Hello everyone, and thanks for joining today's webinar. It's just a bit of introduction about the agenda of today and some instruction on how to manage the panel. You will find in the application a panel for the URL. We also have given a link to provide a survey with some questions. If you want to answer them, they will be anonymous, so we'll be happy to listen to your feedback at the end of the webinar. Also, during the webinar, at any time, you can open that panel. There is also in the application panel a Q&A section where you can submit questions. Me and my team will reply as soon as possible about those questions, but most likely we'll go offline because there's a lot of participants and attendees today, and we probably don't have time to answer all the questions.
So today we have plenty of content to share. Together with me, we have a friend from Nestlé, so we're going to see first a retrospective of 2020 food safety data from my side, and then we'll go through a Q&A with Nestlé. Let's introduce, in fact, today's speakers. First of all, let me thank Paolo for joining. I really appreciate your effort for being here, contributing to today's webinar, which I hope will be an interesting one for all the audience. Would you like to spend a couple of words about yourself and introduce it?
Sure. Thank you, Nicola. First of all, really thank you for having me here. Good morning, everybody. Good evening, depending on where you see this. I'm working at Nestlé since quite some time now. I'm a senior specialist in food safety and quality and working in the digital food safety department. I'm really excited to be here with you today.
Thank you, Paolo. And about myself, I'm Nicola Colombo, the global head of SGS DigiComply. I've been working in the digital industry for many years and bringing digital solutions for regulatory compliance and food safety. Okay, so I think, Paolo, you're now going to have to listen to myself just for a bit. Basically, my idea was to set the ground for this discussion, and then based on what I share, also asking you a few questions that would give us a glimpse on how Nestlé is supporting and operating within the emerging risk process. So what we're going to see is a retrospective on 2020 food safety data. I'll try to be fast, keeping more time for the interactive conversation with you in the latter part of this meeting.
What I'm going to show in the next 15 minutes are basically the data that we are constantly gathering from the food safety intelligence platform we have to present to everyone an eye-level picture of the food safety issues during 2020. Just, let's say, an overview. The sources of this data are basically public, and they come from authority sites, databases, scientific literatures on which we collect, transform, and enrich this data so that we can provide a continuous intelligence to understand and monitor emerging risk. First, one of the most and foremost facts is that we were and we are still, unfortunately, in the pandemic. One of the first panels I want to share, or the first slide I want to share, is more about quantitative.
That is already from comparing 2019 versus 2020 in terms of quantity of incidents and food safety issues. We can see that in 2020 we have less. But obviously, looking only at the number, it doesn't say much. Okay, we can say, "Okay, interesting." But it's important also to try to find why this is happening. But this basically represents some numbers from about more than 100, I think 107 sources of data that report food safety alerts, recalls, outbreaks, or let's say border controls and the like. What we can see, in fact, as I said, is that during the hard lockdown, you remember that most of the Western countries, after China, they went in hard lockdown in March. We can immediately see that within the March timeframe, we see a drastic drop down of the incidents reported, and then we see a ramping up back from July onwards.
But actually, also we see a trend that even steadily increased even toward the end of the year. So what we see for sure in Q2, we can correlate the reduced number of records to a reduced also inspection activities within the authorities because they had to cope with the pandemic emergency. There were certainly other priorities. Nevertheless, for example, just making sure that respiratory and personal protection equipments were coming into every market. So that was just one of the crises of last year. And also foremost, what was also important, we remember that in that time we were actually seeing images of people queuing at the supermarket shelves completely empty.
So one of the main probably issues and emergencies on the table of the authorities were definitely supporting the supply chain that did not get too much disrupted or that at least the primary goods were reaching the consumers. Although just a more funny note, I remember an image of a supermarket where the toilet paper was actually completely run out across the supermarket. So in general, the compulsive purchase instinct of the consumer really showed some weird cases last year. But despite these funny facts impacting the food safety, that's what we can observe. Just as RASFF, in fact, in that time, RASFF alone, which is one of the sources, reported fewer notifications in 2020 than 2019.
But if we move along to the curve, we can see that later in the year, in the latter part of the year, we have seen a restored capacity of inspection, and also what we have observed was new and immediate policies and guidelines from enforcement authorities to protect local supply chains. In that period, we observed cases like some export ban of wheat from Eastern countries, or we have also seen amendment to standards like in France on some cheeses or some specific origin products that were changed or maybe even raised the barrier to protect more the supply chain to basically use more local produce or eventually to loosen up this to allow, let's say, or to avoid disruption in the supply chain, so this is a correlation that we have seen last year by observing the data.
Again, I'm still looking more at the numerical facts and then see what the correlation can be. Just by giving you to give a bit more deeper insight on what are the risk and the incident we can observe last year by product categories where we see a general decrease, although an increase on beverages, especially alcoholic beverages, where we've seen an increase on frauds. That's probably also because of the fact that there was different consumption ways, more online purchase, and so on. Meat and meat products see a decrease, but actually there was a strong increase on poultry-related products. That's where we've seen increases. But overall, yes, that's basically represent the overall status by type of products. I don't go into details of everyone, but certainly that's what we have seen.
Just about the most common hazard that everybody, or most everybody in the food safety knows, we have seen less reported incident cases of Salmonella. We also saw pesticide residues to be less. I mean, basically there has been, again, less incidents, but that means because less tests, less inspections performed truly as a general trend, not just in one country, but everywhere, despite the fact that, for example, on Q4 alone, compared to Q4 the year before, pesticide residues increased by 50% almost. So that means that there was a very strong push for the authorities to make sure that the consumer's safety is restored. That's just to give some numbers and also some other type of hazard that we have observed. Mislabeled products, yeah, we have a certain amount of mentions, but more on the numbers, we see a decrease.
Just to say that over the last years, mislabeled products, especially missing allergens or missing declared allergens, it's something that was always high, normally count up to 20% of the overall recall worldwide. But in fact, also there was a very strong decline. That doesn't mean that there is more safety on the labels and on the information shared by consumers, but probably, again, that's a bit less enforcement. And also what we have seen and we have observed, it's also related to the fact that there was also more online shopping that typically were the information to consumer. It's an area that presents a bit more gray areas because consumers might not really foresee and get the same level of information when they buy online versus when they can really read the label printed on the packaging. Also, migration from food contact materials shows a decrease significant.
So I'm just reporting the main facts, the type of hazard where we've seen the most relevant changes or facts in there. And between parentheses, you can see the decrease in terms of numbers from the previous year. Some relevant issues that I'd like to share with you, we have seen ethylene oxide spike at the end of the year, especially in Europe. That was pretty much the category was there was a presence of this unauthorized substance. In fact, in this period, ethylene oxide used as a biocide was often detected in high concentration. But again, this happened when the enforcement was restored after, let's say, the first wave of the COVID. So that doesn't mean that this issue was not around before. So probably with the same level of enforcement, we could have had a similar curve for this particular issue.
But in fact, that's what we have observed, and we started to see a decline on these products because there is more testing going on this. The source of these contaminants was mostly India, mostly traded in the EU, and using various products such as sesame oil or sesame seeds and so on and so forth. This is worth to mention, again, quantitative risk. We also see some other known issues like mineral oil. It's still currently predominant in the European Union, despite still the number of incidents, it's still very low. What we have observed in terms of analyzing different types of data, including scientific paper, test methods, and so on, we have seen that there is not yet a—I mean, we can see also, we as SGS, we see an increased demand for testing.
So we see more and more clients that are in need for testing of this contaminant. But there is not yet a fully agreed standard for testing. For example, in Europe, JRC has been working with the labs on a specific test method. So I think once that would be a more standardized testing method generally available, we won't see this high number of incidents because of this. There's not so much testing going on. So we would possibly see an increase over time as soon as these standardized testing methods would be approached. Let me say that's interesting also to talk about E- coli outbreaks. Just to mention the frequency of mention also around social media and not only official sources, we have observed in April a peak of mentions because there was the sprouts outbreak in the U.S.
But also interesting to see a spike in mentions of these issues at the end of the year where there was an outbreak in the U.S. But it's normally there is a general connection with a specific food. But one of the outbreaks of E. coli of 2020, despite the deep investigation of CDC and also how in general FDA has enforced and has tools to track and to monitor food safety, there was no specific food item that could be tracked back to the infection of E- coli outbreak at the end of the year. So that's just to say that this we see, we observe that this has tracked more in the conversation we have seen about this outbreak. We have observed also correlation with traceability, transparency, and need for more immediate and faster actions to trace and to track the food safety issues.
That's just to say that we always try to also find a correlation between this data, I mean, what we see, the signals that we observe. And that's one of the most challenging parts of data and technology. It's not just giving an inventory on the numbers of incidents, which is what I've done in the last few slides. But what is also important is how to correlate the signals. I've shown quite a bit of numbers of occurrences. It's like an inventory of the risk and the incident, which is important to a certain extent. But what is also important is to capture and connect different signals. So how can we derive certain conclusions as the ones I shared? There was a drop. Why there was a drop in that number? So this is something. Those are kind of hypotheses that you can also test using semantic technology.
You can test the conclusion. Was there any measure implemented to protect local supply chain? Having a knowledge, millions of documents that you can analyze like we do, similarly to also what we are getting used to with Google. We Google a question and we try to find an answer. But with specific knowledge around food safety, that's also what can be done in a way that we can correlate and explore new assumptions and conclusions exploring this big amount of information. But again, exploring a conclusion, that means that you still have a clue. I see the number of incidents dropping, and I can ask a question like, was there any new measure implemented to protect supply chain? Or how many was the number of inspections dropped in 2020? If there is a document that talks about this, you can certainly find it.
But again, everything starts from something that you know or at least you try to explore because you have a clue to explore. That's one way of gathering a signal. Already useful, already super powerful if you have this technology on hand because you can explore in a second. And that's also what Google has empowered us with an immense knowledge available for everyone. Although Google maybe is not that specialized into knowledge that it's related to food safety and the language of food safety. But what could be the next step? The next step, from a technology standpoint, what can we do to discover emerging trends? And here I go a little bit more technological, let's say, in the talk. What we have done, it's using word embeddings to basically mathematically calculate the distance between keywords or sentences or tokens or even documents.
To make it simple, let's think about these three examples. France and Italy are quite similar in the sense that, by the way, they are actually adjacent countries. They are actually sharing the border. But if we look at it from a mathematical standpoint, you see that they are quite similar. They have a close gradient. Bull and crocodile are not similar. So you see how mathematically they are represented. They are very distant. In terms of France, Paris, and Rome, Italy, the two vectors that represent these two keywords are similar, but they are actually opposite in their code. That means that they are basically far away from that, but it's close to 180, sort of different cities. So how can we apply this principle to a large corpus of data unstructured related to food safety, science, incidents, news articles, consumer journals, blogs? Imagine a huge amount of data.
That's basically allowed with this technique of word embeddings. You can basically create these maps of, let's say, vicinity at scale because machines can do at scale. Nobody can read all of these documents even in the entire lifetime, and by creating this map, what we can do, it's basically by creating in this image, there's basically billions of data points which represent all these keywords. Inside here, there might be 5,000 pesticides, 3,000 different types of products that they are actually organized using this vectorial space representation in a multidimensional way, so it's really billions of data points that can be generated. This is a snapshot that represents a certain stage of the knowledge at a certain time, but what we can start to use these maps to do, it's to find new correlations.
So for example, if within 2020, what we have observed (and here, I can go back one second) we have observed that certain keywords that were more isolated, they get closer to a certain cluster. For example, COVID and health food and transparent labeling and these types of keywords started to get closer into this group. What this happened, basically, it triggers the discovery that there are new correlations. Just by analyzing these keywords, we were able (and that's also where we published a blog article recently) to detect a certain trend that basically gave us the capability to understand the correlation between the pandemic and the urge of policymakers to encourage consumers toward healthier choices. The output of this was also new policies and enforcement of new regulations and guidelines.
Just to mention a few, I mean, with the new U.K. being in the Brexit, they just published a regulation that basically banned the sales of high sugar and high fat products to be sold near the cashier or the point of sale. So the effect of the pandemic among this huge amount of data gave us this correlation thanks to this technique of keywords embedding where we can basically monitor the moving of certain distances of keywords, obviously at scale, and identify new patterns to recognize new potential issues. A similar concept can, for example, be applied to discover that in a certain product, there is a new contaminant.
Let me say that if normally we can observe in rice arsenic and probably let me just say something that doesn't make much sense, but let's say that we start to observe that there is a new contaminant that can be mineral oil, let's say MOSH in rice. This by becoming closer, these two terms, rice and mineral oil, they start to get closer. This movement of distance between these two terms triggers an alert and a discovery of a new fact. That's basically my first part. As I said, I wanted to be faster to finally leave Paolo that has been very patient waiting for his part. So I have some nice questions on this topic to ask Paolo. So thanks again for your patience. And okay, so we can basically get started with your part.
So first of all, I think Paolo, you are part of the Nestlé Institute of Food Safety and Analytical Sciences. I just want to maybe give a few words to understand what is the mission of this institute within Nestlé, which I think it's a center of excellence, I think, especially for me being in Switzerland. It's something that I always been seen as very important.
Thank you, Nicola. And first of all, I mean, I was waiting for this moment, but I was very interested in what you just said. I mean, very interesting speech and very interesting overview of what happened in 2020. Thank you very much. So yes, I mean, no surprise. I mean, food safety is the number one priority for us.
The Nestlé Institute of Food Safety and Analytical Sciences is a long name, helps indeed Nestlé to stay a step ahead by ensuring that the safety of our products is never compromised. The institute, our institute, the Food Safety and Analytical Institute, is part of the fundamental research entity that exists in Switzerland, as you just mentioned, in Lausanne, to be precise. In collaboration with our global R&D network, research institute, other research institute, and governmental agency, we focus on four main areas. The one is really to develop powerful analytical methods which ensure that our products are free from contaminants. The second one is to optimize our process to ensure that our products meet the highest quality and safety standards. The third one is to support product innovation with scientific approach, advanced analytical tools, and latest detection technologies.
And last but not least, we also explore new digital tools that help to drive food safety. I'd like here really to emphasize that this is really part of a bigger machine, if you want, because all the methods that we develop in the institute are actually implemented worldwide in the Nestlé factory. And just to mention a few figures, each year, we use these methods and tools to conduct over four million analyses in our quality analytical assurance center. And this effort is really to help to ensure that the safety and quality of all our products worldwide.
Wow, impressive. And especially, I want to connect with your fourth point about digital tools. I've been sharing this emerging risk, and at the end, it's all about this huge amount of information. So how relevant is for Nestlé the emerging risk management process?
I mean, it's something that you have on the forefront. Absolutely. I mean, again, I mean, safety and the quality of our product is the main priority. And as you can imagine, such a big company like Nestlé that has to deal with a huge variety of products for this kind of company is super important to be able to monitor all emerging food safety risks because this is really an essential part for our business to make sure that we can adapt and address according to the signal that we identify. Our organization has dedicated teams that are working on what we call early warning, and these are distributed across the world and with specific skills that are able to identify specific topics and work on specific subjects such as contaminant, fraud, and regulatory.
Very nice. Very nice. And where I mean, so you mentioned basically signals. I mean, is there a specific place where you look for those relevant signals? Or as I was also mentioning, it's basically you need to dig into this sea or huge big data lake or let's call it ocean of data to find the signals.
Absolutely. You mentioned it. I think your slides before, I mean, was very nice and was helped really to figure exactly the size of the problem. We have to go through any type of source. It's important that we have to consistent observation of input from what our authority is doing, the science, so scientific publication, but also the media. And basically, we collect and analyze data along all the value chain from Farm to Fork.
These inputs that we receive are actually accessed by our expert who may identify potential emerging issues and basically connect the dots so that we can see and identify potential risk for our products. That's why, as I mentioned before, we do have an early warning system that helps us to really pinpoint the signals, the relevant signals that may develop into issues. And of course, the earlier we can spot these potential issues, the better we can prevent and manage them.
Good. Good. And basically, in this context, you mentioned an early warning system, but how can you say for Nestlé the relevancy of the technology? I mean, you mentioned teams and people that are definitely bringing and carry, I guess, an immense knowledge, but how relevant is also the technology for you to succeed with these challenges?
I mean, the technology, especially the new one, the digital solution, are really like a fundamental companion of any human decision. Again, you have seen, you have mentioned the volume of data information that we have to process and to understand. And of course, no expert can do it with this manually. That's why we are looking with more and more interest to these new technologies such as big data, machine learning, and AI because, yes, the world is changing rapidly and the data, so we have to continue to stay abreast of this changing environment and leverage indeed the latest digital solution that exists.
Yeah. And I think it's a big challenge.
And somehow from our, this is more, let's say, our point of view, it's important also to find the proper alliances to achieve because the research and the amount of resources needed to overcome these challenges is big. Good. But among many topics, I think talking more about not just, let's say, the technology side for managing emerging risk, but on a more specific topic, we heard a lot on the news that Nestlé is committing toward recycling packaging. And I guess this is also bringing potentially some new risk into the zone because there's going to be some new materials, new things that you don't know. So what Nestlé is doing to ensure that the recycled materials are safe. Now you are basically on the race to achieve the goal, right?
Yeah, absolutely.
I mean, packaging is one of the biggest commitments Nestlé has done lately and definitely one of the biggest challenges the world has today. And as you might be aware and you just mentioned, by 2020, 100% of our packaging will be recyclable or reusable. And to achieve this, we have to evaluate new packaging material and also work with material that comes from recycled content. And of course, these materials need attention from a food safety perspective because, as you mentioned, are new. We are not that we don't have all that history of working with these. So there is a strong collaboration within our food safety expert and the expert that we have in another institute, again, in Lausanne, the Nestlé Institute of Packaging Sciences.
And we work together with them to ensure that our alternative packaging solutions do not compromise the safety and the quality of our food. We are developing new tools, of course, to assess these new packaging so that we can ensure that they are safe for food contact application. And such a tool includes, again, powerful analytical methods to detect potential contaminants at very low levels directly in the food. We develop several bioassays that allow us to assess the potential biological effect of packaging material. And also, we are creating food-grade recycled plastics and boosting packaging innovation. So with this approach, with this approach to food safety, we ensure that the safety of the packaging is looked after from the very beginning.
Good. Good. Just a clarification. By when you say that Nestlé will go on 100% recycled material?
2025. That's the commitment that our CEO makes.
That's really short-term and interesting. What definitely appears is that you're still in R&D concept. Your food safety culture, it's really going hand in hand within R&D. It's really planting the roots into the R&D and then goes down to the manufacturing. Let me just ask the last question. How Nestlé builds these R&D capabilities to ensure food safety?
Okay. Absolutely. I mean, R&D is very much involved in the food safety, and it starts there, but of course, we embrace food safety across the whole organization, so operation, factory, markets, and so on, and R&D is part of the Nestlé organization. That's why, I mean, the food safety includes, of course, also R&D expert on ingredients, regulatory, chemical, and environmental contaminants, fraud, microbiology, you name it. We have an expert really in every little field of food safety.
Basically capitalizing on their expertise already at the very beginning of every project. That's why we start. We have a project that has quality and safety by design. We can assess the ingredient, the applied technology, and the condition that has to be used to produce our product so that we can ensure the finished product is safe and compliant. In summary, if you want, as an organization, we connect the dots. We anticipate how different factors interact, and we ensure that all the regulatory requirements.
Great. Great. Thank you so much. This is really a nice insight. I think, to all the audience, obviously, a symposium like this sitting on a face-to-face meeting would have been certainly nicer, but we have to stick with these digital interactions. Time-wise, I know Paolo, you have to leave us shortly.
So I just want to mention the audience that we still have a few minutes for Q&A that I will let Paolo go. But before we do that, again, there is a survey link into the webinar panel for the web links. If you want to play that, it will definitely be appreciated. And you can submit now your questions so that I can answer. Most likely, you can probably address most of the questions to myself because, in fact, for time constraint, I need to Paolo go. So thanks, Paolo, again, for joining. It was really nice talking with you, and I hope we'll have other opportunities in the future to discuss around this topic.
With pleasure, Nicola. I really thank you for having me here and giving me this opportunity. It was super interesting, and yeah, we'll stay in contact.
Thank you. Stay safe.
Bye.
Bye-bye.
All right.
Thank you, everyone. So we have let Paolo go for his next meeting. So it was difficult to grasp him from his many priorities he had. There was really a lot of questions. I think we've been quite overwhelmed. So there's many of those. I maybe just really use a couple of minutes to mention a couple of cases. I just picked them up randomly and read as I speak. There was an interesting comment about Salmonella cases from Sagit in 2020 where we said I mentioned basically less inspection. Yes, that could be a root cause. But he says his question is, wouldn't be then an effect to see more illness? Well, yeah, it's probably a good clue that less inspection might have led with more potential illnesses.
But I think my first reaction to that would be difficult to find the data, especially last year when the health system was so much under pressure because of the pandemic emergency. But good point. I mean, it would be good to incorporate the insight also on that piece of data. Another question that I just grasped by scrolling the hundreds we got so far, and we'll try to get back to all of you. It's about that we mentioned food safety risk signals. So besides official incidents from authorities, what do you consider a signal for a potential risk?
I can actually talk, let's say, what is our expertise. Maybe also, I'm sure maybe Paolo would have a slightly different or maybe certainly has his input.
But basically, what we use, as I mentioned, to detect signals are very broad datasets, not only including the official records like RASFF or FDA incidents, but also social media. We use general news and media. We use also other data factors and correlate with even raw materials caused to commodities. And so we correlate different data points to identify potential risk, but then still requires these threads to be then analyzed and discussed. I'm now pushing up on the screen the result of the survey that we have online that everybody can do. And thank you for all the people that have already answered. I just captured, and thanks, Jennifer, my colleague. She pointed out to these. One of the questions is, if you currently perform a systematic scanning of many sources to perform emerging risk, right now we still get half and half.
Interesting to see that probably that's also due to the fact that it really requires, especially if you want to do it manually, it really requires an intense effort. So nobody can really read hundreds of documents, hundreds of sources to stay up to date. No one can really perform this because otherwise we become an obstacle instead of an opportunity. So in fact, it's an area that becomes valuable if technology can perform this in a way that can distill the information and reduce the burden on each food professional. So good insight from here. So I think we are heading toward the end for today. I hope you enjoyed the session. I was really proud to have Nestlé Paolo with us to share his feedback on this.
We'll certainly organize future events like this one, maybe a bit more interactive as now. Again, the technology allows us to make a bit more open to questions and to interactions, this type of virtual events. So thanks, everyone, for joining today. I wish you a good rest of the day, whether you are in the morning or in the evening, depending on where you are watching this webinar. So thanks again, and talk to you soon. Bye.