In This Episode
As opposed to being reactive to our pets developing issues, what if we were proactive in looking for issues they might develop, so that we could help them before disease or illness even began to present itself? With twenty years worth of data provided by one of the world’s biggest veterinary companies, combined with the latest developments in artificial intelligence and machine learning, this week’s guest has had a hand in doing just that.
This week on the Veterinary Innovation Podcast, Shawn and Ivan speak with Dr. Jennifer Ogeer, the Vice President of Medical Affairs at Antech, about RenalTech, the company’s new predictive diagnostic tool, the AI modelling that serves as the backbone of the application, and the importance of building emotional intelligence.
- AI Modelling
- Collaborative Intelligence
Shawn: You are listening to the Veterinary Innovation Podcast. My name is Shawn Wilkie and along with my awesome co-host we interview the innovators in the veterinary space every week. I want to go in and introduce today’s guest.
Meet our guest – Dr. Jennifer Ogeer of Antech
Ivan: Hey, my name is Ivan Zak, and I’m happy to introduce a colleague of mine, Dr. Jennifer. We’re gonna talk about the artificial intelligence in diagnostic space. But I would like to start with introducing Jennifer and her long-lasting business career, which is innovation on itself, I think. So Jennifer’s long-lasting business career started as a DVM at the university of Guelph. She also had a residency and master’s degree in cripple care medicine . She had an exact MBA and master’s in organizational behaviour and leadership and all at university of Guelph. Currently, Jennifer is the vice president of medical affairs in Antech diagnostics, and we met in person and as colleagues at IDAXX, where Jennifer was medical affairs marketing manager. She also is a vice chair of boarder directors “Veterinarians without borders” of Canada, and finance chair of directors of VC? head trust.
Wooh, what a career. On a personal side, she has 7 pets, 5 cats and 2 dogs, and they are all rescuees? And she enjoys athletics such as cycling, running, wheel training, and mixed marshall arts. You are an innovation as a career. So can you maybe introduce us to how did you role in a such short period of time to all of these educations into one direction and how did you kind of did it? along the way?
Jen: Thank you, Ivan. It’s truly been, I think a lifelong pursuit to be able to think about how as of veterinarian I can help people and their pets. And also how I can contribute back to community and so in my journey of my education I have sort of ? opportunities to be able to not only as a doctor and a medically oriented person, be able to incorporate the elements of business intelligence, as well as, you know, what I call a high level of emotional intelligence. And I think particularly in the area of emergency medicine and critical care where many times we’re dealing with patients with life threatening conditions and pet owners who oftentimes are in severe distress, and even in tragic situations that they themselves are dealing with personally, not even just with a pets. It is an important area where we have to understand that, you know, it’s emotions, and psychology and well-being of all pet owners, as well as sort of the individuals we teach in academia, and how we actually can sustain ourselves as veterinary care workers and nurture. And so I think my journey for my education has been a very cohesive path around sort of building that intelligence and our emotional level as well as an educational level, and being thoughtful as a leader and innovating in a path that I’m gonna talk about today, which’s diagnostics and imaging.
Ivan: Well, it certainly impressive. You know, sometimes, sometimes I’m thinking about people getting motivated to do certain things, and then I’m trying to finish my MBA right now, and I’m kind of procrastinating on dissertation, so that’s definitely puts me to shame, and motivates me to finish it. So I appreciate that. So two diagnostic giants, IDEXX and Antech, and you know, you just worked at IDEXX, and now you joined Antech. I know this is hard giving these interviews like this, working for these large companies, especially, especially with all the, you know, PR departments, and so (indistinct speech 4:08). But is there anything differentiating between the two companies? As a veterinarian, I put it from this angle. I’m a vet, and I’m just, you know, I’m starting a new practice, I’m thinking about two diagnostic companies, I’m thinking I want to be with one or another, left aside the pricing and everything. What is new, innovative about Antech that we can learn from you since you just joined and maybe learned a lot of stuff there that would be exciting for veterinarians?
Jenn: So I think, in general, as a veterinarian you also know that the patients we treat oftentimes cannot speak to us. They don’t communicate clinical science, they don’t communicate how they are feeling, and diagnostics really is that portal, it’s that gate that we have, that gives us an opportunity to sort of give that pet that ability to communicate with us indirectly and through understanding of what diagnostic information is giving us. What is going on with that patient, and so I think, as I think about you know as a veterinarian when we approach a diagnostic company, it really is about Where am I able to build a relationship with that organization or that group that will be able to provide me with the tools and the innovations that allow me in a very sort of practical, sort of easy-to-do business relationship based way that we understand what we deal with the challenges as practitioners, and then how do I access that diagnostic information easily so I’m able to actually integrate that into my own practice management systems, and then actually communicate that information back to pet owner in language that is understandable to them, and make sense for what they deal day-to-day with their pets. And so when we think about diagnostic companies there are many of them out there, and I think what differentiates, you know, Antech in particular, as a diagnostic company is we are a company that we are very relationship based where easy to do business with, but we also innovators in what we are doing diagnostically, and we currently have been putting a lot of new innovations on market, and one of them most recent, has been Renal Tech which you probably heard about.
Ivan: Excellent. So renal tech, yeah, so what I heard about it and I would like to hear more, is that this is a predictive measure of certain parameters on the blood work that will tell you the likelihood of the cats of getting renal disease in the future. Yes, so that’s how I understand it. And from what I understand, and this is a pre-topic of the conversation, this is powered by AI, and which is also sort of the area of unknown that one of my developers on my previous team told me there’s developer joke, I hope, they’d appreciate it. So the difference between the artificial intelligence and the machine learning is that machine learning is written in the computer code, and artificial intelligence in power point presentations. So I guess that’s a.. (laughs). But can you tell us more about renal tech, and how does that works?
Jenn: Absolutely. I will not get into the discussion on the difference between AI and machine learning, but and I think most people are aware that AI or artificial intelligence or what some people refer to automated intelligence, really it’s using our computer to mimic human behavior in some way, and this idea of the machine learning really is just a subset of that where we use techniques and these computers figure out things from a lot of details so they actually can give us these applications. And so one of the applications that we have as an innovation from Antech, which I will also mention very excitingly just one of the best company animal award of 2019 from animal farm is renal tech. And so renal tech is a new kind of predictive diagnostic tool, and it’s really based on deep analysis of large amounts of data using artificial intelligence and machine learning. It’s unique in that it’s shifting the paradigm from the detection of disease, which we traditionally do as veterinarians, to actually predicting disease before it occurs. And in this case, renal tech is predicting whether or not a cat will develop chronic kidney disease, sometime within the next two years with greater than 95% accuracy, and I think that is exciting because as veterinary practitioners, we are now actually able to identify whether these cats are gonna develop disease or not. And, I think, for those of us who are practitioners who listen to this podcast, we know that this is a disease that has a very high prevalence within a feline population, it’s one of the leading causes of mortalities in cats. And so if we are not abe to diagnose these cats until they are advance in their disease, that will limit our opportunities to do something therapeutically and manage to be able to help these cats. And so if we are able to predict disease prior to the actual diagnosis, it then allows us to develop these personalized care plans for these cats, that may delay the onset of this disease, as well as it may even slower the progression of this disease. And I think one of the other important elements about that very much relates to diagnostics is that when we think about feline renal health, the kidneys are an important organ in the body, and so there are many other conditions in the body of feline patients that may contribute or predispose a cat to loss of kidney function. And so being able to predict this disease is an important part of this innovation, but what also happens is that if we do diagnose or predict that a cat is a positive renal cat, meaning that he’s likely gonna develop chronic kidney disease, sometime within the next two years, we then can start to investigate for or even concurrent conditions that then help us figure out “Is this cat predisposed to the disease that’s gonna lead it to the developed chronic kidney disease?And I think one of the common ones, most of us think about is something like feline hyperthyroidism. So we’re really thinking about overall feline health of our patients, it’s not just kidney health. And diagnostics are a valuable agent to sort of locking, as I said, sort of giving us that window into what our patients can tell us of what’s going on with their organ systems.
Ivan: It’s really, really interesting, Jennifer. I love AI, I love reading on it, and love the confuse by it, and it has a tendency to do that to most people, but I’d like to ask you a little bit more about your modeling, the AI modeling, you know. Where did you get the data from to build the model, to do the predictive analytics?And what is it looking at? What are the key factors that really trigger the ability of the algorithm to produce results when you get that diagnostic information?
Jen: Renal Tech is actually a global collaboration of many groups that are within the Mars pet care ecosystem, and so that includes Antech Diagnostics, it includes the pet-care science. It includes Mars advance research institute. The data was pulled from the banfield pet hospitals, so we looked at over 20 years of retrospective data. There are about 150 000 cats that were included, that met the criteria to actually build the model, and there were 600 000 hours of computing that we put in from a group that we collaborated with that is called process integration and predictive analytics team that based in Davis, California. And then we also worked with Dr. Johnoson Elliot who is a world-nephrologist and a member of the international renal intersociety board, and he’s faced in the Royal veterinary college in the UK. And so for years, most pet care has actually been using machine learning to analyze pet health data, and they’v e been using it in supportive market science programs, and using it in ways that help us transform how veterinarians predict and maybe even sort of diagnose disease. And so this particular tool is based on using all of these data, it’s used on what’s called the recurrent neural networks, and this recurrent neural networks are then used to identify patterns and trends from these large data basis that includes anonymized patient data. And then we use that and developed algorithm using proprietary equations. So what the tool uses is that it uses six common feline health parameters to deliver a renal tech status on that cat. And so that patient is required to have two data sets with 6 parameters that come from a minimum data based that we would typically get from a routine wellness or preventive care visit. And so those parameters include a white cells count from the complete blood count. It includes a blood urine nitrogen, and creatinine from the chemistry panel. And then, from the urinalysis it includes a urine pH, urine specific gravity, and urine protein. And then it also then incorporates the absolute or approximate age of that patient. So using that algorithm derives from that anonymized medical data that 150 000 cats and over 700 000 banfield patient visits over 20 years that renal tech tool was developed to predict whether these cats will develop chronic kidney disease or not in the next two years with a very high statistical accuracy.
Ivan: It’s incredible
Shawn: It’s amazing.
Ivan: So, from the practical point of view, you know, I’ve listened to this, I think it was some sort of a forum in Russia, and then there was professor from MAT pointed out why russians invented a lot but didn’t innovate a lot. And he made a distinction between invention and innovation. So invention being something really cool being discovered, and innovation is actually putting it into the field, into the practice and craving business around it. So as a practitioner, when I am using an Antech diagnostics, how do I get access to it and what does it look like? Of blood work? Is that something that is generated along the way? What is the actual application of it and what does it look like to the end user?
Jenn: So renal tech actually is provided to all our Antech customers free of charge. So they do not have to request a renal tech status on the feline patients. One that we have withing the data base the two data sets from that patient that includes those six parameters, the algorithm will be run on the data sets, and will be able to generate a renal tech status. And so when that veterinarian gets that renal tech status, which affors into one of three categories, which is that cat can be positive, it could be negative. And then we have a very small percentage of cats that make it inconclusive status, and what we are seeing is about 20% of the more than 100 000 statuses were reported so far are inconclusive. And really it just means that this additional data needed for us to report whether that cat is gonna be positive or negative with statistical certainty. And most of the times, it’s typically that we don’t have the urinalysis data, because maybe that veterinarian could not get a urine on that cat, for example. But in terms of what is this mean we give you the results, again, you know, we receive that as a report that comes once we have the information on the patient, they get it with no additional charge. And then if that cat is a renal tech positive cat, what does that mean in terms of the veterinarians having as that application of that status. What it means they now have an opportunity to really have a conversation at a much deeper level with a pet owner about. Let’s start monitoring this cat more closely. Let’s start looking to see if the’re actually eating good high-quality diet. Is that cat developing you know signs that typically may not be noticed, particularly for in the multi-cat house, like are they going to the little box more often. Maybe they are not eating as much, and maybe they are losing weight. And maybe also if that’s a cat on an annual basis, then come back with that cat, maybe bring that cat sooner so we can re-evaluate that cat that may be 3 months or 6 months, so we can actually identify at what point that cat actually does develop or going to that diagnosis of chronic kidney disease. And so I think this having the ability to have a status on a cat allows us to sort of really engage with that pet owner of that pet on a much deeper level, and really develop these personalized care plans. You know, another great example is, you know, we wanna know the risk factors for cats to develop chronic kidney diseases. If they have moderate distal periodontal disease. And we are seeing more and more awareness of pets becoming our babies, and they spend more time with us, they sleep in our beds with us, we’re probably more aware of what their teeth are like and what their breath smells like. And so if we can actually have that pet owner sort of have a good oral care program at home, and we as veterinarians instituted dental care program with them, where, again, in a position of being able maybe to delay the onset of very serious disease, like chronic kidney disease. So I think they all are very actionable items that are as veterinarians and care givers to these pets and to pet owners so deeply involved and emotionally attached to their pets, that then they also have something that they can do when they feel much more involved day-to-day on what’s happening with that cat.
Shawn: That’s amazing. And, Jennifer, one of the things that really stood up, I guess stood up to me in this conversation so far, is the unique ability for the Mars organization to attack these diagnostic opportunities because of the data that has existed within the organizations that like Netfield, and other companies that fallen into the umbrella. It appears very much that Antech, under the Mars umbrella, has a very unique ability to not just solve this issue, but many other issues. And I guess that one of the things that I would like to ask is there anything else interesting that you guys are digging into that leverages AI that you can talk about?
Jenn: Absolutely. So we are also using artificial intelligence in other ways. We are currently working with the Mars Next generation technology group, with our Antech imaging services, and we are using artificial intelligence to sort of refine the way we automate reading of ACGs. We are also looking at using artificial intelligence and machine learning, and what we call sort of natural language artificial intelligence to sort of help us at how we improve the efficiency of reading radiology reports, for example. Like what is you know a normal or abnormal thorax, for example. You know, the area of point of care cytology which is something that I think is becoming more and more topical in veterinary medicine, are we able to look at this digital images of cytological specimens as we store these images in the cloud, and we look for these patterns. Are we able to define a cell structural cell type more definitively. So I think this idea of artificial intelligence and machine learning is permeating through the entire realm of diagnostics and imaging not just from the element of being a predictive tool, but maybe also helping us in terms of the actual diagnosis. One of the areas that we are also working on with imaging is you know “Can we predict, you know, from sort of what we call pen hip area, is , at a very young age, is that dog likely to develop hip dysplasia, for example. And so all of those, I think, are important applications of machine learning. Can we maybe also use this in patients? Are we able to look to see what’s happening as we plan serial studies or we manage and treat cancer in our veterinary patients? Are we able to use AI to quickly and maybe accurately to determine what are the differences between studies and maybe what the effectiveness of our treatment, as imaging is a huge part of many times, along with radiation therapy for cats-patients. So I think this war of natural language, AI, and machine learning is just a sort of really gonna become the future of diagnostics and being important element of what we use as veterinary practitioners and diagnosticians.
Shawn: This topic is so interesting, I think if we had 3 or 4 hours, we might have a good start to where we could end up, kind of going around with this. Cause I have so many more questions that I would like to ask you, and really one kind of final comment from my side would be that it very much appears that you guys have the perfect storm to really tackle some of these very difficult issues when it comes to taking the data and putting mask computer power behind it. But, with that said, we are really running out of time, which is super unfortunate, but we gotta couple of questions that we would like to throw you away before we wrap up. One is, with an education career, like you’ve had, and so much exposure in the veterinary field, what’s one thing that you’ve come across, whether it’s a book, a ted talk, a youtube video, that you would like to share with our listeners.
Jenn: So, you know, one of the things that I have come across, that I believe has sort of really been an important impact on me as a person has been the idea of collaborative intelligence. It’s this idea of how as humans, we are using AI and joining forces to be able to take technology as well as our clinical acumen, our integrity of knowledge and information we have, to be able to put together this fusion, this ability to integrate technology. But that judgment we have, that intelligence that we develop that social and emotional intelligence, to be able to bring forward to what we do day to day in our lives, but also to what we do for our patients in a very exciting way. And I know it’s something that’s been talked about with you know a lot of the Harvard business review folks, and a lot of social sciences. So I think it’s really exciting. We’re using intelligence in a way that doesn’t meet a gate against what we know as a human beings, but we’re using it to actually make us better at what we do and to actually do it in a way that is efficients, it’s getting us into more personalized approach on how we interact with each other, and what we do for our professions. And so I think this is a really exciting area.
Shawn: That’s great. And then the final question is what other innovator that you know in your network, do you feel we should have on the show?
Jenn: Dr. Paul Fisher, a colleague of mine, has been inspirational innovator to me. His ability to look at natural ability learning artificial intelligence and some machine learning, and implement that into an imaging services and the clinical applications, and the efficiency of what we bring forward to us as veterinarians really has been impactful. He’s continuing to work in this area, and I believe he would be someone who would be a huge innovator in the future, and someone you should speak with, I think as we all appreciate diagnostic imaging specialist in veterinary medicine are at a major shortage, and the work that fallen, his team, are doing in collaboration with the Mars pet care, next generation team is just incredible. And I think he would be a huge asset on this podcast.