Food Safety Risk Assessment: How Scientists Evaluate Chemical and Microbial Hazards

By Dr. Chinta SidharthanReviewed by Benedette Cuffari, M.Sc.

Introduction
What Is Food Safety Risk Assessment?
The Four Stages of Food Safety Risk Assessment
Assessing Chemical and Microbial Food Hazards
Measuring Dietary Exposure and Cumulative Risk
Uncertainty and Vulnerable Populations
From Scientific Assessment to Food Safety Regulation
Conclusions
References
Further reading

What happens after scientists detect a contaminant in food? Exposure estimates, dose-response models, and consumer susceptibility shape the subsequent investigation.

Image Credit: THICHA SATAPITANON / Shutterstock.com

Introduction

Finding a toxic metal or pathogenic bacterium in a food sample begins an evaluation of exposure and potential health effects. Modern laboratory methods can detect substances at minute concentrations, and detection alone says little about whether anyone will be harmed. What matters is how much of the agent reaches consumers, how often, and what that amount does in the body.¹

A product that meets every legal limit may still deserve attention when several contaminants occur together or when consumers are unusually sensitive. In this article, using chemical and microbial examples, we explore how scientists distinguish the presence of a hazard from the likelihood and severity of harm under defined exposure conditions.^1,2

What Is Food Safety Risk Assessment?

A food safety hazard is a biological, chemical, or physical agent in food that can cause harm. Food safety risk describes how likely harm is and how severe it would be under defined exposure conditions.¹ Exposure is estimated from the amount of an agent that a person or population actually ingests, usually expressed per kilogram of body weight per day for chronic chemical exposure.²

Risk assessment is the organized scientific process that connects these concepts of hazards, exposure, and risk. Its purpose is to estimate the probability and magnitude of adverse effects, thereby allowing decision makers to rank problems by their estimated impact on public health instead of reacting to every detection.¹

In microbiology, this assessment is also kept separate from management and communication functions, which handle policy choices and public messaging. This separation protects the integrity of the evaluation while still allowing its results to guide control measures and consumer advice.³

The Four Stages of Food Safety Risk Assessment

Hazard identification aims to determine whether an agent can cause harm. Assessors compile evidence from human studies, animal experiments, and laboratory tests, and may rely on structural similarity to known compounds when data are scarce.²

At this stage, the International Agency for Research on Cancer (IARC) contributes by classifying carcinogenic potential, though its classifications do not quantify risk at a particular dietary exposure.² For microbes, hazard identification focuses on identifying the pathogen and illness it causes.⁴

Hazard characterization, often called dose-response assessment, links the dose to the frequency or intensity of the effect. For chemicals assessed using a threshold approach, animal studies can provide a no-observed-adverse-effect level or dose-response data for estimating a benchmark dose. The no-observed-adverse-effect level or lower confidence limit of the benchmark dose can then be divided by uncertainty factors to derive an acceptable or tolerable daily intake.²

For microbes, quantitative microbiological risk assessment (QMRA) uses mathematical models to relate the ingested dose to the probability of infection or illness. For example, for Listeria monocytogenes, QMRAs commonly use the exponential dose–response model.^4,7

Exposure assessment combines how much of a food people eat with how much of the agent that food contains. For chemicals, it estimates intake from the foods included in the assessment.^2,5 For bacteria, it must also follow growth and decline along the food chain until the food is eaten. Processing, cross-contamination, and consumer handling can also change the ingested dose.^4,7

Risk characterization combines hazard identification and characterization with exposure assessments to describe how likely and how serious the harm might be for a defined population. A hazard quotient compares estimated chemical exposure with an oral reference dose, and a value above 1 indicates potential concern. This quotient is a screening measure, not a numerical probability of illness or proof that harm will occur. Exposure may also be expressed as a proportion or percentage of an acceptable or tolerable intake.^2,6

For carcinogens that act through damage to deoxyribonucleic acid (DNA), the margin of exposure compares a toxicological reference point, often a benchmark dose lower confidence limit, with estimated human exposure. A larger margin generally indicates lower concern; it does not directly quantify cancer probability.² Microbial results may be expressed as illness cases per year, which policymakers may find easier to interpret than a probability per serving.⁴

Image Credit: TSViPhoto / Shutterstock.com

Assessing Chemical and Microbial Food Hazards

Policymakers often assess chemical contaminants such as pesticide residues and metals using health-based guidance values. Chronic assessments examine whether long-term intake exceeds these values. In contrast, acute assessments examine short-term exposure, such as intake during a single meal or day, using an acute reference dose when appropriate.

Some metals accumulate in the body, and their toxicity depends on the element, chemical form, dose, and exposure duration. Lead and mercury can damage the nervous system, while cadmium is associated with kidney damage.²

A study of rice in northern and northeastern Iran shows why detection and risk differ. The study identified chromium and ochratoxin A (OTA) as the main contributors to its calculated cumulative hazard index, with their relative contributions varying by region. The study detected diazinon, an organophosphate insecticide, in grain, but its contribution to the hazard index was negligible under the study's exposure assumptions.⁶

The chromium estimate requires caution: the researchers applied the reference dose for hexavalent chromium to total chromium measurements. This may overestimate chromium-related risk, as chromium in rice is expected to be predominantly trivalent.⁶

Naturally occurring toxins, including those produced by certain fungi, add another dimension to the problem because growing conditions, temperature, humidity, and storage influence their levels. Aflatoxin B1 injures the liver and can cause cancer, while OTA can damage the kidneys, and zearalenone has estrogenic effects.² Moisture control after harvest is a preventive measure that limits fungal growth and mycotoxin production.⁶

Microbial pathogens differ because their numbers can change along the food production cycle. Microbial contamination can vary by orders of magnitude from production; therefore, models must describe growth under realistic temperatures and storage times.⁴ Limited data, food matrix effects, uncertain dose-response relationships, and microbial adaptability have been major obstacles to reliable microbial risk assessment.^3,7

Measuring Dietary Exposure and Cumulative Risk

For chronic chemical exposure, estimated daily intake is calculated by multiplying each food's concentration by the daily amount eaten, summing the contributions across foods, and dividing by body weight.⁵ Consumption figures are obtained from dietary recalls and food records, while total diet studies analyze typical foods prepared as eaten, often pooled into composite samples.²

Because concentration data often include results below the detection limit, researchers calculate lower-, middle-, and upper-bound scenarios to show how handling those results shifts the estimate. In the Chongqing study, researchers replaced values below the detection limit with zero, half the detection limit, or the full detection limit, respectively.⁵

Dose-response information then provides the benchmark against which intake is compared. People can be exposed to several substances at the same time. In a study of Chongqing residents, researchers grouped lead, cadmium, arsenic, and mercury by their shared effects on the kidney and nervous system.

Here, the researchers used a modified Reference Point Index to combine exposure estimates with toxicity reference points and substance-specific uncertainty factors. At mean exposure, each food category's index was below 1, yet the combined index exceeded 1 under most scenarios.⁵

These calculations assume dose addition for the chemicals grouped for assessment. Grouping should be supported by shared toxic effects or other toxicological evidence, as interactions may increase or decrease toxicity, but an additive index alone does not demonstrate such interactions.^2,5 The Iranian rice study used a broader screening approach, summing hazard quotients across metals, diazinon, and mycotoxins. Its overall index depends on the assumption of dose addition across these contaminants and should be distinguished from the endpoint-specific grouping used in Chongqing.^5,6

Image Credit: Nan Tun Nay / Shutterstock.com

Uncertainty and Vulnerable Populations

Every stage of food safety risk assessment carries uncertainty from analytical error, limited sampling, self-reported diets, and incomplete toxicity data. Deterministic calculations yield a single number, while probabilistic assessments treat concentrations and consumption as distributions and produce a range of exposures.

Variability describes real differences between people, foods, or conditions; uncertainty reflects incomplete knowledge. Distinguishing the two helps explain whether better data could reduce the range of estimates.^2,4

Chemical-form assumptions provide another source of uncertainty. In Chongqing, researchers measured total arsenic and mercury and used conversion factors to estimate inorganic arsenic, methylmercury, and inorganic mercury. The resulting exposure estimates depend partly on how closely those assumed proportions represent the foods assessed.⁵

Input choices can also change conclusions considerably. In a Listeria model for cooked ham, adding variability to the dose-response parameter representing host susceptibility and strain virulence raised predicted illnesses more than fifty-fold, from 191 to 10,397 cases. These were alternative model predictions, not observed illness counts or evidence that actual incidence had increased.⁴

Susceptible groups further complicate risk assessment. Children often consume more food relative to body weight than adults, which can increase their estimated chemical exposure. In the Iranian rice study, the hazard index was consistently higher in children because the assumed rice intake per kilogram of body weight was greater.^2,6

Some chemicals can also cross the placenta, and exposure in fetal or infant life may cause long-term health consequences for the offspring. Assessments must account for developmental susceptibility rather than assume that effects are identical across life stages.²

For microbial hazards, susceptibility also varies with factors such as age and underlying health status. Microbial models often assign separate parameters to general and susceptible populations.⁴

New Approach Methodologies, including laboratory cell systems, computational models, and molecular profiling, are being developed to support chemical hazard characterization. For microbial assessment, genomic data and other newer methods also show promise, but data gaps, validation, and added model complexity remain concerns.^1,3,7

From Scientific Assessment to Food Safety Regulation

Assessment findings are not rules in themselves. Risk managers weigh them alongside other considerations, such as economic feasibility, technical capability, and public concern, to set maximum limits, product restrictions, or consumer warnings. Health-based guidance values, including acceptable daily intakes, are derived through scientific assessment and inform these management decisions.²

Food safety assessment also distinguishes between a potential hazard and the actual risk to consumers. For example, while the IARC has classified the sweetener aspartame as possibly carcinogenic to humans, the Joint FAO/WHO Expert Committee on Food Additives retained its acceptable daily intake of 0-40 mg/kg body weight per day. The review presents these conclusions as an example of the difference between identifying carcinogenic potential and assessing risk at dietary exposure levels.²

Assessing chemicals individually may overlook concern from combined exposure. Cumulative assessments can identify which contaminants and foods contribute most to a screening index and help prioritize monitoring and control.

In Chongqing, lead and cadmium drove the assessed neurological and kidney effects, while rice, leafy vegetables, and wheat products were major dietary contributors. These findings support targeted investigation and exposure reduction, with decisions informed by the assessment's assumptions and uncertainties.⁵

Image Credit: Dusan Petkovic / Shutterstock.com

Conclusions

Food safety risk assessment connects evidence about hazards with estimates of consumer exposure and possible health effects. Evidence-based risk assessments weigh toxicity, exposure, vulnerability, and uncertainty together, which allows limited resources to be directed toward problems with the greatest estimated public health impact.^1,2

Accounting for uncertainty, combined exposures, and susceptible groups such as children and pregnant women gives regulators a scientific basis for setting limits and prioritizing controls. Clear communication should explain what the estimates show, which assumptions affect them, and how these findings inform proportionate food safety decisions.^2,4,5

References 

1. Bragotto, A. P. A. (2026). Risk Assessment in Food Safety. Foods, 15(7), 1190. DOI: 10.3390/foods15071190. https://www.mdpi.com/2304-8158/15/7/1190
2. Pekmezci, H., Sipahi, S., & Başaran, B. (2025). Health Risk Assessment of Dietary Chemical Exposures: A Comprehensive Review. Foods 14(23); 4133. DOI: 10.3390/foods14234133. https://www.mdpi.com/2304-8158/14/23/4133
3. Fetsch, A., Esteban Cuesta, I., Alinaghi, M., et al. (2026). Principles and approaches applied for foodborne microbial risk assessment: A scoping review of concepts, limitations, and future directions. Food Microbiology 138; 105081. DOI: 10.1016/j.fm.2026.105081. https://www.sciencedirect.com/science/article/pii/S0740002026000456?via%3Dihub
4. van der Vossen-Wijmenga, W. P., Serra-Castelló, C., den Besten, H. M., & Zwietering, M. H. (2025).Analyzing quantitative microbiological risk assessments: From modeling approaches to informed decision-making and risk communication. Current Opinion in Food Science, 64, 101326. DOI: 10.1016/j.cofs.2025.101326. https://www.sciencedirect.com/science/article/pii/S2214799325000566
5. Chen, J., Chen, J., Li, M., et al. (2025). Probabilistic assessment of the cumulative risk from dietary heavy metal exposure in Chongqing, China using a hazard-driven approach. Scientific Reports, 15(1). DOI: 10.1038/s41598-024-83299-2. https://www.nature.com/articles/s41598-024-83299-2
6. Farboodi, R., Didar, Z., Vazifedoost, M., & Hajirostamloo, B. (2026). Health risk assessment of combined exposure to heavy metals,s diazino,n and mycotoxins in Iranian rice. Scientific Reports, 16. DOI: 10.1038/s41598-025-34310-x. https://www.nature.com/articles/s41598-025-34310-x
7. Orahovac, A., Raičević, N., Martinović, A., et al. (2026). A Bibliometric and Systematic Review of Quantitative Microbial Risk Assessment in Food Safety (1995–2024). Foods, 15(7), 1197. DOI: 10.3390/foods15071197. https://www.mdpi.com/2304-8158/15/7/1197

Further Reading

• All Food Safety Content
• Is Food Packaging Harmful to Health?
• How Food Safety Technology Improves Monitoring Across Global Supply Chains
• How Foodborne Pathogens Survive Food Processing and Preservation
• Foodborne Outbreak Investigations: Tracing Food Poisoning to Its Source

Last Updated: Oct 25, 2026