Balancing food safety and sustainability: trade-off risk assessments and predictive modeling

Abstract

The importance of food safety to public health is reflected in its inclusion in the United Nations Sustainable Development Goals (SDGs)—SDG 2 (Zero Hunger), SDG 3 (Good Health and Well-being), and SDG 12 (Responsible Consumption and Production)—and the World Health Organization’s food safety strategy. Its inclusion across multiple areas underscores how food safety is not an isolated objective but is closely tied to broader public health and sustainability goals. While the public often expects food to be “absolutely” safe, experts recognize that all foods carry a residual risk of causing foodborne illness and that zero risk is neither achievable nor desirable. Advances in diagnostics and surveillance systems (e.g., increases in test sensitivity and specificity) have increased the frequency of hazard detection in foods, including detection of hazards at levels that may pose minimal public health risks. However, efforts to manage these negligible risks can divert attention from more significant threats and may introduce unintended consequences that outweigh the intended benefits. To address this, holistic approaches and trade-off risk assessments are needed, accounting for the interrelationship between the health of humans, animals, and the environment (i.e., One Health) and evaluating both the costs and benefits of food safety measures, including direct expenses, externalities, social or legal constraints, and consumer preferences. Key tools enabling these risk assessments include Monte Carlo simulations and other modeling tools that are also being adopted for food safety applications, such as geographic information system models, agent-based models, and artificial intelligence (AI)-based predictive tools. These efforts can help define quantitative food safety goals that ensure appropriate, but not absolute, safety, so long as implemented controls are validated and verified. Technological advances, such as AI-enabled risk negotiation, offer new opportunities to integrate trade-offs in risk analysis and support more balanced, effective food safety strategies.

Key points

Introduction

Food safety is a key determinant of public health (1). According to the most current estimates by the World Health Organization (WHO), microbial foodborne pathogens cause 600 million cases of illness annually (2). These illnesses result in 420,000 deaths (2)—a third (125,000 deaths/year) in children aged under 5 years (3)—and a loss of 33 million disability-adjusted life years (DALYs) (2). In 2019, the World Bank estimated that foodborne diseases cost US$110 billion/year in productivity loss and medical costs in low- and middle-income countries alone (4). The inclusion of food safety within the United Nations (UN) Sustainable Development Goals (SDGs) (5), which are a part of the 2030 Agenda for Sustainable Development, underscores not only its global importance to public health but also its close interconnection with sustainability. Accordingly, the WHO Global Strategy for Food Safety 2022–2030 aims to reduce the burden of foodborne illness (1), and the WHO is currently working with the World Bank to provide updated disease burden estimates (2).

Various food production and processing strategies, such as good agricultural practices, drying, pasteurization, sanitary equipment design, refrigeration and freezing throughout distribution, and strict microbiological limits for foods, can help reduce the presence of microbial foodborne pathogens and improve food safety. However, many of these approaches come with unintended consequences that can affect and limit their application, ranging from increased food insecurity to a variety of negative environmental impacts, including increased food waste. In addition, some technologies carry high costs (e.g., for equipment, power, and labor) or logistical challenges (e.g., access to skilled labor or necessary maintenance) that hamper their implementation in many parts of the world. Hence, for some food safety approaches, mandated use or overly stringent requirements (e.g., with regard to microbiological testing) could negatively impact food availability and sustainability.

The essentially ubiquitous trade-offs associated with strategies to improve microbial food safety represent an often underappreciated and unaddressed challenge. More specifically, while some groups and individuals may desire food that is “completely” safe, absolute safety, even regarding a specific challenge such as microbial food safety, is neither possible nor desirable (6). This problem has been acknowledged since at least 1993, when the Organisation for Economic Co-operation and Development defined safe food as “a reasonable certainty that no harm will result from intended uses under the anticipated conditions of consumption” (7). Consequently, food safety decisions at all levels (governments, companies, and individuals) must weigh various trade-offs to determine what constitutes “sufficiently safe” food or what food can be considered to show “reasonable certainty that no harm will result from intended use” (7). This approach of defining “sufficiently safe” food is also exemplified by concepts such as the “Appropriate Level of Protection” (ALOP) (8, 9), which involves defining acceptable levels of risk; this approach has been applied to set microbiological benchmarks for food safety (e.g., log reduction targets for foodborne pathogens). Another approach for defining acceptable levels of risk, particularly for food additives, includes setting a one-in-a-million risk level as “acceptable”, with the United States Food and Drug Administration (FDA) declaring that lifetime risks lower than this one-in-a-million level should be considered de minimis under the so-called Delaney Amendment (10). However, the de minimis principle could be considered as an arbitrary threshold and thus may not be considered a satisfactory approach for distinguishing between negligible and non-negligible risk (10). Similarly, defining what is an “appropriate” level of protection can be challenging. Hence, approaches and tools for improving and facilitating food safety decisions must be further developed and applied to better reach consensus on what constitutes acceptable or tolerable risks (11), as well as negligible or non-negligible risks, for different foods across societies. Consistent with this, increasing the use of food chain information, scientific evidence, and risk assessment in making risk management decisions is a strategic pillar of the WHO Global Strategy for Food Safety (1).

This article details how global food systems are increasingly facing more, and sometimes competing, demands, such as (i) supporting public health protection by ensuring food safety while also (ii) striving for sustainability in food systems (including reduction of food waste). For example, discarding any food that tests positive for foodborne pathogens, without considering the likelihood of the food causing foodborne illness, may help safeguard public health but also contributes to food waste and associated consequences (e.g., greenhouse gas emissions). Thus, due to the “interconnectedness” of our food system (as illustrated by the One Health concept), there is a need to optimize, balance, and negotiate strategies that can improve the safety and sustainability of food systems to provide well-being and health to global populations while protecting constituents of food systems (e.g., farmers and food businesses) and the planet (e.g., animals, plants, and the environment). These challenges bring new urgency to the prioritization of risk-based decision-making—which has been previously been detailed by various agencies and groups (7, 12, 13) and will be discussed later in this manuscript—over hazard-based decision-making, which could be described as primarily focusing on pathogen detection and elimination. This focus on risk-based decision-making needs to overcome implementation barriers (e.g., challenges in defining acceptable food safety risk levels) and move beyond current approaches, including ALOP and “As Low as Reasonably Achievable” (ALARA) (14). Furthermore, explicit consideration of residual risk in policymaking will be essential to comprehensively drive implementation of risk-based food safety systems. The use of new approaches—e.g., artificial intelligence (AI) or new modeling tools—can help define acceptable levels of risk and consider residual risks while balancing competing interests in our complex food systems. While the focus of this article is on microbial food safety, many of the concepts detailed here, and particularly the need for trade-off risk assessments, also apply to chemical food safety risks, where some trade-off risk assessments have already been applied (e.g., 15).

Microbial food safety in the context of human health, One Health, and food system sustainability

Human health impacts of foodborne microbial pathogens

Pathogens commonly associated with foodborne illness cases include the bacteria Salmonella, certain Escherichia coli strains, Listeria monocytogenes, and Campylobacter (16), and viruses such as norovirus, hepatitis A virus, and hepatitis E virus (17). While many of these pathogens (e.g., norovirus) typically cause mild foodborne illnesses, characterized by vomiting and diarrhea, several pathogens can cause more severe illnesses. For example, L. monocytogenes typically causes severe invasive infections with symptoms including septicemia, encephalitis, and miscarriage; approximately 90% of listeriosis infections result in hospitalization, and 20% cause death. For other pathogens, specific strains may have a propensity to cause more serious infections. Enterohemorrhagic E. coli (EHEC), for example, are a subgroup of E. coli that can cause particularly severe symptoms, including kidney failure and hemolytic uremic syndrome (18). Other E. coli subtypes may either cause milder symptoms (e.g., diarrhea) or are not typically linked to foodborne disease at all (19). In addition, susceptibility to foodborne infections can differ substantially between population subsegments, with young, elderly, and immunocompromised individuals often carrying a higher risk.

Foodborne pathogens can also cause long-term sequelae with substantial public health impacts. For example, Campylobacter can trigger Guillain-Barré syndrome, a neurological autoimmune disorder whose symptoms include muscle weakness and sometimes paralysis; estimates of its occurrence range from 21.5–172 cases per 100,000 Campylobacter cases (20, 21). Other long-term sequelae of Campylobacter, Salmonella, Shigella, and Yersinia infections can include reactive arthritis (22–24). In addition to direct public health impacts, microbial foodborne illnesses have many other indirect impacts, including contributions to malnutrition and poor infant health (e.g., due to reduced nutrient absorption in individuals with re-occurring foodborne illness episodes) (25).

Microbial food safety and One Health

Many microbial food safety issues are closely linked to the microbial safety and quality of water and to animal health, as detailed in a number of reviews on One Health and food safety (e.g., 1, 26). For example, contaminated water can be an important source of pathogen contamination of different foods at different stages of the food chain, including primary production (e.g., if contaminated water is used for irrigation of produce), food processing, and consumer food preparation. Food animals infected with or carrying foodborne pathogens (e.g., Salmonella) not only represent an important “direct” source of foodborne pathogens (e.g., through contaminated meat, milk, or eggs) but also contribute more indirectly to transmission and dispersal. For example, runoff from animal facilities can introduce foodborne pathogens into water sources used for irrigation (27) as well as directly into pre-harvest environments, such as produce fields. Although less well documented, airborne dispersal of foodborne pathogens from livestock operations may also contribute to contamination of pre-harvest environments (28, 29). Wildlife, too, has been found to lead to the dispersal of foodborne pathogens through contamination of water or direct fecal deposition into pre-harvest environments (including in foodborne disease outbreaks linked to contaminated produce) (30, 31). As microbial food safety can be considered a One Health issue (1, 26), strategies designed to reduce food safety risks are likely to have wide-ranging impacts—both positive and negative (see Figure 1 and Table 1 for examples of food safety interventions and associated trade-offs).

Figure 1

Challenges for food system sustainability

Food safety is integral to UN SDG 2 (Zero Hunger), which aims to end hunger and achieve food security. This goal links to two further goals, SDG 3 (Good Health and Well-being) and SDG 12 (Responsible Consumption and Production), due to the need for safe and nutritious foods and the reduction of foodborne illness rates as well as the need for sustainable consumption and production patterns, which requires maintaining food safety while reducing food waste (5).

Major challenges for sustainable food systems include (i) ensuring access to safe, abundant, and nutritious food, (ii) minimizing damage to natural ecosystems, and (iii) assuring quality of life and prosperity for individuals and communities associated with food production, processing, and distribution (32). These major challenges must be addressed in the context of microbial food safety. Reducing public health impacts associated with microbial food safety is a significant challenge in its own, often classified as a “wicked problem” (33, 34), i.e., one that is challenging to address because of complex, contradictory, and changing requirements that are often difficult to recognize. “Wicked problems” have solutions that “are not true-or-false but better or worse” for which “the problem is never solved definitively” (33).

Some key challenges associated with ensuring food safety include (i) diversity of the food supply (ranging from canned products to fresh meat, seafood, and produce), (ii) diversity of food production around the globe (from highly sophisticated production facilities to informal, and often unregulated, food establishments), (iii) foodborne pathogen contamination that can occur throughout the food value chain, including in agricultural environments, during harvesting, or in processing plants, retail establishments, restaurants, homes, and during various transportation steps, and (iv) the vast magnitude of food that must be produced. To illustrate the last point, in 2022, approximately 26.7 billion chickens were raised globally (35), and the United States consumed approximately 6.2 billion pounds of tomatoes (36). It is also worth noting that food production must continue to increase as the human population grows; it is estimated that the global food demand will increase by over 35% during the first half of the 21st century (i.e., from 2005, 2007, or 2010 through to 2050) (37, 38), exacerbating the challenges in building and maintaining sustainable and safe food systems.

Hazard- and risk-based approaches to microbial food safety

In the literature, it is not uncommon to see the mention of hazard- and risk-based approaches to food safety. Hazard-based approaches have been defined as approaches where the detection of a pathogen (typically a pathogenic species), regardless of level or other factors that impact risk, is used as a basis for legislation and/or risk management action (52). Examples for this may be regulations that define any ready-to-eat (RTE) food that tests positive for L. monocytogenes as adulterated, regardless of levels or whether it would support growth of Listeria. Risk-based approaches, on the other hand, assess the probability of an adverse effect on an organism, including humans, given a certain exposure and scale risk mitigation strategies accordingly. The value of risk-based approaches to food safety is generally well recognized and broadly supported. For example, the WHO 2022–2030 Global Strategy for Food Safety recommends that “When setting and implementing regulatory requirements, the national food control systems should consider the whole food chain and take a risk-based approach” (1). However, a delineation between hazard- and risk-based approaches to microbial food safety is not straightforward. Since zero risk is unattainable, i.e., some residual risk always remains despite food safety interventions (see Figure 2 for an example illustrating residual risk), even so-called hazard-based approaches typically involve decisions about acceptable risk, often defined indirectly through sensitivity of detection methods, sample sizes, and sampling frequency. For instance, a food safety system requiring RTE food to test negative for L. monocytogenes in a single 25 g sample accepts a higher level of risk than one requiring five 25 g samples (125 g total). In practice, food safety systems and approaches are thus probably better viewed as “hazard-focused” or “risk-focused”, as nearly all systems incorporate some degree of risk-based decision-making.

Figure 2

Development and implementation of risk-focused food safety approaches often utilize quantitative microbial risk assessment (QMRA) (as detailed further below). However, other and simpler alternative approaches (e.g., risk profiling) can also be utilized (13); such alternative approaches to a full QMRA can be particularly important, as a frequently cited global challenge in implementing risk-based approaches is the availability of sufficient data and the capacity to collect and analyze it. However, some studies have also illustrated that meaningful QMRAs can be conducted with limited data, such as a recent QMRA of food safety interventions for Campylobacter spp. and Salmonella spp. along the chicken meat supply chain in Burkina Faso and Ethiopia (53).

Definition of an acceptable risk (or an acceptable risk reduction), particularly for risk-focused food safety approaches, is important for establishing action levels [e.g., <x colony-forming unit (CFU) of a given pathogen in a certain food product]. However, as the term “acceptable risk” implies that some level of foodborne illness is acceptable, other terms such as “tolerable risk” or “achievable risk” may be preferable, as outlined in a report on risk assessments for waterborne diseases (54). This report (54) also describes previous efforts to define acceptable microbial risks for water and thus may provide valuable learnings for food safety. Regarding food safety, previous attempts to help define acceptable risks included efforts to define ALOPs for different food safety hazards (e.g., L. monocytogenes), which could then be used to define specific food safety objectives (8, 9). However, the concept of ALOP and the associated nomenclature and definitions appear to have seen only limited global uptake since the World Trade Organization first described this concept in 1995 (55). To address the issue of defining limits for presence and levels of foodborne pathogens, the concept of ALARA is sometimes used (14); this approach, however, tends to focus more on technologically attainable risk reduction rather than societally acceptable risk levels and thus may be less valuable in the further development of risk-based food safety systems that consider societal costs, benefits, and trade-offs.

The choice of hazard- versus risk-focused food safety approaches can have important implications beyond the determination of regulatory limits for foodborne pathogens. For example, hazard-focused systems often emphasize end-product testing as a means of ensuring safety despite well-established evidence that it is largely ineffective at detecting contamination, especially when present at a low prevalence (56), though it remains useful as a verification tool. Risk-focused approaches, on the other hand, facilitate the setting of performance criteria (PC), which specify goals for the effectiveness of food safety interventions, ideally throughout production and processing (57, 58). For example, risk-focused approaches may determine that a 5-log reduction in a target organism (e.g., the most heat-resistant vegetative pathogen in raw milk) may be sufficient for one product, whereas a 4-log reduction may suffice for another (e.g., Salmonella in almonds) (59, 60). While decisions on appropriate PC rely on assumptions about pathogen distribution and initial contamination concentration in raw materials, which should be verified (see Figure 3), decisions on specific PC (and whether they are “fit for purpose”) should also take into account an acceptable risk (e.g., x human illness cases per year), a consideration used when proposing that a 4-log reduction may be appropriate for Salmonella in almonds.

Figure 3

In contrast, hazard-focused approaches may aim to eliminate pathogens in finished products and even in areas of food processing facilities that pose minimal risk to the product (this approach could manifest by regulators not accounting for sampling areas when reviewing industry environmental sampling data). This can misdirect resources: as a somewhat extreme example, food business operators may allocate similar resources to address pathogen contamination in a zone 4 site (far from food contact surfaces) as to contamination in a zone 2 site (near exposed food), despite the much higher risk posed by the latter (61). Similarly, a hazard-focused approach may give equal resources to control L. monocytogenes in products that do and do not support the growth of the organism, with the former representing a higher public health risk. These examples illustrate how a hazard-focused approach that demands equal attention be given to all hazards ignores the concept of opportunity cost, which Frank and Bernanke defined for an activity as “the value of the next best alternative that must be forgone to undertake the activity” (62). In our specific scenario, the opportunity cost is the cost of improving pathogen control in zone 4 by redirecting resources to control the pathogen in zone 2, resulting in greater benefits for public health and food safety. Although opportunity costs are rarely formally analyzed in food safety, they likely influence food business decisions (63). Consequently, a better understanding of these trade-offs could lead to more effective system design. Additionally, hazard-focused systems targeting zero detection may create other unintended consequences, such as less rigorous environmental sampling (e.g., sampling mostly clean surfaces), ultimately weakening validation and verification efforts.

Risk metrics

While risk-focused food safety approaches are frequently applied in public health decision-making, a somewhat “hidden” aspect of risk assessments is the variety of different metrics available to assess risk. These include different numerators, such as the number of foodborne illness cases, deaths, or quality-adjusted life years (QALYs), and different denominators, such as per serving, or per population per year (e.g., per million people per year) (see Figure 4 and Table 2 for further details on risk metrics, including QALYs). Crucially, the choice of risk metric may impact management decisions. For example, assessments using the number of foodborne illness cases in the United States may prioritize managing norovirus (estimated 5,461,731 cases and 149 deaths annually) (16) over L. monocytogenes (estimated 1,591 cases and 255 deaths annually), whereas assessments based on deaths would suggest the opposite; DALYs and QALYs represent one approach for unified risk metrics that take different dimensions of illness into account. Similarly, the choice of denominator can shape conclusions on food safety priorities. For example, based on “risk per serving”, raw milk has a substantially higher risk of causing foodborne illness than pasteurized milk. However, pasteurized milk represents a larger “risk per population” (as compared to raw milk) due to its far greater consumption (6), suggesting that risk reduction efforts focusing on pasteurized products are justified even if raw milk represents a larger per serving risk. A published QMRA of Salmonella in almonds illustrates how risk metrics can guide food safety decisions. This analysis supported the use of a heat treatment process giving a 4-log reduction in this pathogen, which was estimated to result in fewer than one case of salmonellosis per year in the United States (59).

Figure 4

Importantly, risk-based decision-making at the food business level must often account for both public health risks and food business-level financial risks—sometimes out of necessity due to legal requirements such as the Sarbanes-Oxley Act in the United States (https://sarbanes-oxley-act.com/). Examples of food business-level risk metrics may include (i) recall risk and annualized recall cost, which factors in both the likelihood and financial impact of recalls, (ii) outbreak risk and associated annualized outbreak costs, considering the likelihood and financial impact of outbreaks, and (iii) costs of contamination events that do not lead to outbreaks or recalls, such as expenses related to product disposal, rework, or downgrading. The financial impacts of recalls and outbreaks should also account for the impacts of reputational damage.

As with public health metrics, different denominators may be used for business risk metrics, such as cost per unit produced or per facility. Since companies must prioritize addressing various types of risk, such as food safety, cybersecurity, or workplace safety, metrics that quantify the costs of different food safety risk management options can support more rational, data-driven decision-making. This applies both at the food business level and more broadly, for example when trade groups propose or mandate specific food safety practices. Financial risk metrics can also inform policy and regulatory decisions by identifying the most cost-effective risk management strategies. For instance, Ssemanda et al. (76) recently evaluated the cost–effectiveness of interventions to control Campylobacter and Salmonella in the chicken meat supply chain in Burkina Faso and Ethiopia, using intervention costs and avoided DALYs as key metrics (76). Similarly, another study evaluated the cost–benefit ratio of different Campylobacter interventions in poultry supply chains in Belgium (77).

Regulatory agency use of hazard- and risk-focused approaches

Regulatory agencies worldwide generally employ a combination of hazard- and risk-focused approaches to ensure food safety. Hazard-focused approaches are often favored for pathogens that cause severe illness, such as L. monocytogenes, and typically involve regulations that deem any detection of a pathogen in a finished product as evidence that a product is “unsafe” and, hence, adulterated. The current FDA policy on L. monocytogenes in RTE foods is widely regarded as an example of a hazard-focused approach (11). Under this policy, any FDA-regulated RTE product that tests positive for L. monocytogenes, regardless of concentration or whether the food supports pathogen growth, is considered adulterated. In contrast, some countries take a more risk-based approach to regulating L. monocytogenes in RTE foods. These approaches include setting different microbial limits depending on the food’s characteristics. For example, Canada, Chile, and others have adopted a threshold of 100 CFU/g for foods that do not support L. monocytogenes growth, consistent with the Codex Alimentarius Commission guideline (78).

Regulatory agencies may also use hazard- or risk-focused approaches beyond standard-setting, such as for prioritizing inspections and other regulatory actions (11). Importantly, this allows regulatory agencies that operate under a hazard-focused standard (e.g., absence of L. monocytogenes in a 25 g RTE product sample) to implement a more risk-based system by focusing enforcement and testing on high-risk foods, such as RTE foods that support L. monocytogenes growth and have been linked to outbreaks. The United States Department of Agriculture (USDA) “Listeria Rule” (79, 80) provides an example of this approach. Under this rule, RTE deli meats, which have been implicated in several listeriosis outbreaks (81, 82), are classified into control alternatives based on (i) the capacity to control L. monocytogenes contamination and (ii) the product’s ability to limit pathogen growth in the finished product. Alternative 1 products use both a post-lethality treatment to reduce or eliminate L. monocytogenes and an antimicrobial agent or process to inhibit its growth and are consequently considered to represent the lowest public health risk. In contrast, Alternative 3 products rely solely on sanitation practices to control contamination in the environment and product and are considered to represent the highest risk. Although all products are subject to the same legal standard for L. monocytogenes detection (i.e., absence in a 25 g sample is typically interpreted as “zero tolerance”) (11), the USDA uses a risk-ranking algorithm that takes into account the control alternative, as well as production volume, type of product produced, and sampling history, to identify facilities for finished product testing (80). In addition, the USDA prescribes risk-based sampling frequencies for environmental Listeria sampling on food contact surfaces. For example, a 2022 Food Safety and Inspection Service directive indicates that there are no regulatory testing requirements for Alternative 1 products, while testing of food contact surfaces in the post-lethality environment, for L. monocytogenes or an indicator organism, is required for Alternative 3 establishments (83). These regulatory and enforcement strategies are supported by risk assessments that support significant public health benefits from reformulating RTE deli meats with growth inhibitors (84, 85).

Industry use of hazard- and risk-focused approaches

While regulatory agencies and international bodies, such as the Food and Agriculture Organization of the United Nations (FAO), frequently use QMRAs to guide food safety regulations and enforcement, food business operators (and particularly smaller food businesses) may often have limited access to such tools and resources. Food business operators with limited risk assessment expertise may also be prone to utilizing hazard-focused approaches and general food safety plans outlined in guidance documents (86). Consequently, they may use simple action limits, such as the absence of a hazard in samples assessed by end-product testing (with these limits often prescribed by regulatory agencies or buyers, e.g., retailers). Food businesses may also rely on qualitative (rather than quantitative) risk assessment approaches to manage food safety risk and develop and implement food safety systems. These assessments often lack standardized methods and may not formally support food safety decisions related to preventive controls or prerequisite programs. This is particularly evident in areas such as control of environmental hazards (e.g., L. monocytogenes), where decisions about sanitation frequency, maintenance protocols, and environmental monitoring are rarely grounded in formal qualitative or quantitative risk assessments and instead rely more heavily on classification of zones within food production facilities (e.g., as low or high hygiene and food contact or non-food contact surfaces) to define food safety measures including sampling frequency (87, 88).

While many companies implement fairly effective food safety systems using current practices, there are substantial opportunities to improve decision-making. Specifically, there is a need for (i) better and easier-to-use modeling and risk assessment tools suitable for industry and (ii) training and support for development and application of risk assessments; this is particularly important as regulators increasingly expect food safety systems to be “science- and risk-based”. Building science-based systems will, however, require consensus on clear goals regarding foodborne illness risk, defined using either (i) absolute targets (e.g., a maximum allowable probability of illness per serving of a product), (ii) relative targets (e.g., a specific percentage reduction in illness risk for a product or pathogen), or (iii) business risk-focused targets using financial or operational risk metrics. Scientific methods, such as risk assessments, can then be used to identify and evaluate the best strategies for achieving these goals.

Risk assessment and predictive modeling for food safety and sustainability

Modeling tools and quantitative risk assessments play a key role in developing systems that assure a level of food safety appropriate for a society or business, which by necessity must balance multiple objectives, including food safety, public health, and the sustainability of food and agricultural systems. Over the past few decades, numerous studies have developed food safety risk assessments. These assessments are valuable for setting priorities and standards, including by regulatory agencies. For example, a 2003 USDA/FDA risk-ranking report (89) concluded that RTE deli meats posed the highest risk for listeriosis, supporting regulatory and enforcement efforts targeting these products in the United States. More targeted risk assessments and models evaluate the impact of specific regulatory practices and recommended control strategies for reducing foodborne illness. Increasingly, these tools are also being used to support risk management in specific industry sectors, such as the almond industry (59), and even at the level of individual companies and processing facilities (90).

Monte Carlo simulation-based quantitative microbial risk assessments

Monte Carlo simulation-based quantitative microbial risk assessments (MC-QMRAs) were first applied to microbial food safety questions in the early 1990s (91), although key components, such as dose–response modeling, trace back to the 1920s (92). Key applications include (i) risk ranking, (ii) evaluating the predicted impact of regulations, interventions, and control strategies, and (iii) identifying knowledge gaps to prioritize future research and resource allocation.

An MC-QMRA (see Figure 5 for details) consists of two components: (i) a QMRA, which typically estimates the likelihood of adverse health outcomes from exposure to a hazard (e.g., a foodborne pathogen) and (ii) a Monte Carlo simulation, which quantifies variability and uncertainty in the model outcomes. A typical QMRA includes four stages: hazard identification, exposure assessment, dose–response characterization, and risk characterization (Figure 5) (93). Although primarily used to assess public health risks, the stages of a QMRA can be adapted to assess other types of risk, such as business risk.

Figure 5

Within QMRAs, food value chain stages are typically represented using quantitative values or distributions that describe pathogen prevalence/concentration and/or process parameters (e.g., log reduction from heat treatment). Model inputs reflect either variability (natural heterogeneity within a system) and/or uncertainty (lack of knowledge or data) (94). For example, while both Lambertini et al. (75) and Farakos et al. (95) used distributions to represent the variability of Salmonella spp. concentration on pistachios, Farakos et al. (95) also represented uncertainty using this distribution by sampling its parameters (i.e., mean and standard deviation) using a Monte Carlo Markov chain.

Monte Carlo simulations work by iteratively sampling from the defined distributions to generate a range of possible outcomes. A single iteration yields one estimate of a risk metric (e.g., illness risk per 100,000 servings), while many iterations (e.g., 100,000) generate a distribution of the risk metric. This allows estimation of the mean, range, and confidence intervals of risk that can be compared to real-world data on foodborne illness prevalence (if such data are available for a given food–pathogen combination) to gauge the accuracy of the risk estimates. In a second-order MC-QMRA, variability and uncertainty are modeled separately, which helps identify both the likely magnitude of risk and the areas needing improved data (94).

In addition to their use for risk rankings, MC-QMRAs are commonly used to evaluate potential regulatory interventions. Examples include (i) modeling of Salmonella transmission through eggs, which informed United States regulations mandating refrigeration (96, 97), (ii) a joint FAO/WHO QMRA for L. monocytogenes (98), which found that public health risk did not differ substantially between acceptance limits of 0.04 and 100 CFU/g at the point of consumption, assuming certain contamination patterns, and (iii) a recent USDA QMRA that evaluated the impact of controlling Salmonella levels and targeting high-virulence serovars in raw chicken products (99).

Some MC-QMRAs were also specifically developed to support decision-making by industry, which often must manage microbial risks more stringently than required by regulation. For example, Zoellner et al. (90) modeled the risk of listeriosis from frozen vegetables, showing low median risk but high potential risk in worst-case scenarios involving improper cooking by vulnerable consumers—underscoring the importance of considering consumer behavior. Other MC-QMRAs have introduced alternative risk metrics, such as the probability of a positive product test (100, 101). Although published MC-QMRAs have yet to fully incorporate recall or outbreak costs as primary risk metrics, such metrics could provide valuable insights into business-level risks.

Other modeling tools used for microbial risk assessment and risk management

In addition to Monte Carlo simulations, other modeling tools are also being adopted for food safety applications, including geographic information system (GIS) models, agent-based models (ABMs), and AI-based predictive tools. Mechanistic models grounded in molecular biology are also emerging to help predict microbial inactivation (102). Following external validation to assess predictive accuracy, the validated models can provide support for decision-making across the stages of the food value chain, including in complex and heterogenous environments (e.g., in agricultural fields with GIS models); these models also offer opportunities for conducting stakeholder-specific trade-off risk assessments.

GIS-based models

GIS-based models have diverse applications in food safety and can predict pathogen contamination risks based on spatial and environmental data. These models typically involve mapping locations of pathogen-positive sites and interpolating these data across geographical regions to identify high-risk areas and environmental drivers for pathogen contamination. Considered risk factors often vary spatially (e.g., proximity to pastures, roads, or water sources) and temporally (e.g., rainfall or temperature patterns). For example, some studies (103–105) used GIS models to predict pathogen contamination of soil or produce at the field level, accounting for variables such as proximity to land-use classes (e.g., pastures), soil properties, and weather. One study found that the likelihood of isolating L. monocytogenes from fields in the state of New York was influenced by the field’s proximity to certain spatial features, such as roads and water (105).

GIS models can be used not only to determine food safety risk factors and help identify possible interventions to reduce risks but also to understand trade-offs involved in such interventions. For example, Karp et al. (106) reported that the prevalence of enterohemorrhagic E. coli in agricultural fields was associated with both proximity to grazing lands and removal of non-crop vegetation around fields. These findings illustrate that GIS models can be applied to conduct trade-off assessments, informing the implementation of various food safety-related measures relevant to produce, including (i) determining buffer zones between fields and grazing lands, which can involve balancing the loss in arable land with reductions in food safety risks and (ii) determining the impact of non-crop vegetation removal on both food safety risks and ecological disturbances.

Agent-based models

ABMs can be used to characterize the dynamics of pathogens in complex environments by considering events that can directly or indirectly influence pathogen introduction, persistence, die-off, or removal. While ABMs have been used to understand pathogen transmission in agricultural environments, for example, in grazing cattle (107), they have shown particular promise for assessing transmission and control of index organisms (e.g., Listeria spp.) and foodborne pathogens (e.g., L. monocytogenes) in processing facilities (108) and retail environments (109). For example, ABMs have been applied to simulate Listeria transmission in processing environments, revealing that pathogen persistence is higher in zone 4 areas such as loading docks—likely due to damaged surfaces and less frequent sanitation (110). ABMs have also been used to test intervention strategies. For example, Barnett-Neefs et al. (111) reported that risk-based sanitation, which is informed by facility-specific root cause analysis, can be effective for controlling Listeria in produce packing houses.

ABMs can also support decision-making by estimating costs and different trade-offs. For instance, increasing wet cleaning frequency may enhance sanitation but also increase the risk of pathogen spread due to water dispersion (112). ABMs are uniquely equipped to not only consider (i) how wet cleaning can impact pathogen dynamics, including die-off (due to exposure to sanitizers and disinfectants) and growth and/or dispersion (due to the presence of moisture) but also (ii) facility-specific challenges, such as evaluating pathogen dynamics in areas within a facility that are susceptible to pooling water.

AI-based models and predictive tools

AI tools, including machine learning models, are also increasingly being developed for food safety applications (113, 114), including risk assessments. For example, machine learning models have been developed to predict the presence of pathogens in agricultural water, utilizing various inputs such as weather, water quality, and microbial characteristics (115, 116). AI models also have the potential to enhance QMRAs and food safety decision-making more broadly. For example, machine learning models may (i) outperform traditional primary growth models by predicting novel growth phenotypes of bacteria (117), (ii) improve specificity of dose–response models by incorporating genomic data (118), and (iii) support scenario development for QMRAs by identifying emerging hazards, risks, and intervention strategies using natural language processing (119).

Neural networks have also been used to predict the growth/no-growth boundary of foodborne pathogens such as Staphylococcus aureus (120). While these tools can help refine predictions related to pathogen growth in MC-QMRAs, they can also be applied more directly to inform food safety decision-making (e.g., development of food safety plans), for example with improved decision-making on whether a given food supports pathogen growth (78).

Emerging tools for precision food safety

Advances in technologies such as genomics offer opportunities to improve risk assessments and develop precision food safety approaches—methods that are more targeted and data-driven than current practices, akin to precision medicine (121). For example, traditional food safety approaches often rely on broad taxonomic classifications (e.g., species or genus) to define hazards and regulatory targets. However, bacterial species or foodborne pathogens like L. monocytogenes, Salmonella, and Cronobacter include strains with substantial variation in virulence and other phenotypic traits. For certain pathogens for which virulence differences are well characterized—such as E. coli—regulations already target specific subtypes (i.e., E. coli O157 or broader groups of enterohemorrhagic strains, which include E. coli O157) (122). Genomic tools can further refine these approaches by identifying genetic markers—such as specific gene patterns or point mutations—that correlate with virulence differences. For instance, studies on L. monocytogenes have identified point mutations causing premature stop codons in the virulence gene inlA, resulting in reduced invasiveness in human intestinal epithelial cells and reduced virulence in a guinea pig model (123–125). In addition, at least some of the L. monocytogenes strains that carry these mutations are frequently found in food isolates but are rare (and significantly under-represented) among human isolates, further supporting their reduced ability to cause human disease (126–129).

Food safety regulations rarely incorporate strain-level virulence differences. A recent WHO/FAO report acknowledged the potential benefits of such precision but also highlighted challenges, including risk communication and the cost of subtype-specific management (130). While virulence determinants in pathogens like Salmonella are less clearly defined, existing knowledge has been applied in risk assessments to evaluate the benefits of targeting high-risk strains. For example, a risk assessment on human salmonellosis cases from raw chicken parts found that public health risk is concentrated in finished products with high Salmonella levels and specifically, high levels of high‐virulence serotypes; focusing control measures on high-virulence serotypes achieved public health benefits comparable to strategies targeting all Salmonella serotypes (131). Such targeted approaches could reduce food waste (by avoiding unnecessary rejection of low-risk products) and improve control strategies (e.g., by directing vaccination efforts or surveillance toward high-virulence serotypes). Hence, efforts should continue to focus on integration and use of genomic tools (132) in risk assessments to help with the development of more resource-efficient food safety systems, appreciating that some groups (e.g., some food business operators) may be more risk-averse than others and thus may select more conservative approaches to minimize business-related risk (e.g., due to litigation and legal action).

Trade-off risk assessments: current status and future developments

In the long term, a holistic approach to food safety (Figure 6) requires assessing both the benefits and costs of interventions—including the economic, environmental, societal, legal, and logistical considerations associated with different control strategies. Currently, most cost evaluations of food safety interventions occur after classical MC-QMRAs, which focus primarily on food safety outcomes. In some countries, regulatory impact assessments are required for high-cost regulations, but these are often qualitative and lack the methodological rigor (e.g., Monte Carlo simulations) of the initial MC-QMRA. Similarly, industry-led assessments typically consider costs only after the primary risk analysis, even though decisions on implementing food safety measures are often shaped as much by trade-offs as by the expected safety benefits.

Figure 6

Hence, there is a need for formal “trade-off risk assessments” that evaluate public health, environmental, social, and economic consequences of interventions within the risk assessment itself (see Appendix 1 for a list of terms that are relevant to trade-off risk assessments). Trade-off risk assessments represent one analytic approach that can be used to perform risk–benefit analyses (RBA), which can also be performed with more informal approaches. The term “multi-objective optimization” (a tool that can support “multiple-criteria decision-making”) represents another, related concept that has been mentioned in efforts to optimize food systems while considering multiple factors and trade-offs (133). Trade-offs are already regularly evaluated in disaster management, often through less formal approaches. For example, as part of the emergency response to Hurricane Harvey in Texas in 2017 (134), the authorities in Houston issued stay-in-place orders to avoid evacuation-related casualties, which, drawing on prior experience, were expected to outweigh the benefits of evacuations (135). Researchers in other fields have proposed pairing risk assessments with life-cycle assessments (LCAs) to enable holistic and simultaneous analyses of local and global impacts of public health interventions (136–138). In one example, a hybrid QMRA-LCA that assessed two municipal water management plans in Sydney, Australia (137) found that one plan reduced foodborne illnesses (lower DALYs in QMRA) but increased energy use and climate-related health risks (higher DALYs in LCA), illustrating the importance of using complementary tools to fully capture trade-offs. Additional publications that provide further insight into integrating LCAs and risk assessments for microbial and/or chemical hazards include Harder et al. (139) and Kobayashi et al. (140).

In food safety, trade-off risk assessments have been used for chemical hazards. For example, the FDA’s 2014 Quantitative assessment of the net effects on fetal neurodevelopment from eating commercial fish calculated the “net effects” of fish consumption by weighing the risks of methylmercury exposure against the benefits of nutrient intake during pregnancy using dual dose–response models (15). Considering the “net effects” of an action is representative of RBA, which allows trade-off decision-making that addresses whether consumption of a certain food (141) or change to a process (142) is warranted by simultaneously quantifying health-related benefits and consequences. Since RBA typically consider “localized” benefits and risks limited to the subset of a population that consumes a certain food, others have discussed the need to expand RBA to quantify regional and global health and environmental impacts (e.g., biodiversity loss) by incorporating LCA (143) or multi-criteria decision analysis (144); such approaches would also allow RBA to account for economic constraints and public preferences. Another example of a trade-off risk assessment is a study that used a risk–risk analysis approach to estimate DALYs from infant exposure to inorganic arsenic and aflatoxins in cereals, finding that shifting all infant cereal consumption to oats could substantially reduce health burdens (145). Despite these efforts, broader use of trade-off assessments is needed to quantify unintended negative impacts of food safety measures, including those affecting public health, sustainability, and society. For example, Kim et al. (131) recently assessed not only the benefits of stricter Salmonella product standards for raw poultry but also quantified associated food losses. Post-implementation studies have also suggested unintended consequences of food safety practices. For example, efforts to reduce pathogen contamination of produce through removing non-crop vegetation and replacing compost with other fertilizers were reported to possibly impact biodiversity, soil health, and carbon sequestration (45). Other examples of food safety sustainability dilemmas that could be addressed by trade-off risk assessments include balancing potential public health risks of water reuse with the environmental impacts of conserving water as well as balancing food safety risks associated with donation of left-over or expired foods with food security benefits (see Table 1 for additional examples). These scenarios underscore the value of trade-off risk assessments to quantitatively capture the full impact of food safety interventions at multiple scales including, for example, (i) individual-level health risks and benefits, (ii) local and regional sustainability impacts, and (iii) global environmental consequences (e.g., carbon emissions from recalls or product destruction). This will be particularly important as we try to address various impacts of climate change on food safety, which will include a need to balance food safety outcomes with sustainability impacts related to climate change (e.g., carbon emissions). See (146, 147) for further reading on the impacts of climate change on food systems and food safety.

Innovative approaches to socially acceptable food safety

As discussed, conventional risk assessment frameworks often fall short in addressing intersectoral trade-offs and the inherent complexity of achieving and defining appropriate levels of food safety in multi-sectoral food systems. A central challenge is the difficulty of comparing benefits and costs that are measured in different units, such as “illness cases avoided” versus “greenhouse gas emissions generated” or “implementation costs incurred”. Classical approaches attempt to harmonize these using common metrics, such as financial costs or DALYs. For example, one study assessed trade-offs between foodborne illness and the climate change impacts of two municipal water management plans, using DALYs as a shared metric (140).

However, using a single, shared metric for costs and benefits is often impractical and insufficient. In addition, risk perception by the general population also needs to be considered when trying to define acceptable food safety risks. To address this, Ehling-Schulz et al. (148) recently proposed an AI-assisted risk negotiation framework—a structured, multi-step process that includes (i) stakeholder roundtable formation, (ii) problem formulation, (iii) AI-supported risk assessment and valuation, (iv) risk negotiation, (v) communication and implementation, and (vi) outcome evaluation and risk renegotiation. AI tools, such as large language models and predictive analytics, can facilitate each step by reducing bias, enabling consensus-building, and fostering transparent, cross-sectoral engagement. This negotiation-based risk analysis has the potential to transform conventional, siloed risk assessments into participatory and more holistic processes that balance food safety and diverse trade-offs. However, this will require a fundamental shift away from a “zero-risk” paradigm toward a balanced risk negotiation model that acknowledges trade-offs between food safety, food security, sustainability, and economic feasibility. AI can assist in risk negotiation by analyzing large datasets, identifying trends, and simulating potential outcomes. Specifically, AI-powered multi-agent negotiation frameworks allow stakeholders such as policymakers, food producers, and consumer representatives to engage in structured decision-making processes that integrate a multitude of diverse perspectives.

By incorporating AI-assisted risk negotiation, food safety governance can become more adaptive, participatory, and transparent. This has the potential to improve risk communication and stakeholder trust as well as to enhance global food system resilience by fostering equitable and sustainable food safety solutions that accommodate societal needs. It is, however, important to consider that generative AI can be susceptible to fallacies including fabrication, and reasoning and mathematical errors (149). Thus, despite the advances made possible by generative AI, it is important that stakeholders maintain active oversight during the AI-assisted risk negotiation process to detect and rectify errors. Furthermore, a recent survey (150) found that approximately 39% of the US population aged 18 to 64 has used generative AI, suggesting the potential for more widespread uptake, which is essential to ensure that AI-assisted risk negotiation can benefit from, and provide benefits to, all participating stakeholders.

Future directions

While risk assessments have a long history of supporting food safety decision-making, most have focused on quantifying risks and evaluating risk reduction strategies, often without formally accounting for the associated costs or trade-offs or addressing factors considered more subjective, such as consumer risk perception (which may often reflect social and cultural factors, such as the importance of specific foods and food preparation practices for a given society). In contrast, fields such as emergency preparedness are moving toward impact forecasting (151), which involves considering both the likelihood and severity of hazards to facilitate trade-off decision-making; specific examples and discussion of trade-off decision-making in emergency preparedness include Dale at al. (152), Wild et al. (153), and Davidson et al. (154). Where feasible, using a single metric such as QALYs to quantify both benefits and unintended consequences can help rationalize trade-off decisions. Yet, even with unified risk metrics, society may assign different values to risks and costs incurred through different means. For example, in the field of health technology assessment, QALYs are used in cost–benefit analyses to help identify cost-effective healthcare interventions, but they can neglect societal preference for fairness by disregarding healthcare interventions for conditions that are not cost-effective to treat (e.g., rare, high-risk conditions with lower capacity for recovery) (155, 156). In the context of food safety, QALYs gained through food safety interventions may not be valued the same as those gained through reduced carbon emissions. Thus, we need further development of frameworks that support risk negotiation and allow societies to balance competing priorities and values (1).

While there are advantages to globally consistent food safety standards (e.g., for trade), trade-off decisions will likely have to consider whether a single food safety standard is socially acceptable in different regions with distinct societal expectations and needs. What is considered an acceptable risk or sustainable outcome in one context may not be in another. Tools such as multicriteria decision analysis can integrate diverse data and preferences, including health impacts, costs, and societal acceptability (140), though these methods may also introduce subjectivity in how interests are weighted. Importantly, providing policymakers with extensive options for risk-reduction interventions and their associated costs may not always lead to optimal decisions. As Mccaughey & Bruning (157) noted, policymakers and risk managers often exhibit bounded rationality and “satisficing”—choosing the first option that seems sufficient rather than the best possible one. Therefore, future efforts are needed to strengthen interdisciplinary research that incorporates social sciences, economics, life sciences, and systems biology to facilitate the development of food value chains that meet nutritional and safety needs while aligning with broader societal values. Improved communication between risk managers and assessors will also be important, ensuring that assessments and their metrics support rational, inclusive decision-making. Technological advances, such as GIS, AI, and genomics, can enhance the efficiency and precision of trade-off risk assessments (e.g., by accounting for virulence differences among strains within a given species). However, participatory and transparent approaches will be equally, if not more, essential in assuring evolving global food systems that meet safety, nutrition, sustainability, and accessibility goals.

References

• 1

  World Health Organization. WHO Global Strategy for Food Safety 2022–2030: towards stronger food safety systems and global cooperation. Geneva: WHO (2022). Available at: https://www.who.int/publications/i/item/9789240057685

  • Google Scholar

• 2

  World Health Organization. WHO estimates of the global burden of foodborne diseases: foodborne disease burden epidemiology reference group 2007–2015. Geneva: WHO (2015). Available at: https://iris.who.int/bitstream/handle/10665/199350/9789241565165_eng.pdf

  • Google Scholar

• 3

  World Health Organization. WHO’s first ever global estimates of foodborne diseases find children under 5 account for almost one third of deaths [online]. (2015). Available at: https://www.who.int/news/item/03-12-2015-who-s-first-ever-global-estimates-of-foodborne-diseases-find-children-under-5-account-for-almost-one-third-of-deaths

  • Google Scholar

• 4

  JaffeeSHensonSUnnevehrLGraceDCassouE. The Safe Food Imperative: accelerating progress in low- and middle-income countries. Washington, DC: World Bank (2019). doi: 10.1596/978-1-4648-1345-0

  • CrossRef
  • Google Scholar

• 5

  World Food Programme. WFP and the sustainable development goals [online]. Available at: https://www.wfp.org/sdgs

  • Google Scholar

• 6

  ZwieteringMHGarreAWiedmannMBuchananRL. All food processes have a residual risk, some are small, some very small and some are extremely small: zero risk does not exist. Curr Opin Food Sci (2021) 39:83–92. doi: 10.1016/j.cofs.2020.12.017

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 7

  Organisation for Economic Co-operation and Development. Safety evaluation of foods derived by modern biotechnology: concepts and principles. OECD (1993). Available at: https://www.iatp.org/sites/default/files/Safety_Evaluation_of_Foods_Derived_by_Modern_B.htm

  • Google Scholar

• 8

  GorrisL. Performance objectives and performance criteria – Two sides of the food chain. Mitt Lebensm Hyg (2004) 95:21–7. Available at: https://www.icmsf.org/wp-content/uploads/2018/02/021-027_Gorris.pdf

  • Google Scholar

• 9

  de SwarteCDonkerRA. Towards an FSO/ALOP based food safety policy. Food Control (2005) 16:825–30. doi: 10.1016/j.foodcont.2004.10.023

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10

  PetersonM. What is a de minimis risk? Risk Manag (2002) 4:47–55. doi: 10.1057/palgrave.rm.8240118

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 11

  FarberJMZwieteringMWiedmannMSchaffnerDHedbergCWHarrisonMAet al. Alternative approaches to the risk management of Listeria monocytogenes in low risk foods. Food Control (2021) 123:107601. doi: 10.1016/j.foodcont.2020.107601

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 12

  BuchananRL. Principles of risk analysis as applied to microbial food safety concerns. Mitt Lebensm Hyg (2004) 95:6–12. Available at: https://www.icmsf.org/wp-content/uploads/2018/02/006-012_Buchanan.pdf

  • Google Scholar

• 13

  Codex Alimentarius Commission. Principles and guidelines for the establishment and application of microbiological criteria related to foods (2013 edition). United Nations Food and Agriculture Organization (1993). Available at: https://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXG%2B21-1997%252FCXG_021e.pdf

  • Google Scholar

• 14

  DoménechEMartorellS. Definition and usage of food safety margins for verifying compliance of Food Safety Objectives. Food Control (2016) 59:669–74. doi: 10.1016/j.foodcont.2015.05.042

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 15

  United States Food and Drug Administration. A quantitative assessment of the net effects on fetal neurodevelopment from eating commercial fish (as measured by IQ and also by early age verbal development in children). US FDA (2014). Available at: https://www.fda.gov/media/88491/download

  • Google Scholar

• 16

  ScallanEHoekstraRMAnguloFJTauxeRVWiddowsonMARoySLet al. Foodborne illness acquired in the United States-Major pathogens. Emerg Infect Dis (2011) 17:7–15. doi: 10.3201/eid1701.P11101

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 17

  KoopmansMDuizerE. Foodborne viruses: an emerging problem. Int J Food Microbiol (2004) 90:23–41. doi: 10.1016/S0168-1605(03)00169-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 18

  KaperJBNataroJPMobleyHLT. Pathogenic Escherichia coli. Nat Rev Microbiol (2004) 2:123–40. doi: 10.1038/nrmicro818

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 19

  LeimbachAHackerJDobrindtU. E. coli as an all-rounder: the thin line between commensalism and pathogenicity. In: DobrindtUHackerJHSvanborgC, editors. Between pathogenicity and commensalism (1st edition). Berlin: Springer (2013). 3–32. doi: 10.1007/82_2012_303

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 20

  McCarthyNGieseckeJ. Incidence of Guillain-Barré syndrome following infection with Campylobacter jejuni. Am J Epidemiol (2001) 153:610–4. doi: 10.1093/aje/153.6.610

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 21

  Scallan WalterEJCrimSMBruceBBGriffinPM. Incidence of Campylobacter-associated Guillain-Barré syndrome estimated from health insurance data. Foodborne Pathog Dis (2020) 17:23–8. doi: 10.1089/fpd.2019.2652

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 22

  American College of Rheumatology. Reactive arthritis [online]. Available at: https://rheumatology.org/patients/reactive-arthritis

  • Google Scholar

• 23

  TownesJM. Reactive arthritis after enteric infections in the United States: the problem of definition. Clin Infect Dis (2010) 50:247–54. doi: 10.1086/649540

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 24

  Pogreba-BrownKAusthofETangXTrejoMJOwusu-DommeyABoydKet al. Enteric pathogens and reactive arthritis: systematic review and meta-analyses of pathogen-associated reactive arthritis. Foodborne Pathog Dis (2021) 18:627–39. doi: 10.1089/fpd.2020.2910

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25

  CheckleyWBuckleyGGilmanRHAssisAMGuerrantRLMorrisSSet al. Multi-country analysis of the effects of diarrhoea on childhood stunting. Int J Epidemiol (2008) 37:816–30. doi: 10.1093/ije/dyn099

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 26

  KnielKEKumarDThakurS. Understanding the complexities of food safety using a “One Health” approach. Microbiol Spectr (2018) 6:401–11. doi: 10.1128/microbiolspec.pfs-0021-2017

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 27

  ShaQForstnerMRJHahnD. Diversity of Salmonella in biofilms and water in a headwater ecosystem. FEMS Microbiol Ecol (2013) 83:642–9. doi: 10.1111/1574-6941.12021

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 28

  YanamalaSMillerMFLoneraganGHGraggSEBrashearsMM. Potential for microbial contamination of spinach through feedyard air/dust growing in close proximity to cattle feedyard operationS. J Food Saf (2011) 31:525–9. doi: 10.1111/j.1745-4565.2011.00330.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 29

  KumarGDWilliamsRCAl QublanHMSriranganathanNBoyerRREifertJD. Airborne soil particulates as vehicles for Salmonella contamination of tomatoes. Int J Food Microbiol (2017) 243:90–5. doi: 10.1016/j.ijfoodmicro.2016.12.006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 30

  LangholzJAJay-RussellMT. Potential role of wildlife in pathogenic contamination of fresh produce. Hum Wildl Interac (2013) 7:140–57. doi: 10.26077/e5gg-r037

  • CrossRef
  • Google Scholar

• 31

  ParkSNavratilSGregoryABauerASrinathISzonyiBet al. Multifactorial effects of ambient temperature, precipitation, farm management, and environmental factors determine the level of generic Escherichia coli contamination on preharvested spinach. Appl Environ Microbiol (2015) 81:2635–50. doi: 10.1128/AEM.03793-14

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 32

  EvenBThaiHTMPhamHTMBénéC. Defining barriers to food systems sustainability: a novel conceptual framework. Front Sustain Food Syst (2024) 8:1453999. doi: 10.3389/fsufs.2024.1453999

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 33

  RittelHWJWebberMM. Dilemmas in a general theory of planning. Policy Sci (1973) 4:155–69. doi: 10.1007/BF01405730

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 34

  LönngrenJvan PoeckK. Wicked problems: a mapping review of the literature. Int J Sust Dev World Ecol (2021) 28:481–502. doi: 10.1080/13504509.2020.1859415

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 35

  Statista. Number of chickens worldwide from 1990 to 2023 (in million animals)* [online] (2025). Available at: https://www.statista.com/statistics/263962/number-of-chickens-worldwide-since-1990/

  • Google Scholar

• 36

  Statista. Per capita consumption of fresh vegetables in the United States in 2023, by vegetable type (in pounds)* [online] (2024). Available at: https://www.statista.com/statistics/257345/per-capita-consumption-of-fresh-vegetables-in-the-us-by-type/

  • Google Scholar

• 37

  Food and Agriculture Organization of the United Nations. Feeding the world in 2050. Rome: FAO (2009). Available at: https://www.fao.org/4/k6021e/k6021e.pdf

  • Google Scholar

• 38

  Van DijkMMorleyTRauMLSaghaiY. A meta-analysis of projected global food demand and population at risk of hunger for the period 2010–2050. Nat Food (2021) 2:494–501. doi: 10.1038/s43016-021-00322-9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39

  PetruzziLCampanielloDSperanzaBCorboMRSinigagliaMBevilacquaA. Thermal treatments for fruit and vegetable juices and beverages: a literature overview. Compr Rev Food Sci Food Saf (2017) 16:668–91. doi: 10.1111/1541-4337.12270

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 40

  PiersonMDSmootLA. Nitrite, nitrite alternatives, and the control of Clostridium botulinum in cured meats. Crit Rev Food Sci Nutr (1983) 17:141–87. doi: 10.1080/10408398209527346

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 41

  SongPWuLGuanW. Dietary nitrates, nitrites, and nitrosamines intake and the risk of gastric cancer: a meta-analysis. Nutrients (2015) 7:9872–95. doi: 10.3390/nu7125505

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 42

  De MeyEDe MaereHPaelinckHFraeyeI. Volatile N-nitrosamines in meat products: potential precursors, influence of processing, and mitigation strategies. Crit Rev Food Sci Nutr (2017) 57:2909–23. doi: 10.1080/10408398.2015.1078769

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  BondonnoCPZhongLBondonnoNPSimMBlekkenhorstLCLiuAet al. Nitrate: the Dr. Jekyll and Mr. Hyde of human health? Trends Food Sci Technol (2023) 135:57–73. doi: 10.1016/j.tifs.2023.03.014

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 44

  BurtonNCGibbinsJ. Assessment of peracetic acid exposure among federal poultry inspectors. United States Department of Health and Human Services (2016). Available at: https://www.cdc.gov/niosh/hhe/reports/pdfs/2014-0196-3254.pdf

  • Google Scholar

• 45

  KarpDSBaurPAtwillERDe MasterKGennetSIlesAet al. The unintended ecological and social impacts of food safety regulations in California’s central coast region. Bioscience (2015) 65:1173–83. doi: 10.1093/biosci/biv152

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 46

  MartinNHTrmčićAHsiehTHBoorKJWiedmannM. The evolving role of coliforms as indicators of unhygienic processing conditions in dairy foods. Front Microbiol (2016) 7:1549. doi: 10.3389/fmicb.2016.01549

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47

  AnelichLECMLuesRFarberJMParreiraVR. SARS-CoV-2 and risk to food safety. Front Nutr (2020) 7:580551. doi: 10.3389/fnut.2020.580551

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48

  KitzRWalkerTCharleboisSMusicJ. Food packaging during the COVID-19 pandemic: consumer perceptions. Int J Consum Stud (2022) 46:434–48. doi: 10.1111/ijcs.12691

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 49

  KrugMDChapmanBDanylukMD. Establishing a lot through sanitation clean breaks in produce packing facilities (2020 edition). University of Florida Institute of Food and Agricultural Sciences (2013). Available at: https://www.citrusindustry.net/wp-content/uploads/2024/02/FS234.pdf

  • Google Scholar

• 50

  ZaragozaA. As Big Beer moves in, activists in Mexicali fight to keep their water. National Public Radio (2018). Available at: https://www.npr.org/sections/thesalt/2018/03/26/596448290/as-big-beer-moves-in-activists-in-mexicali-fight-to-keep-their-water

  • Google Scholar

• 51

  HouserMDorfmanJHRejesusRM. The long-term effects of meat recalls on futures markets. Appl Econ Perspect Policy (2019) 41:235–48. doi: 10.1093/aepp/ppy010

  • CrossRef
  • Google Scholar

• 52

  BarlowSMBoobisARBridgesJCockburnADekantWHepburnPet al. The role of hazard- and risk-based approaches in ensuring food safety. Trends Food Sci Technol (2015) 46:176–88. doi: 10.1016/j.tifs.2015.10.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53

  SsemandaJNden BestenHMWvan WagenbergCPAZwieteringMH. Quantitative assessment of food safety interventions for Campylobacter spp. and Salmonella spp. along the chicken meat supply chain in Burkina Faso and Ethiopia. Int J Food Microbiol (2024) 415:110637. doi: 10.1016/j.ijfoodmicro.2024.110637

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54

  American Academy of Microbiology. Clean Water: what is acceptable microbial risk? Washington, DC: American Society for Microbiology (2007). Available at: https://www.ncbi.nlm.nih.gov/books/NBK560444/

  • Google Scholar

• 55

  World Trade Organization. The WTO Agreement on the application of sanitary and phytosanitary measures (SPS Agreement). WTO (1995). Available at: https://www.wto.org/english/tratop_e/sps_e/spsagr_e.htm

  • Google Scholar

• 56

  ZwieteringMHJacxsensLMembréJMNautaMPeterzM. Relevance of microbial finished product testing in food safety management. Food Control (2016) 60:31–43. doi: 10.1016/j.foodcont.2015.07.002

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 57

  Unites States National Archives and Records Administration’s Office of the Federal Register, United States Government Publishing Office. Current good manufacturing practice, hazard analysis, and risk-based preventive controls for human food [part 117]. In: Code of Federal Regulations. OFR (2015). Available at: https://www.ecfr.gov/current/title-21/chapter-I/subchapter-B/part-117

  • Google Scholar

• 58

  LepperJAhnSSchneiderKRDanylukMDGoodrich-SchneiderR. The Food Safety Modernization Act of 2011: final rule for preventive controls for human food: FSHN17-6/FS301, 1/2018. EDIS (2018) 2018(1). doi: 10.32473/edis-fs301-2018

  • CrossRef
  • Google Scholar

• 59

  FarakosSMSPouillotRJohnsonRSpungenJSonIAndersonNet al. A quantitative assessment of the risk of human salmonellosis arising from the consumption of almonds in the United States: the impact of preventive treatment levels. J Food Prot (2017) 80:863–78. doi: 10.4315/0362-028X.JFP-16-403

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 60

  United States Food and Drug Administration. Risk assessments: Salmonella on tree nuts [online] (2017). Available at: https://www.fda.gov/food/risk-and-safety-assessments-food/risk-assessments-salmonella-tree-nuts

  • Google Scholar

• 61

  United Fresh Food Safety & Technology Council. Guidance on environmental monitoring and control of Listeria for the fresh produce industry. Second edition. Washington, DC: United Fresh Produce Association (2018). Available at: https://www.freshproduce.com/siteassets/files/reports/food-safety/guidance-on-environmental-monitoring-and-control-of-listeria.pdf

  • Google Scholar

• 62

  FrankRHBernankeBS. Principles of macroeconomics (4th edition). New York, NY: McGraw-Hill Irwin (2009).

  • Google Scholar

• 63

  XueYGengXKipropEHongM. How do spillover effects influence the food safety strategies of companies? New orientation of regulations for food safety. Foods (2021) 10:1–17. doi: 10.3390/foods10020451

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 64

  SolbergCTSørheimPMüllerKEGamlundENorheimOFBarraM. The devils in the DALY: prevailing evaluative assumptions. Public Health Ethics (2020) 13:259–74. doi: 10.1093/phe/phaa030

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 65

  NakaoYKawakamiHMiyazakiSSaitoMLuoYYamamotoKet al. Evaluation of health utility in trial-based cost-utility analyses for major cardiovascular disease: protocol for a systematic review. BMJ Open (2023) 13:e067045. doi: 10.1136/bmjopen-2022-067045

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 66

  DevleesschauwerBScharffRLKowalcykBBHavelaarAH. Burden and risk assessment of foodborne disease. In: RobertsT, editor. Food safety economics: incentives for a safer food supply (1st edition). Cham: Springer International Publishing (2018). 83–106. doi: 10.1007/978-3-319-92138-9_6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 67

  FengXKimDDCohenJTNeumannPJOllendorfDA. Using QALYs versus DALYs to measure cost-effectiveness: how much does it matter? Int J Technol Assess Health Care (2020) 36:96–103. doi: 10.1017/S0266462320000124

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 68

  GoldMRStevensonDFrybackDG. HALYs and QALYs and DALYs, oh my: similarities and differences in summary measures of population health. Annu Rev Public Health (2002) 23:115–34. doi: 10.1146/annurev.publhealth.23.100901.140513

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 69

  PozoVFSchroederTC. Evaluating the costs of meat and poultry recalls to food firms using stock returns. Food Policy (2016) 59:66–77. doi: 10.1016/j.foodpol.2015.12.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 70

  GomezCBMarksBP. Monetizing the impact of food safety recalls on the low-moisture food industry. J Food Prot (2020) 83:829–35. doi: 10.4315/JFP-19-553

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 71

  RiberaLAPalmaMAPaggiMKnutsonRMasabniJGAncisoJ. Economic analysis of food safety compliance costs and foodborne illness outbreaks in the United States. HortTechnology (2012) 22:150–6. doi: 10.21273/HORTTECH.22.2.150

  • CrossRef
  • Google Scholar

• 72

  BartschSMAstiLNyathiSSpikerMLLeeBY. Estimated cost to a restaurant of a foodborne illness outbreak. Public Health Rep (2018) 133:274–86. doi: 10.1177/0033354917751129

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 73

  ThomasMKVriezenRFarberJMCurrieASchlechWFazilA. Economic cost of a Listeria monocytogenes outbreak in Canada 2008. Foodborne Pathog Dis (2015) 12:966–71. doi: 10.1089/fpd.2015.1965

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 74

  PangHLambertiniEBuchananRLSchaffnerDWPradhanAK. Quantitative microbial risk assessment for Escherichia coli O157:H7 in fresh-cut lettuce. J Food Prot (2017) 80:302–11. doi: 10.4315/0362-028X.JFP-16-246

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 75

  LambertiniEBaroueiJSchaffnerDWDanylukMDHarrisLJ. Modeling the risk of salmonellosis from consumption of pistachios produced and consumed in the United States. Food Microbiol (2017) 67:85–96. doi: 10.1016/j.fm.2017.06.003

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 76

  SsemandaJNden BestenHMWDioneMMAmenuKKnight-JonesTJDZwieteringMHet al. Cost-effectiveness of interventions toward improving microbial food safety of chicken meat along supply chains in Burkina Faso and Ethiopia. Int J Food Microbiol (2025) 431:111086. doi: 10.1016/j.ijfoodmicro.2025.111086

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 77

  GellynckXMessensWHaletDGrijspeerdtKHartnettEViaeneJ. Economics of reducing Campylobacter at different levels within the Belgian poultry meat chain. J Food Prot (2008) 71:479–85. doi: 10.4315/0362-028X-71.3.479

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 78

  Codex Alimentarius Committee. Guidelines on the application of general principles of food hygiene to the control of Listeria monocytogenes in foods. Rome: United Nations Food and Agriculture Organization (2007). Available at: https://www.fao.org/fao-who-codexalimentarius/sh-proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FStandards%252FCXG%2B61-2007%252FCXG_061e.pdf

  • Google Scholar

• 79

  GlassKBedaleWUnruhD. It’s time to reformulate deli meats to reduce the risk of Listeria monocytogenes. Food Safety Magazine. (2024). Available at: https://www.food-safety.com/articles/9992-its-time-to-reformulate-deli-meats-to-reduce-the-risk-of-listeria-monocytogenes

  • Google Scholar

• 80

  United States Department of AgricultureFood Safety and Inspection Service. Controlling Listeria monocytogenes in post-lethality exposed ready-to-eat meat and poultry products. USDA FSIS (2014). Available at: https://www.fsis.usda.gov/guidelines/2014-0001

  • Google Scholar

• 81

  GottliebSLNewbernECGriffinPMGravesLMHoekstraRMBakerNLet al. Multistate outbreak of listeriosis linked to Turkey deli meat and subsequent changes in US regulatory policy. Clin Infect Dis (2006) 42:29–36. doi: 10.1086/498113

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 82

  CartwrightEJJacksonKAJohnsonSDGravesLMSilkBJMahonBE. Listeriosis outbreaks and associated food vehicles, United States 1998–2008. Emerg Infect Dis (2013) 19:1–9. doi: 10.3201/eid1901.120393

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 83

  United States Department of Agriculture Food Safety and Inspection Service. FSIS Directive 10,240.4, revision 4: Listeria rule verification activities. USDA FSIS (2022). Available at: https://www.fsis.usda.gov/sites/default/files/media_file/2020-08/10240.4.pdf

  • Google Scholar

• 84

  PradhanAKIvanekRGröhnYGeornarasISofosJNWiedmannM.Quantitative risk assessment for Listeria monocytogenes in selected categories of deli meats: impact of lactate and diacetate on listeriosis cases and deaths. J Food Prot (2009) 72:978–89. doi: 10.4315/0362-028X-72.5.978

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 85

  United States Department of AgricultureFood Safety and Inspection Service. Comparative risk assessment for Listeria monocytogenes in ready-to-eat meat and poultry deli meats report. USDA FSIS (2010). Available at: https://www.fsis.usda.gov/sites/default/files/media_file/2020-07/Comparative_RA_Lm_Report_May2010.pdf

  • Google Scholar

• 86

  EFSA Panel on Biological Hazards (BIOHAZ)RicciAChemalyMDaviesRFernández EscámezPSGironesRet al. Hazard analysis approaches for certain small retail establishments in view of the application of their food safety management systems. EFSA J (2017) 15:e04697. doi: 10.2903/j.efsa.2017.4697

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 87

  United States Food and Drug Administration. Control of Listeria monocytogenes in ready-to-eat foods: guidance for industry. Draft guidance. US FDA (2017). Available at: https://www.fda.gov/media/102633/download

  • Google Scholar

• 88

  European Association of Fruit and Vegetable Processors. Hygiene guidelines for the control of Listeria monocytogenes in the production of quick-frozen vegetables. Brussels: PROFEL (2020). Available at: https://profel-europe.eu/_library/_files/PROFEL_Listeria_mono_guidelines_November2020.pdf

  • Google Scholar

• 89

  United States Food and Drug Administration, United States Department of Agriculture. Quantitative assessment of relative risk to public health from foodborne Listeria monocytogenes among selected categories of ready-to-eat foods. US FDA (2003). Available at: https://www.fda.gov/media/124721/download?attachment

  • Google Scholar

• 90

  ZoellnerCWiedmannMIvanekR. An assessment of listeriosis risk associated with a contaminated production lot of frozen vegetables consumed under alternative consumer handling scenarios. J Food Prot (2019) 82:2174–93. doi: 10.4315/0362-028X.JFP-19-092

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 91

  BuchananRL. Microbiological risk assessment: origins, current status, and future challenges. Food Safety Magazine (2021). Available at: https://www.food-safety.com/articles/1754-cover-story-microbiological-risk-assessment-origins-current-status-and-future-challenges

  • Google Scholar

• 92

  HaasCN. Microbial dose response modeling: past, present, and future. Environ Sci Technol (2015) 49:1245–59. doi: 10.1021/es504422q

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 93

  United States Department of Agriculture Food Safety and Inspection Service, United States Environmental Protection Agency. Microbial risk assessment guideline: pathogenic organisms with focus on food and water. USDA FSIS (2012). Available at: https://www.fsis.usda.gov/sites/default/files/media_file/2020-07/Microbial_Risk_Assessment_Guideline_2012-001.pdf

  • Google Scholar

• 94

  PouillotRDelignette-MullerML. Evaluating variability and uncertainty separately in microbial quantitative risk assessment using two R packages. Int J Food Microbiol (2010) 142:330–40. doi: 10.1016/j.ijfoodmicro.2010.07.011

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 95

  FarakosSMSPouillotRDavidsonGRJohnsonRSpungenJSonIet al. A quantitative risk assessment of human salmonellosis from consumption of pistachios in the United States. J Food Prot (2018) 81:1001–14. doi: 10.4315/0362-028X.JFP-17-379

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 96

  HopeBKBakerAREdelEDHogueATSchlosserWDWhitingRet al. An overview of the Salmonella enteritidis risk assessment for shell eggs and egg products. Risk Anal (2002) 22:203–18. doi: 10.1111/0272-4332.00023

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 97

  United States Department of Agriculture Food Safety and Inspection Service. Salmonella Enteritidis risk assessment: shell eggs and egg products. USDA FSIS (1998). Available at: https://www.fsis.usda.gov/sites/default/files/media_file/2020-07/pdfrisk1.pdf

  • Google Scholar

• 98

  Food and Agriculture Organization of the United Nations, World Health Organization. Risk assessment of Listeria monocytogenes in ready to eat foods. Rome: FAO (2004). Available at: https://www.fao.org/fileadmin/templates/agns/pdf/jemra/mra4_en.pdf

  • Google Scholar

• 99

  United States Department of Agriculture Food Safety and Inspection Service. Quantitative risk assessment for Salmonella in raw chicken and raw chicken products. USDA FSIS (2024). Available at: https://www.fsis.usda.gov/sites/default/files/media_file/documents/Chicken_SRA_July2024.pdf

  • Google Scholar

• 100

  ChenROrsiRHGuariglia-OropezaVWiedmannM. Development of a modeling tool to assess and reduce regulatory and recall risks for cold-smoked salmon due to Listeria monocytogenes contamination. J Food Prot (2022) 85:1335–54. doi: 10.4315/JFP-22-025

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 101

  PintoGReyesGABarnett-NeefsCJungYJQianCWiedmannMet al. Development of a flexible produce supply chain food safety risk model: comparing tradeoffs between improved process controls and additional product testing for leafy greens as a test case. J Food Prot (2025) 88:100393. doi: 10.1016/j.jfp.2024.100393

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 102

  GaoSLewisGDAshokkumarMHemarY. Inactivation of microorganisms by low-frequency high-power ultrasound: a simple model for the inactivation mechanism. Ultrason Sonochem (2014) 21:454–60. doi: 10.1016/j.ultsonch.2013.06.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 103

  StrawnLKFortesEDBihnEANightingaleKKGröhnYTWoroboRWet al. Landscape and meteorological factors affecting prevalence of three food-borne pathogens in fruit and vegetable farms. Appl Environ Microbiol (2013) 79:588–600. doi: 10.1128/AEM.02491-12

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 104

  ChapinTKNightingaleKKWoroboRWWiedmannMStrawnLK. Geographical and meteorological factors associated with isolation of Listeria species in New York state produce production and natural environments. J Food Prot (2014) 77:1919–28. doi: 10.4315/0362-028X.JFP-14-132

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 105

  WellerDWiedmannMStrawnLK. Spatial and temporal factors associated with an increased prevalence of Listeria monocytogenes in spinach fields in New York state. Appl Environ Microbiol (2015) 81:6059–69. doi: 10.1128/AEM.01286-15

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 106

  KarpDSGennetSKilonzoCPartykaMChaumontNAtwillERet al. Comanaging fresh produce for nature conservation and food safety. Proc Natl Acad Sci USA (2015) 112:11126–31. doi: 10.1073/pnas.1508435112

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 107

  DawsonDEKeungJHNapolesMGVellaMRChenSSandersonMWet al. Investigating behavioral drivers of seasonal Shiga-Toxigenic Escherichia coli (STEC) patterns in grazing cattle using an agent-based model. PloS One (2018) 13:e0205418. doi: 10.1371/journal.pone.0205418

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 108

  ZoellnerCJenningsRWiedmannMIvanekR. EnABLe: an agent-based model to understand Listeria dynamics in food processing facilities. Sci Rep (2019) 9:495. doi: 10.1038/s41598-018-36654-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 109

  JungYJQianCBarnett-NeefsCIvanekRWiedmannM. Developing an agent-based model that predicts Listeria spp. transmission to assess Listeria control strategies in retail stores. J Food Prot (2024) 87:100337. doi: 10.1016/j.jfp.2024.100337

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 110

  MokhtariAVan DorenJM. An agent-based model for pathogen persistence and cross-contamination dynamics in a food facility. Risk Anal (2019) 39:992–1021. doi: 10.1111/risa.13215

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 111

  Barnett-NeefsCWiedmannMIvanekR. Examining patterns of persistent Listeria contamination in packinghouses using agent-based models. J Food Prot (2022) 85:1824–41. doi: 10.4315/JFP-22-119

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 112

  KaruppuchamyVHeldmanDRSnyderAB. A review of food safety in low-moisture foods with current and potential dry-cleaning methods. J Food Sci (2024) 89:793–810. doi: 10.1111/1750-3841.16920

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 113

  QianCMurphySIOrsiRHWiedmannM. How can AI help improve food safety? Annu Rev Food Sci Technol (2023) 14:517–38. doi: 10.1146/annurev-food-060721-013815

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 114

  ChhetriKB. Applications of artificial intelligence and machine learning in food quality control and safety assessment. Food Eng Rev (2024) 16:1–21. doi: 10.1007/s12393-023-09363-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 115

  WellerDLLoveTMTWiedmannM. Comparison of resampling algorithms to address class imbalance when developing machine learning models to predict foodborne pathogen presence in agricultural water. Front Environ Sci (2021) 9:701288. doi: 10.3389/fenvs.2021.701288

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 116

  ChungTYanRWellerDLKovacJ. Conditional forest models built using metagenomic data accurately predicted Salmonella contamination in northeastern streams. Microbiol Spectr (2023) 11:e00381–23. doi: 10.1128/spectrum.00381-23

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 117

  TonnerPDDarnellCLEngelhardtBESchmidAK. Detecting differential growth of microbial populations with Gaussian process regression. Genome Res (2017) 27:320–33. doi: 10.1101/gr.210286.116

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 118

  NjagePMKLeekitcharoenphonPHaldT. Improving hazard characterization in microbial risk assessment using next generation sequencing data and machine learning: predicting clinical outcomes in shigatoxigenic Escherichia coli. Int J Food Microbiol (2019) 292:72–82. doi: 10.1016/j.ijfoodmicro.2018.11.016

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 119

  KamilMZTaleb-BerrouaneMKhanFAmyottePAhmedS. Textual data transformations using natural language processing for risk assessment. Risk Anal (2023) 43:2033–52. doi: 10.1111/risa.14100

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 120

  Fernández-NavarroFValeroAHervás-MartínezCGutiérrezPAGarcía-GimenoRMZurera-CosanoG. Development of a multi-classification neural network model to determine the microbial growth/no growth interface. Int J Food Microbiol (2010) 141:203–12. doi: 10.1016/j.ijfoodmicro.2010.05.013

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 121

  KovacJBakkerHCarrollLMWiedmannM. Precision food safety: a systems approach to food safety facilitated by genomics tools. Trends Analyt Chem (2017) 96:52–61. doi: 10.1016/j.trac.2017.06.001

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 122

  United States Department of Agriculture Food Safety and Inspection Service. Expansion of FSIS shiga toxin-producing Escherichia coli (STEC) testing to additional raw beef products. USDA FSIS (2022). Available at: https://www.federalregister.gov/d/2022-25140

  • Google Scholar

• 123

  NightingaleKKWindhamKMartinKEYeungMWiedmannM. Select Listeria monocytogenes subtypes commonly found in foods carry distinct nonsense mutations in inlA, leading to expression of truncated and secreted internalin A, and are associated with a reduced invasion phenotype for human intestinal epithelial cells. Appl Environ Microbiol (2005) 71:8764–72. doi: 10.1128/AEM.71.12.8764-8772.2005

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 124

  NightingaleKKIvyRAHoAJFortesEDNjaaBLPetersRMet al. inlA premature stop codons are common among Listeria monocytogenes isolates from foods and yield virulence-attenuated strains that confer protection against fully virulent strains. Appl Environ Microbiol (2008) 74:6570–83. doi: 10.1128/AEM.00997-08

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 125

  ManuelCSSteltenAVWiedmannMNightingaleKKOrsiRH. Prevalence and distribution of Listeria monocytogenes inlA alleles prone to phase variation and inlA alleles with premature stop codon mutations among human, food, animal, and environmental isolates. Appl Environ Microbiol (2015) 81:8339–45. doi: 10.1128/AEM.02752-15

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 126

  OrsiRHBakkerHCWiedmannM. Listeria monocytogenes lineages: genomics, evolution, ecology, and phenotypic characteristics. Int J Med Microbiol (2011) 301:79–96. doi: 10.1016/j.ijmm.2010.05.002

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 127

  MauryMMTsaiYHCharlierCTouchonMChenal-FrancisqueVLeclercqAet al. Uncovering Listeria monocytogenes hypervirulence by harnessing its biodiversity. Nat Genet (2016) 48:308–13. doi: 10.1038/ng.3501

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 128

  MauryMMChenal-FrancisqueVBracq-DieyeHHanLLeclercqAValesGet al. Spontaneous loss of virulence in natural populations of Listeria monocytogenes. Infect Immun (2017) 85:10–1128. doi: 10.1128/IAI.00541-17

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 129

  MauryMMBracq-DieyeHHuangLValesGLavinaMThouvenotPet al. Hypervirulent Listeria monocytogenes clones’ adaption to mammalian gut accounts for their association with dairy products. Nat Commun (2019) 10:2488. doi: 10.1038/s41467-019-10380-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 130

  Food and Agriculture Organization of the United Nations, World Health Organization. Listeria monocytogenes in ready-to-eat (RTE) foods: attribution, characterization and monitoring. Rome: FAO (2022). doi: 10.4060/cc2400en

  • CrossRef
  • Google Scholar

• 131

  KimMBarnett-NeefsCChavezRAKealeyEWiedmannMStasiewiczMJ. 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. J Food Prot (2024) 87:100304. doi: 10.1016/j.jfp.2024.100304

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 132

  StruelensMJLuddenCWernerGSintchenkoVJokelainenPIpM. Real-time genomic surveillance for enhanced control of infectious diseases and antimicrobial resistance. Front Sci (2024) 2:1298248. doi: 10.3389/fsci.2024.1298248

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 133

  BerdenJHanley-CookGTChimeraBCakmakEKNicolasGBaudryJet al. Synergies between food biodiversity, processing levels, and the eat-lancet diet for nutrient adequacy and environmental sustainability: a multi-objective optimization using the EPIC cohort. Am J Clin Nutr (2025) 123:101115. doi: 10.1016/j.ajcnut.2025.11.006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 134

  DyeKCEggersJPShapiraZ. Trade-offs in a tempest: stakeholder influence on hurricane evacuation decisions. Organ Sci (2014) 25:1009–25. doi: 10.1287/orsc.2013.0890

  • CrossRef
  • Google Scholar

• 135

  DomonoskeC. Why didn’t officials order the evacuation of Houston. National Public Radio (2017). Available at: https://www.npr.org/sections/thetwo-way/2017/08/28/546721363/why-didn-t-officials-order-the-evacuation-of-houston

  • Google Scholar

• 136

  De HaesHAUSleeswijkAWHeijungsR. Similarities, differences and synergisms between HERA and LCA - an analysis at three levels. Hum Ecol Risk Assess (2006) 12:431–59. doi: 10.1080/10807030600561659

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 137

  KobayashiYPetersGMAshboltNJHeimerssonSSvanströmMKhanSJ. Global and local health burden trade-off through the hybridisation of quantitative microbial risk assessment and life cycle assessment to aid water management. Water Res (2015) 79:26–38. doi: 10.1016/j.watres.2015.03.015

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 138

  OlsenSIChristensenFMHauschildMPedersenFLarsenHFTùrslùvJ. Life cycle impact assessment and risk assessment of chemicals - a methodological comparison. Environ Impact Assess Rev (2001) 21:385–402. doi: 10.1016/S0195-9255(01)00075-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 139

  HarderRHolmquistHMolanderSSvanströmMPetersGM. Review of environmental assessment case studies blending elements of risk assessment and life cycle assessment. Environ Sci Technol (2015) 49:13083–93. doi: 10.1021/acs.est.5b03302

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 140

  KobayashiYPetersGMKhanSJ. Towards more holistic environmental impact assessment: hybridisation of life cycle assessment and quantitative risk assessment. Proc CIRP (2015) 29:378–83. doi: 10.1016/j.procir.2015.01.064

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 141

  MotaJDOGuillouSPierreFMembréJM. Public health risk-benefit assessment of red meat in France: current consumption and alternative scenarios. Food Chem Toxicol (2021) 149:111994. doi: 10.1016/j.fct.2021.111994

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 142

  TondoECGonçalvesCTH. Using risk-benefit analysis to control Salmonella in chicken meat. Food Qual Saf (2021) 5:970–1298. doi: 10.1093/fqsafe/fyab027

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 143

  MembréJMSantillana FarakosSNautaM. Risk-benefit analysis in food safety and nutrition. Curr Opin Food Sci (2021) 39:76–82. doi: 10.1016/j.cofs.2020.12.009

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 144

  RuzanteJMGriegerKWoodwardKLambertiniEKowalcykB. The use of multi-criteria decision analysis in food safety risk-benefit assessment. Food Prot Trends (2017) 37:132–9.

  • Google Scholar

• 145

  FarakosSMSPouillotRSpungenJFlanneryBVan DorenJMDennisS. Implementing a risk-risk analysis framework to evaluate the impact of food intake shifts on risk of illness: a case study with infant cereal. Food Addit Contam Part A (2021) 38:718–30. doi: 10.1080/19440049.2021.1885752

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 146

  TchonkouangRDOnyeakaHNkoutchouH. Assessing the vulnerability of food supply chains to climate change-induced disruptions. Sci Total Environ (2024) 920:171047. doi: 10.1016/j.scitotenv.2024.171047

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 147

  MirónIJLinaresCDíazJ. The influence of climate change on food production and food safety. Environ Res (2023). 216:114674. doi: 10.1016/j.envres.2022.114674

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 148

  Ehling-SchulzMFilterMZinsstagJKoutsoumanisKEllouzeMTeichmannJet al. Risk negotiation: a framework for One Health risk analysis. Bull World Health Organ (2024) 102:453–6. doi: 10.2471/BLT.23.290672

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 149

  SunYShengDZhouZWuY. AI hallucination: towards a comprehensive classification of distorted information in artificial intelligence-generated content. Humanit Soc Sci Commun (2024) 11:1–14. doi: 10.1057/s41599-024-03811-x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 150

  BickABlandinADemingDJLouisS. The rapid adoption of generative AI. Cambridge, MA: United States National Bureau of Economic Research. (2024) Available at: https://www.nber.org/system/files/working_papers/w32966/w32966.pdf

  • Google Scholar

• 151

  MerzBKuhlickeCKunzMPittoreMBabeykoABreschDNet al. Impact forecasting to support emergency management of natural hazards. Rev Geophys (2020) 58:e2020RG000704. doi: 10.1029/2020RG000704

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 152

  DaleMWicksJMylneKPappenbergerFLaegerSTaylorS. Probabilistic flood forecasting and decision-making: an innovative risk-based approach. Nat Hazards (2014) 70:159–72. doi: 10.1007/s11069-012-0483-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 153

  WildAJWilsonTMBebbingtonMSColeJWCraigHM. Probabilistic volcanic impact assessment and cost-benefit analysis on network infrastructure for secondary evacuation of farm livestock: a case study from the dairy industry, Taranaki, New Zealand. J Volcanol Geotherm Res (2019) 387:106670. doi: 10.1016/j.jvolgeores.2019.106670

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 154

  DavidsonRANozickLKWachtendorfTBlantonBColleBKolarRLet al. An integrated scenario ensemble-based framework for hurricane evacuation modeling: part 1—decision support system. Risk Anal (2020) 40:97–116. doi: 10.1111/risa.12990

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 155

  RichardsonJSchlanderM. Health technology assessment (HTA) and economic evaluation: efficiency or fairness first. J Mark Access Health Policy (2019) 7(1):1557981. doi: 10.1080/20016689.2018.1557981

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 156

  CaroJJBrazierJEKarnonJKolominsky-RabasPMcGuireAJNordEet al. Determining value in health technology assessment: stay the course or tack away?Pharmacoeconomics (2019) 37:293–9. doi: 10.1007/s40273-018-0742-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 157

  MccaugheyDBruningNS. Rationality versus reality: the challenges of evidence-based decision making for health policy makers. Implement Sci (2010) 5:1–13. doi: 10.1186/1748-5908-5-39

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 158

  Food and Agriculture Organization of the United Nations, World Health Organization. Principles and guidelines for incorporating microbiological risk assessment in the development of food safety standard, guidelines and related texts. Rome: FAO (2002). Available at: https://www.who.int/publications/m/item/guidelines-for-incorporating-microbiological-risk-assessment-food-safety-standards

  • Google Scholar

• 159

  FerdousJBensebaaFMilaniASHewageKBhowmikPPelletierN. Development of a generic decision tree for the integration of Multi-Criteria Decision-Making (MCDM) and Multi-Objective Optimization (MOO) methods under uncertainty to facilitate sustainability assessment: a methodical review. Sustainability (2024) 16:2684. doi: 10.3390/su16072684

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 160

  SendínJOHAlonsoAABangaJR. Efficient and robust multi-objective optimization of food processing: a novel approach with application to thermal sterilization. J Food Eng (2010) 98:317–24. doi: 10.1016/j.jfoodeng.2010.01.007

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 161

  BanachJLZwieteringMHvan der Fels-KlerxHJ. Multi-criteria decision analysis to evaluate control strategies for preventing cross-contamination during fresh-cut lettuce washing. Food Control (2021) 128:108136. doi: 10.1016/j.foodcont.2021.108136

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 162

  RuzanteJMWhitingRCDennisSBBuchananRL. Microbial risk assessment. In: DoyleMPBuchananRL, editors. Food microbiology: fundamentals and frontiers (4th edition). Washington, DC: ASM Press (2012). 225–6. doi: 10.1128/9781555818463.ch41

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 163

  CelliniSRKeeJE. Cost-effectiveness and cost-benefit analysis. In: NewcomerKEHatryHPWholeyJS, editors. Handbook of practical program evaluation (4th edition). Hoboken, NJ: John Wiley & Sons (2015). 636–72. doi: 10.1002/9781119171386.ch24

  • Pubmed Abstract
  • CrossRef
  • Google Scholar