What Is Horizon Scanning in Food Safety?
Horizon scanning in food safety finds emerging hazards, fraud pressure and regulatory change before they reach your supply chain. What it is, what it covers, which sources to watch and how to run it.
Horizon scanning in food safety is how a food business finds the next hazard before it becomes the next alert: a continuous, documented search of regulatory, scientific, enforcement and media signals for early evidence of contamination, fraud pressure or regulatory change. This guide explains what horizon scanning is, whether it includes early warning, what it should cover, which sources to watch, how to keep alerts relevant and how to run it step by step — with the methods EFSA, the UK Food Standards Agency and FAO use. For the risks currently on the horizon, see our overview of emerging food safety risks for 2026.
What is horizon scanning in food safety?
Horizon scanning in food safety is the systematic, continuous search of regulatory, scientific, enforcement, trade and media information for early signs of hazards, fraud pressure and regulatory change that could reach a food supply chain. The UK Government defines horizon scanning as “a systematic examination of information to identify potential threats, risks, emerging issues and opportunities”, a definition Food Standards Scotland uses for its own programme.[1] The European Food Safety Authority puts the emphasis on timing: horizon scanning seeks early weak signals that may evolve into emerging issues but are not yet found in current research or media.[2] Both definitions describe the same discipline. A scan does not wait for a confirmed incident, a published limit or a recall; it looks for the first, incomplete evidence that something is changing, and it records what was seen and what was done about it, so that the decision can be explained later.
In practice this separates horizon scanning from routine monitoring. Monitoring checks known hazards against known limits. Horizon scanning asks which hazards, origins and rules should be added to that list next.
Does horizon scanning include early warning?
Horizon scanning should include early warning, but the term is used for two different things. One is a feed of official notifications: border rejections, alerts and recalls about hazards that have already been found in products in trade. That information is essential for compliance, but by its nature it reports a problem after it has happened somewhere else. The other is early warning in the sense EFSA uses: the search for weak signals that may evolve into emerging issues but are not yet found in current research or media.[2] Those signals sit upstream of any notification — a harvest loss, a price spike, a local enforcement finding, a new scientific result. Further along the same timeline, EFSA defines an emerging risk as one resulting from a newly identified hazard to which significant exposure may occur, or from an unexpected new or increased significant exposure or susceptibility to a known hazard.[2] FAO’s foresight looks further still, at issues that will or could emerge in the medium to long term.[3]
A food business needs the near end of this timeline most: signals early enough to act on, specific enough to name a material, an origin or a supplier. A scan that only relays notifications tells a team what others have already found.
Why does horizon scanning matter for food businesses?
Horizon scanning matters because a hazard analysis, a vulnerability assessment or a threat assessment is only as current as the information that feeds it. Specifications, test plans and supplier approvals are written against the hazards a team already knows about, so a new contaminant, a new adulterant or a new limit stays invisible until something prompts a review. The 2023 cinnamon case shows the cost of that gap. The US Food and Drug Administration traced elevated lead and chromium in cinnamon applesauce pouches to lead chromate in cinnamon from Ecuador; its highest lead result was 5,110 ppm, more than 2,000 times the 2.5 ppm level being considered by Codex for bark spices, and its leading hypothesis was economically motivated adulteration.[4] FDA notes that lead chromate has historically been added to spices to increase weight and enhance colour. By March 2024 the CDC had received reports of 519 cases.[4] A hazard with that history is exactly what a horizon scan is meant to keep on the list.
Our analysis of the cinnamon recall looks at how early the signals around that hazard appeared.
Early signals, filtered to what you buy.
iComplai’s horizon scanning watches the signals that come before a notification as well as the notifications themselves, and links each one to your ingredients, origins and suppliers. The weekly view shows what has changed for you, not everything that has changed.
See emerging risk monitoringWhat should a food safety horizon scan cover?
A food safety horizon scan should cover every route by which a material can become unsafe, illegal or misdescribed, not only the hazards already on the test plan. That means chemical contaminants and process contaminants; pesticide residues and changing maximum residue levels; mycotoxins, whose pressure moves with weather and harvests; microbiological hazards; allergens and labelling; food contact materials; animal and plant diseases that change what can be traded from where; and fraud — dilution, substitution and false claims of origin, species or production method. It should also cover regulation itself: a new limit, a withdrawn approval or a change in import controls can turn a compliant lot into a non-compliant one without any change in the product. The useful test for scope is simple. If an event would change a specification, a test, a supplier approval or a label, it belongs in the scan.
Two areas carry most of the volume for many buyers: pesticide residues and food fraud, where price and supply pressure often move before the hazard is confirmed.
Which sources should a horizon scan monitor?
A horizon scan should monitor official, scientific, enforcement and open sources together, because each sees a different stage of the same problem. Official sources include rapid alert systems such as RASFF, national recall and border-control databases, EFSA scientific opinions and the Official Journal of the EU for legal changes. Scientific sources include peer-reviewed journals and reference-laboratory notes. Enforcement intelligence comes from inspections, sampling programmes and investigations; Food Standards Scotland draws on publicly available and proprietary sources, direct engagement with government departments, local authorities and industry, and on enforcement activity and incident investigation.[1] Open sources — local news, trade press and market data — often carry the first signal, in the language of the producing country. The UK Food Standards Agency built a signal prioritisation dashboard that consolidates signals from over 50 open sources each day and translates foreign-language sources automatically.[6]
Most teams start with the official layer; our guide to RASFF notifications covers how to receive them without checking the portal by hand.
How do you avoid alert overload in horizon scanning?
Alert overload happens when a scan collects widely but filters weakly: every notification, article and opinion reaches the team, and the few that concern its own materials are buried among the rest. The cause is arithmetic. A wide scan has to read many sources every day, in many languages, while a single business buys a limited list of materials from a limited list of origins. Most of what arrives is accurate and irrelevant. The fix is relevance filtering, not fewer sources. Each signal is matched to a material, an origin and, where possible, a supplier. It is compared with the usual level for that combination, so that a real rise stands out from routine noise. And it is ranked, so the review starts with the handful of changes that could alter a specification, a test plan or an approval. Narrowing the sources instead brings back the blind spot that horizon scanning exists to close.
Filtering by supplier as well as by material turns the scan into an input for supplier risk monitoring: a signal about one origin becomes a question for the suppliers who source from it.
How does horizon scanning work, step by step?
Horizon scanning works as a repeating cycle of six steps. First, set the scope: the materials, origins, suppliers and markets that matter, and the hazards and rules that apply to them. Second, collect signals continuously from the sources above, in every relevant language. Third, filter: remove duplicates, translate, and match each signal to a material, an origin and a hazard. Fourth, assess and prioritise — how credible is the signal, how severe would the hazard be, and how exposed is this business? Fifth, act: a supplier question, a targeted test, a specification change, or a decision to watch. Sixth, document the signal, the assessment and the action, and review the scope when products, suppliers or rules change. The record matters as much as the scan. An auditor or a customer will ask not only what was known, but when, and what was done about it.
Steps four and five are where a scan connects to a formal risk assessment: the signal changes the likelihood or the severity, and the assessment decides what changes in response.
How do EFSA, the FSA and FAO approach horizon scanning?
Authorities run horizon scanning as a standing function, not a project. Under Articles 23 and 34 of Regulation (EC) No 178/2002, EFSA is required to identify emerging risks, and it works through an Emerging Risks Exchange Network with Member States, a stakeholder discussion group and a network of EU agencies, publishing an annual report on its emerging-risk activities; the report for 2025 covered 55 topics, 43 of them analysed in depth.[2] The UK Food Standards Agency describes its signal prioritisation dashboard as a horizon scanning application that identifies global food, feed and food contact material signals for the identification of emerging hazards. [6] FAO works at the longest range, through foresight reports such as Thinking about the future of food safety (2022).[3] The common thread is method: many sources, a defined process for prioritising what they find, and a published record of the outcome. A food business can borrow that method at its own scale.
Should horizon scanning be done manually or with software?
Most food businesses need both: software for collection and filtering, and people for judgement. A manual scan is limited by the number of sources a team can read and the languages it reads them in, and it is hard to keep steady through holidays and audits. The first report of an incident can appear locally, in the language of the producing country, before an alert reaches a European or US system. The EU’s ethylene oxide incident shows the scale such a signal can reach: residues of ethylene oxide in sesame seeds from India led to a meeting of the EU’s food and feed crisis coordinators on 9 October 2020.[5] Software can collect, translate, de-duplicate and match signals to materials and origins every day. What it cannot do alone is decide what matters for a particular specification, supplier or customer.
When you compare horizon scanning software, ask four questions. Does it show signals that come before an official notification, or only the notifications themselves? Does it filter to your materials, origins and suppliers, or leave the team to sort every alert? Does it read sources in the producing countries’ languages? And does it keep an auditable record of what was flagged and when?
How does iComplai approach horizon scanning?
iComplai approaches horizon scanning as early warning plus relevance filtering. The platform monitors more than 150,000 data sources across 240 countries — authority notifications, research papers, social media and news — which comes to roughly 100,000 signals a day.[7] Notifications are part of that, but the scan also watches what moves before them: harvest losses, commodity price spikes and trade disruptions that change contamination and fraud pressure. In food fraud risk prediction, a harvest-loss signal flagged in February 2025 preceded a 274% rise in hazelnut mycotoxin notifications by October 2025.[8] Every signal is then linked to the raw materials, origins and suppliers a customer actually uses, so a team reviews the changes that concern its own portfolio rather than every alert published that week. Prediction models extend the view forward, forecasting pesticide residue, fraud and notification risks up to 12 months ahead, depending on the risk type.[7]
The aim is the one this guide started with: a signal early enough to act on, specific enough to name a material, an origin or a supplier, and recorded with the date it was seen.
Horizon scanning in food safety: FAQ
Is horizon scanning the same as monitoring RASFF?
No. RASFF and similar alert systems report hazards that have already been found in products in trade. Horizon scanning includes those alerts, but it also looks earlier and wider: scientific findings, enforcement actions, regulatory proposals, local media and market signals that may become alerts later.
Does horizon scanning give early warning of food safety problems?
Only if it looks upstream of official notifications. A scan built on alerts, border rejections and recalls reports hazards that have already been found in trade. Early warning comes from signals that appear before that point, such as harvest losses, price spikes, local enforcement findings and new scientific results, matched to the materials and origins a business buys.
How often should a food business run a horizon scan?
Signals should be collected continuously, because the first report of an incident can appear on any day. The review is usually set on a fixed rhythm, such as weekly, with an immediate review when a signal touches a material, origin or supplier the business actually uses.
Who should own horizon scanning in a food company?
It usually sits with quality assurance or food safety, with input from procurement and regulatory affairs. Procurement knows origins and suppliers, regulatory affairs follows legal change, and quality assurance turns a signal into a test, a specification change or a supplier question.
Can horizon scanning help with food fraud?
Yes. Fraud is driven by price, shortage and opportunity, and those pressures often show up in market, harvest and enforcement information before a fraud is confirmed. A scan that tracks them can feed a vulnerability assessment with current information instead of last year’s.
How far ahead does horizon scanning look?
There is no fixed horizon. EFSA describes horizon scanning as the search for weak signals that are not yet found in current research or media, while FAO uses foresight for issues that could emerge in the medium to long term. For a food business, the useful range is early enough to act before an alert or recall.
Start with the materials you buy
We can show what a horizon scan of your own materials, origins and suppliers surfaced over the last weeks.
Talk to iComplaiWhere these claims come from
[1] UK Food Standards Scotland. Horizon scanning (citing the UK Government definition). foodstandards.gov.scot
[2] European Food Safety Authority. Emerging risks and horizon scanning (topic page, last reviewed 3 August 2026). efsa.europa.eu
[3] Food and Agriculture Organization of the United Nations. Food safety foresight. fao.org
[4] US Food and Drug Administration. Investigation of elevated lead and chromium levels: cinnamon applesauce pouches (November 2023). fda.gov
[5] European Commission. Ethylene oxide incident / sesame seeds. food.ec.europa.eu
[6] UK Food Standards Agency blog. Strategic surveillance and food signal prioritisation (6 September 2021). food.blog.gov.uk
[7] iComplai. Platform overview: data sources, daily signal volume and prediction horizon. icomplai.com
[8] iComplai. Food fraud risk prediction: the hazelnut harvest-loss case. icomplai.com