and others 2012; Davidson and others 2011). Our finding of a relatively low cost associated with \u003cem\u003eToxoplasma gondii\u003c/em\u003e and associations with beef and lamb differ from international studies. A study in the United States identified pork and \u003cem\u003eT. gondii\u003c/em\u003e as the second most important pathogen-food combination in most food rankings (Batz and others 2012), while pork was identified as contributing around half the cost from \u003cem\u003eT. gondii\u003c/em\u003e in a study in the Netherlands (Mangen and others 2015). Our model of \u003cem\u003eT. gondii\u003c/em\u003e results in relatively low costs per case (Glass and others 2023), with poor data quality limiting our ability to refine these estimates.\u003c/p\u003e\u003cp\u003eA limitation of our work is that only six of ten pathogens (available in the costing model) were included in the elicitation due to time constraints, and that the elicitation did not consider differences in pathogen-food associations by age. A priority for further work is conducting an elicitation for norovirus, which was estimated to cost Australia AUD 128\u0026nbsp;million circa 2019 (Glass and others 2023). Owing to high levels of person-to-person transmission for norovirus (Vally and others 2014), this elicitation may need to repeat an assessment of transmission pathways to ensure pathogen-food associations are appropriately contextualised. Although \u003cem\u003eCryptosporidium\u003c/em\u003e was not prioritised for inclusion in our costing model, it is commonly included in international studies and could be considered in future modelling.\u003c/p\u003e\u003cp\u003eA key strength of this work is the combination of a detailed costing model with SEJ, which enables costing of food contributions to foodborne disease. While SEJ can improve group judgements, it cannot guarantee perfect estimates. Using calibration questions helps, but only under the assumption that the calibration questions are representative for the target questions. Another untested aspect of using SEJ protocols with facilitated discussion is the effect of a remote discussion (as opposed to a face-to-face setting) on the performance scores.\u003c/p\u003e\u003cp\u003eNevertheless, SEJ provides a valuable approach to identifying food categories associated with foodborne disease pathogens and, when combined with costing data, enables cost estimations of disease burden. Our analysis for Australia indicated poultry as a key cause of foodborne illness, responsible for almost half of the total costs due to \u003cem\u003eCampylobacter\u003c/em\u003e, Non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e, \u003cem\u003eYersinia enterocolitica\u003c/em\u003e, \u003cem\u003eListeria monocytogenes\u003c/em\u003e, and STEC. The identified costs and food-pathogen links will guide priorities for improving food safety in Australian production systems.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAcknowledgements\u003c/h2\u003e\n\u003cp\u003eWe would like to acknowledge Dr Sandra Hoffmann (Economic Research Service, United States Department of Agriculture), Dr Sara Monteiro Pires (National Food Institute, Technical University of Denmark) and the Food Safety \u0026amp; Microbiology team at Food Standards Australia New Zealand (FSANZ) for helping to identify the substantive knowledge needed for formulating the calibration questions. \u0026nbsp;Thirteen experts completed all rounds of the expert elicitation process: Kate Astridge, Mark Chan, Karen Ferres, Stacy Kane, Thea King, Allison McNamara, Stewart Quinn, Tom Ross, Robin Sherlock, Henry Tan, Alison Turnbull, Mark Turner, and Helen Withers. \u0026nbsp;This work was funded by FSANZ under contract 2022-23/37.\u003c/p\u003e\n\u003ch2\u003eAuthorship contribution statement\u003c/h2\u003e\n\u003cp\u003eAnca Hanea, Angus McLure, Ben Daughtry, Snezana Smiljanic, Kathryn Glass \u003cstrong\u003eConceptualization, Methodology, Writing- Reviewing and Editing\u003c/strong\u003e, Anca Hanea, Angus McLure, Kathryn Glass: \u003cstrong\u003eData curation, Formal analysis, Writing- Original draft preparation\u003c/strong\u003e\u003c/p\u003e\n\u003ch2\u003eConflict of interest\u003c/h2\u003e\n\u003cp\u003eThe Authors declare no conflicts of interest.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eFunding Statement\u003c/h2\u003e\n\u003cp\u003eThis work was funded by Food Standards Australia New Zealand (FSANZ) under contract 2022-23/37\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAustralian Bureau of Statistics. Customised Report: Causes of Death 2001-2010. 2010.\u003c/li\u003e\n\u003cli\u003eAustralian Bureau of Statistics. Labour Force, Australia, Detailed: Detailed monthly and quarterly Labour Force Survey data, including hours, regions, families, job search, job duration, casual, industry and occupation, Sept 2021. 2021.\u003c/li\u003e\n\u003cli\u003eAustralian Government Department of Prime Minister and Cabinet: Office of Best Practice Regulation. Best Practice Regulation Guidance Note: Value of Statistical Life; 2019. 2019.\u003c/li\u003e\n\u003cli\u003eAustralian Government Institute of Health and Welfare. Principal Diagnosis data cubes 2020. 2020.\u003c/li\u003e\n\u003cli\u003eBatz M.B., Hoffmann S. and Morris Jr J.G. Ranking the disease burden of 14 pathogens in food sources in the United States using attribution data from outbreak investigations and expert elicitation. Journal of food protection 2012; 75:1278-1291.\u003c/li\u003e\n\u003cli\u003eBeshearse E., Bruce B.B., Nane G.F., et al. . Attribution of Illnesses Transmitted by Food and Water to Comprehensive Transmission Pathways Using Structured Expert Judgment, United States. Emerging Infectious Diseases 2021; 27:182-195.\u003c/li\u003e\n\u003cli\u003eButler A.J., Thomas M.K. and Pintar K.D.M. Systematic Review of Expert Elicitation Methods as a Tool for Source Attribution of Enteric Illness. Foodborne Pathogens and Disease 2015; 12:367-382.\u003c/li\u003e\n\u003cli\u003eColson A.R. and Cooke R.M. Cross validation for the classical model of structured expert judgment. Reliability Engineering \u0026amp; System Safety 2017; 163:109-120.\u003c/li\u003e\n\u003cli\u003eCooke R.M. \u003cem\u003eExperts in uncertainty: Opinion and subjective probability in science\u003c/em\u003e. New York: Oxford University Press, 1991.\u003c/li\u003e\n\u003cli\u003eCooke R.M. and Goossens L.L.H.J. TU Delft expert judgment data base. Reliability Engineering \u0026amp; System Safety 2008; 93:657-674.\u003c/li\u003e\n\u003cli\u003eDaniel N., Casadevall N., Sun P., Sugden D. and Aldin V. The Burden of Foodborne Disease in the UK, 2018. Food Standards Agency, 2020.\u003c/li\u003e\n\u003cli\u003eDavidson V.J., Ravel A., Nguyen T.N., Fazil A. and Ruzante J.M. Food-Specific Attribution of Selected Gastrointestinal Illnesses: Estimates from a Canadian Expert Elicitation Survey. Foodborne Pathogens and Disease 2011; 8:983-995.\u003c/li\u003e\n\u003cli\u003eFood Standards Australia New Zealand. Primary Production and Processing (PPP) Requirements for Horticulture (Berries, Leafy Vegetables and Melons): Supporting document 2 to the Decision Regulation Impact Statement: Cost-benefit analysis. Volume P1052, 2022.\u003c/li\u003e\n\u003cli\u003eFord L., Kirk M., Glass K. and Hall G. Sequelae of Foodborne Illness Caused by 5 Pathogens, Australia, Circa 2010. Emerging Infectious Diseases 2014; 20:1865-1871.\u003c/li\u003e\n\u003cli\u003eFord L., Moffatt C.R.M., Fearnley E., et al. . The Epidemiology of Salmonella enterica Outbreaks in Australia, 2001\u0026ndash;2016. Frontiers in Sustainable Food Systems 2018; 2.\u003c/li\u003e\n\u003cli\u003eGlass K., McLure A., Bourke S., et al. . The Cost of Foodborne Illness and Its Sequelae in Australia Circa 2019. Foodborne Pathogens and Disease 2023; 20:419-426.\u003c/li\u003e\n\u003cli\u003eHald T., Vose D., Wegener H.C. and Koupeev T. A Bayesian Approach to Quantify the Contribution of Animal‐Food Sources to Human Salmonellosis. Risk Analysis 2004; 24:255-269.\u003c/li\u003e\n\u003cli\u003eHall G., Yohannes K., Raupach J., Becker N. and Kirk M. Estimating Community Incidence of \u003cem\u003eSalmonella\u003c/em\u003e, \u003cem\u003eCampylobacter\u003c/em\u003e, and Shiga Toxin\u0026ndash;producing \u003cem\u003eEscherichia coli\u003c/em\u003e Infections, Australia. Emerging Infectious Diseases 2008; 14:1601-1609.\u003c/li\u003e\n\u003cli\u003eHanea A., McBride M., Burgman M., et al. . I nvestigate D iscuss E stimate A ggregate for structured expert judgement. International Journal of Forecasting 2017; 33:267-279.\u003c/li\u003e\n\u003cli\u003eHanea A.M. and Nane G.F. An In-Depth Perspective on the Classical Model. In\u003cem\u003e Expert Judgement in Risk and Decision Analysis\u003c/em\u003e. Hanea AM, Nane GF, Bedford T, and French S, (eds.). Cham: Springer, 2021, pp 225-256.\u003c/li\u003e\n\u003cli\u003eHavelaar A.H., Galindo \u0026Aacute;.V., Kurowicka D. and Cooke R.M. Attribution of Foodborne Pathogens Using Structured Expert Elicitation. Foodborne Pathogens and Disease 2008; 5:649-659.\u003c/li\u003e\n\u003cli\u003eHavelaar A.H., Kirk M.D., Torgerson P.R., et al. . World Health Organization Global Estimates and Regional Comparisons of the Burden of Foodborne Disease in 2010. PLOS Medicine 2015; 12:e1001923.\u003c/li\u003e\n\u003cli\u003eHoffmann S., Devleesschauwer B., Aspinall W., et al. . Attribution of global foodborne disease to specific foods: Findings from a World Health Organization structured expert elicitation. PLOS ONE 2017; 12:e0183641.\u003c/li\u003e\n\u003cli\u003eKirk M. National Gastroenteritis Survey II (NGSII). 2008.\u003c/li\u003e\n\u003cli\u003eKirk M., Ford L., Glass K. and Hall G. Foodborne Illness, Australia, Circa 2000 and Circa 2010. Emerging Infectious Diseases 2014; 20:1857-1864.\u003c/li\u003e\n\u003cli\u003eKullback S. and Leibler R.A. On information and sufficiency. The annals of mathematical statistics 1951; 22:79-86.\u003c/li\u003e\n\u003cli\u003eKumagai Y., Pires S.M., Kubota K. and Asakura H. Attributing human foodborne diseases to food sources and water in Japan using analysis of outbreak surveillance data. Journal of food protection 2020; 83:2087-2094.\u003c/li\u003e\n\u003cli\u003eLagerweij G.R., Pijnacker R., Friesema I.H.M., Mughini Gras L. and Franz E. Disease burden of food-related pathogens in the Netherlands, 2019. Bilthoven: National Institute for Public Health and the Environment, RIVM, 2020.\u003c/li\u003e\n\u003cli\u003eMangen M.-J.J., Bouwknegt M., Friesema I.H.M., et al. . Cost-of-illness and disease burden of food-related pathogens in the Netherlands, 2011. International Journal of Food Microbiology 2015; 196:84-93.\u003c/li\u003e\n\u003cli\u003eManipis K., Mulhern B., Haywood P., Viney R. and Goodall S. Estimating the willingness-to-pay to avoid the consequences of foodborne illnesses: a discrete choice experiment. The European Journal of Health Economics 2023; 24:831-852.\u003c/li\u003e\n\u003cli\u003eMolan A., Nosaka K., Hunter M. and Wang W. Seroprevalence and associated risk factors of Toxoplasma gondii infection in a representative Australian human population: The Busselton health study. Clinical Epidemiology and Global Health 2020; 8:808-814.\u003c/li\u003e\n\u003cli\u003eMughini Gras L., Smid J.H., Wagenaar J.A., et al. . Risk Factors for Campylobacteriosis of Chicken, Ruminant, and Environmental Origin: A Combined Case-Control and Source Attribution Analysis. PLoS ONE 2012; 7:e42599.\u003c/li\u003e\n\u003cli\u003eMullner P., Jones G., Noble A., Spencer S.E.F., Hathaway S. and French N.P. Source Attribution of Food-Borne Zoonoses in New Zealand: A Modified Hald Model. Risk Analysis 2009; 29:970-984.\u003c/li\u003e\n\u003cli\u003eO\u0026apos;Hagan A., Buck C.E., Daneshkhah A., et al. . \u003cem\u003eUncertain judgements: eliciting experts\u0026apos; probabilities\u003c/em\u003e. Chichester: Wiley, 2006.\u003c/li\u003e\n\u003cli\u003eOzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2011. Communicable Disease Intelligence 2015; 39:E236-64.\u003c/li\u003e\n\u003cli\u003eOzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2012. Communicable Disease Intelligence 2018; 42.\u003c/li\u003e\n\u003cli\u003eOzFoodNet Working Group. Monitoring the incidence and causes of disease potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2016. Communicable Disease Intelligence 2021a; 45.\u003c/li\u003e\n\u003cli\u003eOzFoodNet Working Group. Monitoring the incidence and causes of diseases potentially transmitted by food in Australia: Annual report of the OzFoodNet network, 2013-2015. Communicable Disease Intelligence 2021b; 45.\u003c/li\u003e\n\u003cli\u003ePires S.M., Evers E.G., Van Pelt W., et al. . Attributing the Human Disease Burden of Foodborne Infections to Specific Sources. Foodborne Pathogens and Disease 2009; 6:417-424.\u003c/li\u003e\n\u003cli\u003eReddy S.P., Wang H., Adams J.K. and Feng P.C. Prevalence and characteristics of Salmonella serotypes isolated from fresh produce marketed in the United States. Journal of Food Protection 2016; 79:6-16.\u003c/li\u003e\n\u003cli\u003eRose E.B., Steele M.K., Tolar B., et al. . Attribution of \u0026lt;i\u0026gt;Salmonella enterica\u0026lt;/i\u0026gt; to Food Sources by Using Whole-Genome Sequencing Data. Emerging Infectious Diseases 2025; 31.\u003c/li\u003e\n\u003cli\u003eSapp A.C., Amaya M.P., Havelaar A.H. and Nane G.F. Attribution of country level foodborne disease to food group and food types in three African countries: Conclusions from a structured expert judgment study. PLOS Neglected Tropical Diseases 2022; 16:e0010663.\u003c/li\u003e\n\u003cli\u003eScallan E., Hoekstra R.M., Angulo F.J., et al. . Foodborne illness acquired in the United States\u0026mdash;major pathogens. Emerging infectious diseases 2011; 17:7.\u003c/li\u003e\n\u003cli\u003eScharff R.L. Economic burden from health losses due to foodborne illness in the United States. Journal of food protection 2012; 75:123-131.\u003c/li\u003e\n\u003cli\u003eScharff R.L. Food Attribution and Economic Cost Estimates for Meat- and Poultry-Related Illnesses. Journal of Food Protection 2020; 83:959-967.\u003c/li\u003e\n\u003cli\u003eSmith A., Moorhouse E., Monaghan J., Taylor C. and Singleton I. Sources and survival of Listeria monocytogenes on fresh, leafy produce. Journal of Applied Microbiology 2018; 125:930-942.\u003c/li\u003e\n\u003cli\u003eVally H., Glass K., Ford L., et al. . Proportion of illness acquired by foodborne transmission for nine enteric pathogens in Australia: an expert elicitation. Foodborne pathogens and disease 2014; 11:727-733.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1\u003c/strong\u003e: Food or commodity with the strongest association with each pathogen included in the expert elicitation with median, 5% and 95% percentile of the distributions. Two foods are provided for \u003cem\u003eToxoplasma gondii\u003c/em\u003e due to very similar median estimates.\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHazard \u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFood or commodity\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLower bound (5th percentile)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBest estimate (50th percentile)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUpper bound (95th percentile)\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNon-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e spp.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eVegetables\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e3.7\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e26.23\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e48.56\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eCampylobacter\u003c/em\u003e spp.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003ePoultry\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e40.49\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e69.18\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e89.49\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eListeria monocytogenes\u003c/em\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eVegetables\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e8.964\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e23.54\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e39.26\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eToxoplasma gondii\u003c/em\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eBeef\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e5.031\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e22.05\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e57.86\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eToxoplasma gondii\u003c/em\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eLamb\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e1.973\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e22.46\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e59.28\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSTEC\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eBeef\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e7.315\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e36.3\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e62.79\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eYersinia\u003c/em\u003e spp.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003ePork\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e11.83\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e54.5\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e81.84\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eVibrio\u003c/em\u003e spp.\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eMolluscs\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e51.94\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e76.65\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e94.71\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003cem\u003eBacillus\u003c/em\u003e \u003cem\u003ecereus\u003c/em\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003eGrains and seeds\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 113px;\"\u003e\n \u003cp\u003e8.104\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e35.01\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 120px;\"\u003e\n \u003cp\u003e75.46\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2\u003c/strong\u003e: Estimated annual costs circa 2019 for six prioritised pathogens by age group and food commodity (90% uncertainty intervals). Costs are thousands of AUD, inflation adjusted to June 2023\u003c/p\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026lt;5\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e5\u0026ndash;64\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e65+\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll Ages\u003csup\u003e\u0026dagger;\u003c/sup\u003e\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBeef\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e6,050\u0026nbsp;\u003cbr\u003e\u0026nbsp;(2,770 - 21,300)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e32,400\u0026nbsp;\u003cbr\u003e\u0026nbsp;(14,500 - 99,900)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e17,200\u0026nbsp;\u003cbr\u003e\u0026nbsp;(7,280 - 62,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e55,900\u0026nbsp;\u003cbr\u003e\u0026nbsp;(25,500 - 182,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCrustaceans\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e1,200\u0026nbsp;\u003cbr\u003e\u0026nbsp;(118 - 10,800)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e2,440\u0026nbsp;\u003cbr\u003e\u0026nbsp;(222 - 21,900)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e3,170\u0026nbsp;\u003cbr\u003e\u0026nbsp;(316 - 28,200)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e7,080\u0026nbsp;\u003cbr\u003e\u0026nbsp;(688 - 64,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eDairy\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e7,550\u0026nbsp;\u003cbr\u003e\u0026nbsp;(2,990 - 22,700)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e30,800\u0026nbsp;\u003cbr\u003e\u0026nbsp;(11,300 - 94,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e21,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(8,390 - 65,700)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e60,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(23,900 - 181,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eEggs\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e4,860\u0026nbsp;\u003cbr\u003e\u0026nbsp;(495 - 15,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e16,300\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,680 - 50,600)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e12,500\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,280 - 37,100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e34,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(3,540 - 101,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFinfish\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e2,740\u0026nbsp;\u003cbr\u003e\u0026nbsp;(539 - 13,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e5,250\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,250 - 25,300)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e7,080\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,360 - 33,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e15,400\u0026nbsp;\u003cbr\u003e\u0026nbsp;(3,330 - 73,500)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eFruit\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e3,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,100 - 12,700)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e9,100\u0026nbsp;\u003cbr\u003e\u0026nbsp;(3,170 - 29,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e9,530\u0026nbsp;\u003cbr\u003e\u0026nbsp;(2,750 - 33,500)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e22,800\u0026nbsp;\u003cbr\u003e\u0026nbsp;(7,390 - 77,100)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGrains and seeds\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e1,070\u0026nbsp;\u003cbr\u003e\u0026nbsp;(81.6 - 6,250)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e3,830\u0026nbsp;\u003cbr\u003e\u0026nbsp;(403 - 20,900)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e2,550\u0026nbsp;\u003cbr\u003e\u0026nbsp;(161 - 14,200)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e7,530\u0026nbsp;\u003cbr\u003e\u0026nbsp;(690 - 39,600)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eLamb\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e4,020\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,600 - 19,100)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e22,800\u0026nbsp;\u003cbr\u003e\u0026nbsp;(8,280 - 86,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e11,600\u0026nbsp;\u003cbr\u003e\u0026nbsp;(4,100 - 55,900)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e38,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(14,600 - 160,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eMolluscs\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e1,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(192 - 10,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e2,170\u0026nbsp;\u003cbr\u003e\u0026nbsp;(429 - 21,500)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e2,630\u0026nbsp;\u003cbr\u003e\u0026nbsp;(508 - 27,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e5,990\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,200 - 63,200)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNuts\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e547\u0026nbsp;\u003cbr\u003e\u0026nbsp;(41.9 - 3,660)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e1,940\u0026nbsp;\u003cbr\u003e\u0026nbsp;(202 - 13,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e1,310\u0026nbsp;\u003cbr\u003e\u0026nbsp;(81.7 - 8,010)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e3,840\u0026nbsp;\u003cbr\u003e\u0026nbsp;(350 - 23,900)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePork\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e6,510\u0026nbsp;\u003cbr\u003e\u0026nbsp;(2,810 - 21,600)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e30,900\u0026nbsp;\u003cbr\u003e\u0026nbsp;(13,200 - 96,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e17,800\u0026nbsp;\u003cbr\u003e\u0026nbsp;(6,910 - 62,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e55,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(24,100 - 179,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePoultry\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e27,800\u0026nbsp;\u003cbr\u003e\u0026nbsp;(14,200 - 47,800)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e198,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(98,000 - 341,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e99,700\u0026nbsp;\u003cbr\u003e\u0026nbsp;(52,900 - 161,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e328,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(170,000 - 537,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVegetables\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e13,900\u0026nbsp;\u003cbr\u003e\u0026nbsp;(6,530 - 29,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e54,200\u0026nbsp;\u003cbr\u003e\u0026nbsp;(24,600 - 117,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e38,100\u0026nbsp;\u003cbr\u003e\u0026nbsp;(17,900 - 79,700)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e107,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(51,200 - 221,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOther\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e4,560\u0026nbsp;\u003cbr\u003e\u0026nbsp;(1,000 - 17,500)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e34,900\u0026nbsp;\u003cbr\u003e\u0026nbsp;(6,310 - 102,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e15,600\u0026nbsp;\u003cbr\u003e\u0026nbsp;(2,390 - 56,400)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e55,600\u0026nbsp;\u003cbr\u003e\u0026nbsp;(10,000 - 176,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAll Food\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e78,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(59,300 - 104,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 138px;\"\u003e\n \u003cp\u003e404,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(301,000 - 566,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e237,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(190,000 - 304,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 134px;\"\u003e\n \u003cp\u003e721,000\u0026nbsp;\u003cbr\u003e\u0026nbsp;(574,000 - 946,000)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"5\" valign=\"top\" style=\"width: 660px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u003csup\u003e\u0026dagger;\u003c/sup\u003e\u0026nbsp;\u003c/strong\u003eTotals reflect the median of model simulations, so may not equal the sum of individual columns or rows, and uncertainty intervals (UIs) of totals will typically be narrower than the sum of the component UIs. Cost for \u0026lsquo;All Food\u0026rsquo; is 100% of cost of foodborne disease for the six pathogens and differs to sums across food commodities as sums of attribution proportions across food commodities do not sum to exactly 100% for all pathogens.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Food Standards Australia New Zealand","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"disease burden, pathogen-specific costs of foodborne illness, attribution of foodborne illnesses to food commodities, structured expert judgement, IDEA, Classical Model","lastPublishedDoi":"10.21203/rs.3.rs-8110329/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8110329/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eFoodborne disease and its sequelae exert a significant cost to Australia through healthcare costs, lost productivity, and occasional fatal illness. While estimating the cost of illness for all foodborne pathogens or for specific pathogens has value in quantifying this disease burden, it is also informative to estimate costs by food commodity and to identify priority areas for improving food safety. We combined a cost of illness model for foodborne illness in Australia with an expert elicitation of the food commodities associated with illness for key pathogens. The total cost of the six modelled pathogens was 721\u0026nbsp;million (June 2023 AUD), with campylobacteriosis having the greatest overall cost (AUD 420\u0026nbsp;million). Across food categories, AUD 328\u0026nbsp;million was attributed to poultry, AUD 107\u0026nbsp;million to vegetables, while dairy, beef, and pork each had costs over AUD 55\u0026nbsp;million. Strong associations were found between \u003cem\u003eCampylobacter\u003c/em\u003e and poultry (69% of campylobacteriosis cases attributed to poultry) and \u003cem\u003eYersinia\u003c/em\u003e and pork (54% of yersiniosis cases attributed to pork). This study highlights poultry as a key cause of foodborne illness in Australia, responsible for almost half of the total costs due to \u003cem\u003eCampylobacter\u003c/em\u003e, non-typhoidal \u003cem\u003eSalmonella\u003c/em\u003e, \u003cem\u003eYersinia enterocolitica\u003c/em\u003e, \u003cem\u003eListeria monocytogenes\u003c/em\u003e, and STEC.\u003c/p\u003e","manuscriptTitle":"Attribution of pathogen-specific costs of foodborne illness to food commodity groups - combining a costing model with expert judgement","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-17 08:44:32","doi":"10.21203/rs.3.rs-8110329/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"703c843b-8485-4240-a756-f25fb5751064","owner":[],"postedDate":"November 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":57968550,"name":"Health Economics \u0026 Outcomes Research"},{"id":57968551,"name":"Epidemiology"}],"tags":[],"updatedAt":"2025-11-17T08:44:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-17 08:44:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8110329","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8110329","identity":"rs-8110329","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}