The food industry continues to be complicated as new food products with unique ingredients are introduced into the marketplace, supply chain interruptions remain a challenge, and the demand for minimally processed foods grows. Ensuring that appropriate food safety parameters are utilized during the manufacture of food products, with the right supporting documentation, is paramount to providing customers with safe and wholesome food.

The Food Safety Modernization Act's (FSMA's) Current Good Manufacturing Practice, Hazard Analysis, and Risk-based Preventive Controls for Human Food regulation requires that preventive controls for food are validated. Per the regulation, validation "…must include obtaining and evaluating scientific and technical evidence (or, when such evidence is not available or is inadequate, conducting studies) to determine whether the preventive controls, when properly implemented, will effectively control the hazards." Validations for thermal processes are conducted to demonstrate that a specific thermal process can adequately eliminate or reduce the pathogen of concern to a reasonably safe level.

Performance standards are the specific pathogen reduction levels that must be attained during processing to ensure that proper food safety has been achieved. In the case of traditional canning, these limits have been established over decades of scientific testing of canned foods. The 12-D concept, which describes a thermal process that results in a 12-log reduction of Clostridium botulinum spores, has been established for low-acid canned foods as an acceptable performance standard that results in a very low health risk for consumers.

However, in the case of non-canning processing (heat treatment or an alternative), these performance standards can vary depending on the product type. For some non-canned products such as meats, poultry, and nuts, the performance standards have been established by previous scientific testing. For other food products where thermal processes such as baking or blanching are applied, sufficient scientific data is not available or is still in development to establish the proper performance standards for those respective items.  Applying previously established high-moisture products' cook time and temperature parameters to low-moisture products has been proven to be ineffective. Other variables must be considered, such as increased thermal resistance of Salmonella spp. in lower-moisture products, as demonstrated by some recent scientific investigations1 and the effect of impingement cooking on the surfaces of meat products.2,3

Performance standards can be chosen based on available scientific literature, scientific studies performed by the company, or by using risk-based pathogen modeling. In the case where there is no industry-available scientific literature to reference, food companies are expected to provide their own scientific studies and justifications for the safety of their food products.

Industry-Accepted Minimum 5-log Reduction Performance Standard

In the absence of other readily available performance standards, it has been common practice to utilize a minimum 5-log reduction as the performance standard. The minimum 5-log reduction performance standard has been previously established by the U.S. Food and Drug Administration (FDA) as acceptable in both the 2001 Juice Hazard Analysis and Critical Control Points (HACCP) regulation and for in-shell egg treatments. A minimum 5-log reduction in pathogens is also the accepted goal for the pasteurization processes.

Based on these examples, a minimum 5-log reduction in the pathogens of concern has been generally adopted by the industry as a good starting point for processing performance standards in the absence of other scientific information. Although this performance standard is generally accepted, it may not be applicable to the specific product and process being analyzed. Additionally, a 5-log reduction may not be required if other supporting information shows that the product will be safe.

Pathogens of Concern

One of the primary decisions to make when deciding on performance standards for a specific product or product category is to choose the correct pathogen(s) of concern. Multiple pathogens are of health concern to humans. Some of the most common foodborne pathogens are Campylobacter, Clostridium botulinum, Clostridium perfringens, Escherichia coli O157:H7, Listeria monocytogenes, Salmonella, and Staphylococcus aureus. Each of these pathogens have unique characteristics that separate them in terms of susceptibility to various processing conditions.

Some pathogens are more prevalent in certain product categories; however, the prevalence of the pathogen in the product category is only one consideration. Other considerations that pertain to choosing the correct pathogen(s) of concern include product attributes such as water activity and pH, storage conditions (ambient versus frozen), and heat resistance of the pathogens.

A raw material hazard analysis should be conducted to determine the likelihood of pathogen presence in the product or ingredients. A good reference to help determine the pathogen of concern is Appendix 1: Potential Hazards for Foods and Processes4 of FDA's Center for Food Safety and Applied Nutrition (CFSAN) guidance document Hazard Analysis and Risk-Based Preventive Controls for Human Food.

Industry-Established Performance Standards

Scientific testing and validation to establish performance standards have already been completed for certain food groups. These performance standards are usually found within scientific literature or government regulations. Each performance standard was developed using select parameters that must be met for each situation. These parameters include the minimum time and temperature requirements, oven types and cooking methods, specific processing conditions such as humidity, or even certain formulation stipulations such as maximum fat content or minimum salt content. Care must be taken to ensure that any required conditions are properly achieved to utilize those specific performance standards. If the parameters are not met, then the recommended performance standards might not be applicable for the specific product or process being used.

Meats and Poultry

Meat items such as beef, pork, and poultry are regulated by the U.S. Department of Agriculture (USDA). USDA's Food Safety and Inspection Service (FSIS) division has published its own guidance documents that state the performance standards needed to meet USDA's level of required food safety for meat and poultry items. At present, the most recent guidance document published by USDA to describe expected performance standards is FSIS Cooking Guideline for Meat and Poultry Products (Revised Appendix A), issued in December 2021.5 This document states that the pathogen of concern for cooked meats is Salmonella.

Per the guidance document: "For cooked products, FSIS recommends that establishments use Salmonella as an indicator of lethality because the thermal destruction of Salmonella in cooked products would indicate the destruction of most other pathogens" (64 FR 732). If the establishment's scientific support demonstrates that the lethality treatment achieves sufficient reduction in Salmonella, then it does not need to provide additional support that adequate reduction of other pathogens such as Shiga toxin-producing E. coli (STEC), Campylobacter, Listeria monocytogenes, Trichinae spiralis or Toxoplasma gondii is achieved. As stated in the FSIS Compliance Guideline HACCP Systems Validation, "Establishments should not use pathogens other than Salmonella as indicators of lethality for cooked products unless the alternate pathogen displays similar or higher resistance to the lethality processes."

Other preservation methods may require a different pathogen(s) of concern to be considered. The guidance document provides other information pertaining to the processing methods and reasoning behind the required performance standards for meats. Such information includes time/temperature tables to achieve the required performance standards. It is recommended to review the guidance document for full understanding of the information given in this section.


Industry-established performance standards are available for certain nut varieties. These performance standards may translate to other nut varieties, but this should be verified by a thermal processing expert before using the performance standard for validation purposes.

The Almond Board of California set a minimum performance standard based on a risk assessment and published this information in a final rule in the Federal Register in 2007.6 The final rule states, "…handlers must subject their almonds to a process that achieves a minimum 4-log reduction in Salmonella bacteria prior to shipment. The risk assessment model demonstrated that a minimum 4-log reduction provides an appropriate level of consumer protection. Thus, the Board concluded that a 4-log reduction was an appropriate standard for almonds."

Per the Consumer Brands Association's Industry Handbook for Safe Processing of Nuts,7 FDA "…currently suggests minimum 5-log reduction for peanuts and pistachios, unless data are available to support that less than 5-log is adequate."

Multiple Hurdle Concept

The multiple hurdle concept can be utilized as an acceptable pathway to achieve the minimum performance standard that has been chosen. This concept describes a series of processes or preservative factors (i.e., heating, decreasing water activity, decreasing pH, adding preservatives, etc.) that individually have a pathogen reduction step that is lower than the minimum performance standard target, but in combination will all add up to meet or exceed the minimum performance standard. To best utilize this concept, each process step should reside in the same facility, and storage time between steps should be minimized or eliminated to prevent unintended bacterial growth.

This concept is also predicated on keeping the transfer areas or equipment in a clean and safe condition to prevent contamination from increasing the bacterial load again. Each process or preservative factor is required to have its own validation proving out its individual pathogen reduction. Additionally, critical limits would need to be established to assure that each process or preservative factor will maintain the minimum conditions or parameters as tested in the validation. After the individual validations are complete, the processor can add the pathogen reductions for each process to show that the total minimum performance standard has been achieved. Numerous articles and references are available that can be researched to better understand the concept.8

Pathogen Modeling

In instances where a 5-log reduction is not appropriate or achievable for a product, it is crucial that tools are utilized to determine the risk level associated with the pathogen levels of the product in question. To ensure a 5-log reduction, processing conditions can be altered (i.e., longer dwell times, higher process temperatures, etc.) to meet the minimum 5-log reduction and subsequent validation testing that is repeated using the new processing parameters. However, it is possible that quality could be affected with longer processing times or higher temperatures, leading to food that is no longer palatable or, worse, making it unsaleable product.

In these situations, pathogen modeling can be utilized to determine whether a lower performance standard is applicable with the specific product, processing conditions, and set of ingredients to produce an acceptably low health risk for consumers.

Pathogen modeling programs can be used to predict the risk level of the respective food product. These programs are built using a risk-based approach in mind and utilize databases of testing results to predict the growth or inactivation of a pathogen(s), using specified processing or formulation parameters. For some, a pre-determined level of bacterial load is assumed. However, depending on the ingredients used and their sourcing, the starting levels of bacteria in the product can vary. Most have adjustable input parameters, but all usually have limitations to the inputs that are allowed. Modeling is useful only for scenarios that fall within the parameters that have been tested and should not be used for parameters that fall outside of the input limitations.

Battelle PRIA (Probabilistic Risk Informed Analysis) Model9

The Battelle model is a Probit dose response distribution model that is conservative in nature. The model uses pathogen prevalence data for the ingredients used in the formulation, with pathogen reduction measures used to predict the potential number of illnesses that could occur over a certain time. It is relatively easy to use and allows for flexibility in adjusting scenarios to fit different formulations and pathogen-reduction processes.

Only ingredients in the formulation that are micro-sensitive would need to be considered for the model because non-micro-sensitive ingredients, such as salt or those that have gone through a prior acceptable validation process, will not have any significant effect on the result.

To determine the prevalence of pathogens contained in the respective ingredients in the formulation, scientific literature or other industry-wide information searches should be conducted. Some ingredients have been evaluated by industry to show the prevalence of pathogens contained in the ingredient from different regions; however, not all ingredients have been evaluated for pathogen prevalence.

In the case that the formulation contains a micro-sensitive ingredient that has not been evaluated for pathogen prevalence, a conservative prevalence value must be used. As an example, wheat flour has been shown to have Salmonella prevalence values of 1.3–2 percent. As a conservative measure, the higher value should be utilized in the model. If a bakery formulation has been reformulated to contain rice flour instead of wheat flour, then prevalence data for rice flour should be evaluated. In this instance, however, there is little to no prevalence data for rice flour at present. Based on the ingredient, it could be theorized that rice flour should have less contamination issues with Salmonella than wheat flour; however, in the absence of prevalence data for rice flour, the 2 percent prevalence value for wheat flour can be used as a conservative measure, since it is a relatively similar ingredient.

Any assumptions that are made (such as in the example above) should be recorded for future reference. If future evaluations find that the pathogen prevalence for the specific ingredient is higher than that used in the model, then the model will need to be performed again with the new prevalence data.

If an inoculated validation study was performed on the product using certain processing conditions, in addition to the prevalence data, then the results from this study should be added to the model. The study should use inoculation of the pathogen(s) of concern or validated surrogate organism(s) at levels high enough to show the expected log reductions in the product. A statistical analysis of the inoculated validation study's log-reduction results will provide the necessary minimum, average, maximum, and standard deviation values for the model. This data can be used to compare the effects of the processing conditions that were used for the inoculated validation study to the overall food safety level for the product being processed.

Using this information in addition to other required inputs, the model will run a predetermined number of simulations to show the probability of consumer illnesses per million servings. Increasing numbers of simulations will provide better averaged data, but will take a longer time. Another simulation should be performed, using no microbiological controls, to establish a baseline probability of potential illness.

Using the results from the simulations and the baseline simulation, a percent risk reduction can be determined by dividing the potential illnesses from the simulation results into the number of illnesses from the baseline result. Using the percent risk reduction will allow a comparison of different processing conditions and formulations to determine the safety level of the scenario.

The company will need to determine an acceptable risk level that it would be willing to accept to ensure food safety. This determination can be informed by several factors, including risk levels associated with established performance standards covered earlier for certain food types that have been shown to be sufficient to ensure food safety. Once this has been determined, the results from the pathogen reduction modeling must be compared against the acceptable food safety risk level to decide if the modeling would result in a safe product.

If the validation testing showed a log reduction of less than 5, but the pathogen reduction modeling resulted in an acceptable risk level (as determined by several factors, mentioned above), then the processing parameters used for the validation will still provide a safe product. This methodology provides the justification for targeting a performance standard of less than a 5-log reduction for the validation.

FDA-iRISK Model10

FDA's CFSAN also provides a pathogen risk model for use. Per CFSAN's information about the model, "FDA-iRISK is a web-based system designed to analyze data concerning microbial and chemical hazards in food and return an estimate of the resulting health burden on a population level. The data required to execute this analysis include the food and its associated consumption data and processing/preparation methods, the hazard and its dose-response curve, and the anticipated health effects of the hazard when ingested by humans. Each of these elements contributes an essential piece of information to the model on which the final estimate of risk is based."

Other Pathogen Modeling programs

Other pathogen modeling programs can be used for predicting pathogen growth and reduction, as well as for predicting risk levels. These models, such as ComBase Predictor,11 USDA's Pathogen Modeling Program (PMP),12 the University of Wisconsin's Therm 2.0 Modeling,13 and others are more specific for certain food types, but do not estimate the final health risk of the consumer. These models can be very useful, but may be limited in the product types, processing methods, and product parameters (i.e., pH, salt content, water activity range, etc.).

As previously stated, ensuring that the consumer receives a safe food is the most important aspect for food production. Using suitable performance standards in relation to the appropriate pathogen(s) of concern to develop thermal processing parameters for food is imperative to ensure food safety.


  1. Lucore, L., A. J. Gualtieri, and S. Abd, 2017. "A Thermal Process Lethality Model for Low Water Activity Foods." Food Protection Trends 37, no. 1, 43–55.
  2. Hildebrandt, I. et al. "Process Humidity Affects Salmonella Lethality at the Surface and Core of Impingement-Cooked Meat and Poultry Products." Journal of Food Protection 84, no. 9 (2021): 1512–1523.
  3. Sindelar, J. J., R. Hanson, K. A. Glass, A. Milkowski, R. P. McMinn, and J. Nehls. "Validating a Method to Ensure the Destruction of Salmonella on Product Surfaces During Impingement Cooking." Meat and Muscle Biology 5, no. 1 (2021): 1–20.
  4. FDA CFSAN. 2018. "Appendix 1: Potential Hazards for Foods and Processes." Hazard Analysis and Risk-Based Preventive Controls for Human Food
  5. USDA FSIS. December 2021. FSIS Cooking Guideline for Meat and Poultry Products (Revised Appendix A)
  6. USDA Agricultural Marketing Service. 2007. Final Rule: Almonds Grown in California; Outgoing Quality Control Requirements. Federal Register 72, no. 61 (March 30, 2007): 15021–15036.
  7. Consumer Brands Association. 2010. Industry Handbook for Safe Processing of Nuts
  8. Leistner, I. "Basic aspects of food preservation by hurdle technology." International Journal of Food Microbiology 55 (2000): 181–186. aspects of food preservation.pdf.
  9. Battelle. "Probabilistic Risk Informed Analysis."
  10. FDA CFSAN. Joint Institute for Food Safety and Applied Nutrition (JIFSAN) and Risk Sciences International (RSI). FDA-iRISK® version 4.2. 2021.
  11. ComBase. "A Web Resource for Quantitative and Predictive Food Microbiology."
  12. USDA Agriculture Research Service. "Pathogen Modeling Program (PMP)."
  13. University of Wisconsin. Center for Meat Process Validation. "Therm 2.0."

Adam Woodworth is Manager for Thermal Processing at Conagra Brands.

Stephanie Nguyen is Director for Food Safety and Thermal Processing at Conagra Brands.