A perspective on data sharing in digital food safety systems.

In this age of data, digital tools are widely promoted as having tremendous potential for enhancing food safety. However, the potential of these digital tools depends on the availability and quality of data, and a number of obstacles need to be overcome to achieve the goal of digitally enabled "smarter food safety" approaches. One key obstacle is that participants in the food system and in food safety often lack the willingness to share data, due to fears of data abuse, bad publicity, liability, and the need to keep certain data (e.g., human illness data) confidential. As these multifaceted concerns lead to tension between data utility and privacy, the solutions to these challenges need to be multifaceted. This review outlines the data needs in digital food safety systems, exemplified in different data categories and model types, and key concerns associated with sharing of food safety data, including confidentiality and privacy of shared data. To address the data privacy issue a combination of innovative strategies to protect privacy as well as legal protection against data abuse need to be pursued. Existing solutions for maximizing data utility, while not compromising data privacy, are discussed, most notably differential privacy and federated learning.

Developing an Agent-Based Model that Predicts Listeria spp. Transmission to Assess Listeria Control Strategies in Retail Stores.

Contamination of fresh produce with Listeria monocytogenes can occur throughout the supply chain, including at retail, where Listeria spp., including L. monocytogenes, may be introduced and spread via various routes. However, limited tools are available for retailers to assess practices that can enhance control of Listeria transmission to fresh produce. Therefore, we developed an agent-based model that can simulate Listeria transmission in retail produce sections to optimize environmental sampling programs and evaluate control strategies. A single retail store was used as a model environment, in which various routes of Listeria introduction into and transmission between environmental surfaces were modeled. Model prediction (i.e., Listeria prevalence) was validated using a published longitudinal study for all surfaces that were included in both the model and the validation data. Sensitivity analysis using the Partial Rank Correlation Coefficient showed that (i) initial Listeria concentration from incoming produce, (ii) transfer coefficient from produce to employee's hands, and (iii) transfer coefficient from consumer to produce were the top three parameters that were significantly (p < 0.0018) associated with the mean Listeria prevalence across all agents, suggesting that the accuracy of these parameters are important for prediction of overall Listeria prevalence at retail. Cluster analysis grouped agents with similar contamination patterns into six unique clusters; this information can be used to optimize the sampling plans for retail environments. Scenario analysis suggested that (i) more stringent supplier control as well as (ii) practices reducing Listeria transmission via consumer's hands may have the largest impact on reducing finished product contamination. Overall, we show that an agent-based model can serve as a foundational tool to help with decision-making on Listeria control strategies at retail.

Risk Assessment Predicts Most of the Salmonellosis Risk in Raw Chicken Parts is Concentrated in Those Few Products with High Levels of High-Virulence Serotypes of Salmonella.

Salmonella prevalence declined in U.S. raw poultry products since adopting prevalence-based Salmonella performance standards, but human illnesses did not reduce proportionally. We used Quantitative Microbial Risk Assessment (QMRA) to evaluate public health risks of raw chicken parts contaminated with different levels of all Salmonella and specific high- and low-virulence serotypes. Lognormal Salmonella level distributions were fitted to 2012 USDA-FSIS Baseline parts survey and 2023 USDA-FSIS HACCP verification sampling data. Three different Dose-Response (DR) approaches included (i) a single DR for all serotypes, (ii) DR that reduces Salmonella Kentucky ST152 virulence, and (iii) multiple serotype-specific DR models. All scenarios found risk concentrated in the few products with high Salmonella levels. Using a single DR model with Baseline data (µ = -3.19, σ = 1.29 Log CFU/g), 68% and 37% of illnesses were attributed to the 0.7% and 0.06% of products with >1 and >10 CFU/g Salmonella, respectively. Using distributions from 2023 HACCP data (µ = -5.53, σ = 2.45), 99.8% and 99.0% of illnesses were attributed to the 1.3% and 0.4% of products with >1 and >10 CFU/g Salmonella, respectively. Scenarios with serotype-specific DR models showed more concentrated risk at higher levels. Baseline data showed 92% and 67% and HACCP data showed >99.99% and 99.96% of illnesses attributed to products with >1 and >10 CFU/g Salmonella, respectively. Regarding serotypes using Baseline or HACCP input data, 0.002% and 0.1% of illnesses were attributed to the 0.2% and 0.4% of products with >1 CFU/g of Kentucky ST152, respectively, while 69% and 83% of illnesses were attributed to the 0.3% and 0.6% of products with >1 CFU/g of Enteritidis, Infantis, or Typhimurium, respectively. Therefore, public health risk in chicken parts is concentrated in finished products with high levels and specifically high levels of high-virulence serotypes. Low-virulence serotypes like Kentucky contribute few human cases.

Examining Patterns of Persistent Listeria Contamination in Packinghouses Using Agent-Based Models.

ABSTRACT: Persistent Listeria monocytogenes contamination may occur in a packinghouse if the pathogen successfully infiltrates the facility and reaches a harborage site, where it may be difficult to remove and may contaminate produce within the facility. There is a need for simulation-based decision support tools that can predict which equipment sites are more likely to undergo persistent contamination and simulate potential corrective actions to prevent this contamination. Thus, we adapted for longer term simulation two existing applications of an agent-based model of Listeria spp. hourly contamination dynamics in produce packinghouses. Next, we developed a novel approach to identify and analyze persistent and transient Listeria contamination patterns on simulated agents representing equipment sites and employees. Testing of corrective actions showed that methods that involved targeted, facility-specific, risk-based sanitation were the most effective in reducing both the likelihood and duration of persistent contamination. Generic approaches to controlling Listeria (e.g., more concentrated sanitizers) are unlikely to be successful and suggest that use of sanitation schedules produced through facility-specific root cause analysis and hygienic design are key in reducing persistence. Hourly Listeria contamination patterns also suggest that transient contamination may be mistaken for persistent contamination, depending on the frequency of environmental sampling. Likewise, as concentrations of Listeria on most contaminated agents were predicted to be very low, there is also a possibility to mistake persistence for transient contamination of sites, or even miss it outright, due to false-negative environmental Listeria monitoring results. These findings support that agent-based models may be valuable decision support tools, aiding in the identification of contamination patterns within packinghouses and assessing the viability of specific corrective actions.

Using agent-based modeling to compare corrective actions for Listeria contamination in produce packinghouses.

The complex environment of a produce packinghouse can facilitate the spread of pathogens such as Listeria monocytogenes in unexpected ways. This can lead to finished product contamination and potential foodborne disease cases. There is a need for simulation-based decision support tools that can test different corrective actions and are able to account for a facility's interior cross-contamination dynamics. Thus, we developed agent-based models of Listeria contamination dynamics for two produce packinghouse facilities; agents in the models represented equipment surfaces and employees, and models were parameterized using observations, values from published literature and expert opinion. Once validated with historical data from Listeria environmental sampling, each model's baseline conditions were investigated and used to determine the effectiveness of corrective actions in reducing prevalence of agents contaminated with Listeria and concentration of Listeria on contaminated agents. Evaluated corrective actions included reducing incoming Listeria, modifying cleaning and sanitation strategies, and reducing transmission pathways, and combinations thereof. Analysis of Listeria contamination predictions revealed differences between the facilities despite their functional similarities, highlighting that one-size-fits-all approaches may not always be the most effective means for selection of corrective actions in fresh produce packinghouses. Corrective actions targeting Listeria introduced in the facility on raw materials, implementing risk-based cleaning and sanitation, and modifying equipment connectivity were shown to be most effective in reducing Listeria contamination prevalence. Overall, our results suggest that a well-designed cleaning and sanitation schedule, coupled with good manufacturing practices can be effective in controlling contamination, even if incoming Listeria spp. on raw materials cannot be reduced. The presence of water within specific areas was also shown to influence corrective action performance. Our findings support that agent-based models can serve as effective decision support tools in identifying Listeria-specific vulnerabilities within individual packinghouses and hence may help reduce risks of food contamination and potential human exposure.