EPITalk: Behind the Paper

Saved by the Smell – Canine Olfaction as a Means for COVID-19 Detection

Annals of Epidemiology

Send us a text

Dr. Sebastian Meller explores the potential, and challenges, of using dogs’ olfactory capability to detect COVID-19 infection in humans. "Canine olfactory detection of SARS-CoV-2-infected humans—a systematic review” is published in the September 2023 issue (Vol. 85) of Annals of Epidemiology.

Read the full article here:
https://www.sciencedirect.com/science/article/pii/S104727972300087X

Episode Credits:

  • Executive Producer: Sabrina Debas
  • Technical Producer: Paula Burrows
  • Annals of Epidemiology is published by Elsevier.



Patrick Sullivan:

Hello, you're listening to EPITalk: Behind the Paper, a monthly podcast from the Annals of Epidemiology. I'm Patrick Sullivan, Editor-in-Chief of the journal, and in this series we take you behind the scenes of some of the latest epidemiologic research featured in our journal. Today, we're talking with Dr. Sebastian Meller about his article "Canine olfactory detection of SARS-CoV-2- infected humans-- a systematic review." You can find the full article online in the November 2023 issue of the journal at www. annalsofepidemiologyorg.

Patrick Sullivan:

Dr. Sebastian Meller is a postdoctoral researcher at the Department of Small Animal Medicine and Surgery at the University of Veterinary Medicine, Hanover, Germany. His research focuses on neuroscience, epileptology and behavior in the field of clinical veterinary research and translational medicine. Dr. Meller, thank you so much for joining us today. Yeah, thank you for having me. So you published this meta-analysis that pulls together a theme of sort of One Health, which is where veterinary medicine and human population health come together, and also SARS-CoV-2, which is still a really present issue in our public health. Can you talk a little bit about the purpose of the study that you did? What question were you trying to answer?

Sebastian Meller:

Yeah, so we had the first study or first systematic review, to bring together the many claims that were out there in the research community that dogs might detect the coronavirus, and so that was quite new. Of course it emerged right in the pandemic. I think that's logical, and so we wanted to see if the dog is really able to function as a diagnostic or as a screening tool and to see how a dog performs versus like established methods, like PCR techniques, which we considered the gold standard. And of course we had some criteria to exclude and include the studies, and we were also looking at that. There are enough, let's say, samples provided so positive, PCR positive and PCR negative samples, and then we tried to make a story out of it.

Patrick Sullivan:

Great. So what was the main finding or the key answer to your question, which is is there evidence that dogs do sort of a good job, or a sufficient job, at detecting SARS-CoV-2 infections?

Sebastian Meller:

Yeah, so, indeed, dogs are able to detect coronavirus.

Sebastian Meller:

That was like the main finding, but it is very important to mention that the studies and the study designs and the approaches were quite different, and I think this comes from the fact that we were quite in the beginning of the pandemic and a lot of research groups were just looking for something which could help quickly to somehow control the pandemics, and so there were a lot of working groups worldwide that suddenly came up with that idea and, yes, so in the end that research, or my research, showed that, yeah, indeed, the dog is a quite good screening or test system.

Sebastian Meller:

So we found that 80% of studies showed higher sensitivity than 80% and 90% of studies showed a specificity of higher than 90% and even low biased studies. So we had some low biased studies. They showed that, yeah, they showed a good performance of over 80% sensitivity and 90% specificity too. So that was quite nice to see that even the unbiased studies showed a quite good performance for the dogs. Yeah, but we also realized that, as I mentioned in the beginning, we had some issues how to standardize right, or we saw that a lot of studies had a lot of standardization issues. So that's also one important finding we need more standardization and certification processes for dogs to detect medical diseases.

Patrick Sullivan:

Two important issues that sort of came up in the article were one about bias in studies and which studies were more or less biased. How do you operationalize that? How do you decide, when you're doing this analysis, the bias of the studies and why is that important?

Sebastian Meller:

Yeah, so we used one tool that was the QUADAS-2 tool, which is created for systematic reviews and is created for the assessment of diagnostic tests. So that was one part that we used. So we looked at dogs as a diagnostic test in this respect. But of course dogs are not only diagnostic tests, they are also living beings. That's why we also use another system, but we can talk about that maybe later. But what in the beginning we used was the QUADAS-2 tool, which had us to see if studies are biased or not. So we were looking there at how was the patient selection, how was the index test, that means, the dog, how it was used or how dogs were deployed. We were looking at the reference standard, which was the PCR test, and also at the flow and timing domain. We also had a look concerning a more temporal dimension, concerning when tests are shown to dogs after they have been taken before from the people. So then we looked at the bias and also the applicability and we used this semi-quantitative tool to do the bias assessment in first place.

Patrick Sullivan:

So you have a set of things about the way the experiments were done or might be done that you think are important going in, and some of those will favor less bias studies and some will favor more bias studies. Yes, exactly. And then you have just yourself, or multiple people, assessing the individual articles?

Sebastian Meller:

Yes, I did the main work actually, so I assessed the articles, but I also had a colleague who did the cross check. Excellent, If we used the QUADAS-2 tool correctly and we had a high degree of correlation between our results. So we knew that the tool worked actually and we asked the right questions.

Patrick Sullivan:

I also just have to note that you get an award for, I think, the first article in Annals of Epidemiology ever to have smiley faces and frowny faces in your table.

Sebastian Meller:

So yeah, yeah, so that was quite unconventional.

Patrick Sullivan:

Unconventional, but it gives a little color to the article.

Sebastian Meller:

Yeah, it lives a little bit color exactly, and it was also in the beginning for me it was quite weird to use smiley faces because that's not the thing you use normally, but on the other side it was like a fresh new thing to use smiley faces.

Patrick Sullivan:

Yeah, and I mean it conveys what you want to convey, so that's good.

Sebastian Meller:

Exactly, and that's a point that's a semi-quantitative way of assessing and that's a good way to assess semi-quantitative data, which actually, for systematic review, is the right scale, let's say, to use for assessment. Yeah, so we also used another tool, but we can come to that in an instance.

Patrick Sullivan:

Yeah, we can go and talk about, because you did use two approaches. One was this QUADAS- 2 tool, but then there was also kind of a quantitative system that you use. Can you say a little bit more about that system and what it was used for?

Sebastian Meller:

Yes, first of all, we wanted to do a systematic review. So it was important for us to have a quality feature for the systematic review, which is, as I mentioned before, a semi-quantitative tool like QUADAS- 2. And normally there are some opinions about it and some rationales about it that you should not use score systems for systematic reviews, like scores that you can count, let's say. Well, we had a dilemma because then we thought about, ok, what to do? Because on the one side, the dog when we use QUADAS- 2, the dog, we see the dog as a diagnostic system. But the dog is not only a diagnostic system like other diagnostic systems, it's also living being. It's a living being who's working with its nose, with its olfactory tool.

Sebastian Meller:

And that's why the other idea was to use a quantitative tool, that score system, because this is a score system that assesses the quality of sense detection work. And then we decided to do both, to just put both together, to also stress that dogs are not devices, dogs are living beings. We wanted to show that dichotomy and that's why we use both tests, even if the score system is not adequate for systematic reviews. Anyhow, we thought that it is a big picture of the whole test system living being dog, and that's the rationale why we use both tests. But of course the score system should be considered a add-on to the actual systematic review. Diagnostic test quality assessing tool, QUADAS- 2.

Patrick Sullivan:

Yeah, and I think that's a really just interesting point for people that are thinking about. You know, we often get in the pattern of saying, like well, the tool that everybody uses, this, but is this or that. But you really sort of took into account the nature of the intervention or the nature of the sort of public health tool here and exactly, just added a second. You think about it as triangulation of all the issues that are involved. So, exactly, yeah, so can you talk about any of the limitations of your review or what you would? You could think of them as weaknesses or you could think of them as opportunities to do next steps. But so what's on answer and what do you think some of the limitations are?

Sebastian Meller:

Yeah, so one of the limitations I think that should be mentioned is I am also one of the researchers who did canine olfaction studies in SARS-CoV-2 detection and of course those studies also fell into the inclusion criteria. So that is one of the limitations, I would say. But I think it's really important to say that I anyhow looked at the guidelines that I made. I really stick to the guidelines and of course that's challenging but I stick to the guidelines.

Sebastian Meller:

I know the dog community, olfactory detection community, I know what groups are doing. I know that, not everything what is written in the papers. So of course what is written in the papers is, of course, the truth, but of course there are a lot of things between the lines that happen and I know those things. But anyway, I really referred only to the published information and I did not add any information that were known but not published. So it's really I really was really important to me to just look like, okay, just have rules and I stick to the rules and I just worked that through and so I think I did it well. I think we also had some statisticians in our team who also gave the confirmation of how I was working, that this was okay, anyhow, that's one of the limitations, of course, so I was also part of it, but it was important also to us to just, you know, to state, to make a statement, also to assess scientifically if this is really going into the right direction or not, the whole thing with the detection dogs.

Patrick Sullivan:

I think this is a general issue which is, in a lot of research communities tend to have a fairly small number of investigators and we know each other, we see each other at conferences, and so I think that the important thing is acknowledging that that could be a source of bias, you know, and saying I'm going to set up these tools in very objective ways, so that that will be my touch point. So I really appreciate you raising that issue, because it's one that we don't really talk about that often, but especially in a smaller, more sort of focused kind of research community, I think.

Sebastian Meller:

Yeah exactly, so we have quite a small research community. That's that's one of the points that are important to stress, that, yes, so we are also most interested in our subjects, right? So of course, we've right also the we have the knowledge to know what is really important, what really matters in this research field. So that's why, yeah, those things happen, yeah, but it's important to mention. Yeah, of course.

Patrick Sullivan:

Yeah, it's important. It's just important to acknowledge and be systematic. That's all, yeah. So are you aware of programs that are using dogs for COVID screening currently?

Sebastian Meller:

So actually, yes, there are some programs, or there were some programs. For example, in Dubai they use dogs at the airport, and even it was not only in a scientific way, so they really use the dogs as an established tool. Let's say we also did something quite interesting, I think, but that was a study, but a field study. So we did a concert, we organized concerts here in Germany with a huge concert company and we organized concerts just to check if dogs could be used in the entrance of a concert or of a venue, to see if dogs could, for example, screen people in a real life scenario, and that worked quite well. So there are, let's say, the way it's paved to use dogs in those in this ways. But anyhow, we need more infrastructure, the better infrastructure, training infrastructure, which is similar, like in the explosive detection or drug detection dogs. So that's something that has to still be established.

Patrick Sullivan:

Yeah, you mentioned this idea that standardization of the training and certification of the dogs would be important. Next step so you've sort of shown a scientific principle here and you've shown that it meets one of our criteria around the robustness, which is that it plays out in multiple settings and or studies done in different places. So you sort of build that piece. But then really to move into program, you mentioned standardization of training and certification of dogs. So just building, I guess it would build on ideas like the explosives detections, but what does it look like to put those in place and are there any efforts underway to do that, either for the future of this pandemic, this ongoing pandemic, or for future public health challenges?

Sebastian Meller:

So I need to admit, we have been talking a lot with politicians here in Europe and I need to admit, well, there were some some politicians and some some administrative structures which also supported us to do such studies, but somehow, you know, there was not a point in which there was a decision to establish this infrastructure for training, for example, of dogs, of medical detection dogs. I think it's still quite a. It's difficult to say, but I think people maybe do not trust that much dog. Maybe then they trust a test which was artificially produced and not produced by biology, you know, but by humans. So they do trust more of those things.

Sebastian Meller:

So we don't it did not have any opportunities to establish this outside, let's say, outside the scientific space, but anyhow, anyway, the scientific space is just a lot of things are happening right now, even concerning, for example, lung COVID or post COVID syndrome. We are also doing some research and going for that, going more in this direction. Something happened in the scientific space, so I'm quite interested, but the first goal we had this did not work out to use really dogs to help to break through the infection chains of the pandemic.

Patrick Sullivan:

I mean there may be work that can continue to be done, more on a behavioral science side, maybe about the acceptability of this, because I think people don't bat an eye. I mean, it was an airport earlier this week and somebody came through with a dog and like people didn't even look up from what they were reading. The dog's just checking out the bags and it's obviously for that sort of explosives purpose. So I think, trying to understand, maybe through qualitative methods or through surveys, what people's concerns are, and if it's a useful tool, then I think it's a separate question to ask how we introduce it. But clearly, in the case of airport safety and explosives detection and I think drug detection maybe in like customs and immigration, this is acceptable and people just sort of do their thing and the dogs do their thing, and it's not a big deal. So it may just be a little bit of acclimation Exactly Great.

Patrick Sullivan:

So I want to turn now to what we call behind the paper, and so I think when we talk to colleagues, at whatever stage of career, the questions are always like how do you come up with ideas, how do you overcome the things that are challenging about doing these and how do you work with colleagues and all those pieces that aren't part of the research question but they're important about how we get work done. So I wonder you sort of alluded to some of them but you see, you have this idea and you can get the data and it sounds like you're in a network. But what was the biggest challenge, just sort of, in getting started software, people's involvement, skepticism of colleagues, like what came up, that you sort of had to get in place before you could move ahead and do the work.

Sebastian Meller:

So I think the biggest challenge for me was the reading of the papers, but not just reading, just reading again and again, and again and again and again and reading between the lines, and I really was keen on not missing any detail. And that was really, really difficult, because I don't know if you have ever seen a supplementary table that is so huge. So if I have put it into the main manuscript, let's say, then I think it would have exploded. But really I really looked at each detail and that was actually the thing.

Patrick Sullivan:

That was like, yeah, but you did for transparency, though. Then I think the supplementary table ends up being an online appendix, like so people want to go and see that yes so it did come in, I didn't see that.

Patrick Sullivan:

I think that I do think, sometimes, figuring out in the systematic reviews that I've been involved in, it's also a case that you start out with a set of things that you're interested in and then the first six or eight papers that you read, you realize like, oh, like this thing is also of interest and maybe I should.

Sebastian Meller:

So it's a little bit of an iterative process, exactly, and then you start again to read it again, but at the same time to stick to your guidelines, and that's really a yeah, it may be quite intense, just to stick to your guidelines, but then to realize, oh, there's something else which could be really important. Ok, I need to start again and to read it under a different light again. Exactly, this machinery was quite labor-intense, but the other things around me, like software or colleagues or something that was quite so they let me work. We worked together and we had meetings and that worked quite well. So there was no problem. It's great. Just about the material itself.

Patrick Sullivan:

I have to ask, as someone who started not in research training but in clinical training in veterinary medicine myself and people always ask me why did you go to veterinary school and then do epidemiology work and people? So really a lot of your work is related to dogs. So what inspired you or what was your path to get in to be in a college of veterinary medicine and to do this work around with dogs in terms of the direction of your research career, yeah, so after or during my veterinary studies, I realized that I wanted to research.

Sebastian Meller:

Somehow something happened in my head that I realized OK, I am quite more interested in what, the background, what is happening in the background, let's say. So that was one way, and I started with preclinical studies and then came back to the clinic, let's say, and now I'm a postdoc at the interface of clinic and basic research. So it's really an interesting field, especially in dogs, because dogs are great. So we are living with dogs together for thousands of years and there are great models for diseases for humans, but not only this. They are like our, they are living with us, they are our family members and that's really interesting to also have a look on them, to also research them, to bring the importance of those living beings also more into the foreground.

Sebastian Meller:

I can tell you something I had one inspiring meeting in my high school time. There was an elderly lady that was a good friend of my father they had, they were in a group, a photography group, let's say, and that was their hobby and they were like traveling around Europe from cities and making pictures of cultural and of historical stuff, and then they were meeting once a week and just discussing pictures in a group. So that was quite a very creative space and I have been there sometimes, and as a little child. Then when I grew older, I of course was not there that much. But she asked then somehow my father, what about my exams for high school, how is it going? And my father told her that I also do a part of my exams in French because I lived at the French-German border and she was French, she was a translator in French and German, and so she asked my father if I would accept her to come once a week to our home just to discuss about subjects, reading newspapers, discuss about the different things in the newspapers, and so that was a really inspiring space for me and she was really inspiring. And suddenly she asked me what would she do after your high school? And I was telling her I want to be a veterinarian, and then she did not say oh, ok, interesting, so you will be a clinician, interesting, what will you do there? No, she was.

Sebastian Meller:

Her first sentence was wow, there is so much to discover in veterinary medicine. And I did not think about discovering something in veterinary medicine. Of course you can. You always have to discover as you grow in your work you discover new things. That's important. But I did not think about the research, research aspect of discovering something that there is so much to discover, and that was like a little sparkle from a very inspiring person to me. That became, yeah, became a flame, became a fire, and then I think that had also a big, great role for me. You know meetings with, for example, elderly, wise people. So that's an advice to the young folks put away your phone and talk to the people around you. It's quite important, I think, and sometimes, when it's getting hard, I sometimes think of her and think of her sentence that there's so much to discover and that's something that is showing me the big picture. So that's, that's the energy source, I think, sometimes. So that was really inspiring. Maybe this is too long, just a small sorry.

Patrick Sullivan:

No, I think this is.

Patrick Sullivan:

I think this is absolutely perfect because I think how we get to where we are, two things One, how we get to where we are, and sharing that is a generous thing, because there will be some people who are, you know, well earlier in their careers.

Patrick Sullivan:

And I think what you're sort of saying is that there's a value in sort of in talking to people like whether they're in our field or not about what you do, of explaining what you do and being inspired by, you know, by other people's thoughts. And I think in these, in these podcasts, we sort of ask about this very thing, which is like you know, how did you end up getting here? And I think these kinds of events one are nice to acknowledge Exactly. I hope if she's a podcast listener, you know that you can share it onwards, but I think it's inspiring for people just to listen to those things that happen in our lives that feel important, because they often are important and lead us to important places. So who or what do you think has been the biggest influence on your professional work and path?

Sebastian Meller:

Yes, so for my current work, I think that head of department here Holger is his name he has, I think, one of the biggest influence on the work. So he's an incredibly creative and really positive guy and a real out of the box thinker and, yeah, he makes actually you to like to leave your comfort zone and I think this is one of the most important things to have the energy for changing something and to change something to the better. And in the first moment one thinks when he has a new idea that it's crazy, but then suddenly everything makes sense. So I don't know how he's doing that, but he has quite talented concerning this.

Patrick Sullivan:

What a nice acknowledgement of his role. And you know, my old career person perspective is as you move on, think about what that meant to you and how you can create that kind of environment when you get in that position. Along the same lines, if you could think back to yourself you talked about like an earlier stage in your education and when you were a younger person, if you could talk to that younger person and give your younger self a piece of advice, what would it be?

Sebastian Meller:

So I think, well, maybe don't worry too much. That's one of an advice I would tell to myself. I would maybe also tell myself trust the process. So, whatever, whatever process it is, just give it more trust. And I also think that, well, I would maybe also say that if it feels not comfortable, it is the right way to investigate this feeling and not to run away from it. And I think now, right now I am there that you know, if it just does not feel like 100% right, just don't turn away, just investigate it. Maybe there's an opportunity, you know. That's, I think, important advice.

Patrick Sullivan:

So that brings us to the end of our episode. We covered a very broad span of topics and I'm so grateful for you, Dr Miller, for joining us today. It was a real pleasure to have you and hopefully we'll see some more of your work in the published literature and get a chance to talk again. Thank you so much.

Sebastian Meller:

Thank you so it was a pleasure to me. Thank you.

Patrick Sullivan:

I'm your host, Patrick Sullivan. Thanks for tuning in to this episode and see you next time on Epi Talk Brought to you by Annals of Epidemiology, the official journal of the American College of Epidemiology. For a transcript of this podcast or to read the article featured in this episode and more from the journal, you can visit us online at wwwannalsofepidemiologyorg.

People on this episode