Last spring, a dozen horses died on the track at Churchill Downs, home of the world’s most famous horseracing event, the Kentucky Derby. A report later released by the Horseracing Integrity and Safety Authority found no single cause for the deaths but made several recommendations, including expanding the use of wearable technology to detect potentially fatal injuries.
Washington State University has been at the forefront to refine this technology through its partnership for the past four years with a company that makes a wearable sensor for horses that weighs less than a deck of cards. A team of WSU researchers recently published the results of a study that tested how that sensor performed while analyzing the strides of hundreds of horses competing at Churchill Downs and another track in Kentucky last year. Abnormal stride movements revealed bone and muscle damage hidden to the naked eye that had a high risk of developing into a fatal injury.
Denise McSweeney, the lead author of the study and an equine surgery resident at WSU’s College of Veterinary Medicine, joins us to share her findings.
The following transcript was created by a computer and edited by a volunteer:
Dave Miller: This is Think Out Loud on OPB. I’m Dave Miller. A dozen horses died at Churchill Downs last spring. That is the famed track of the Kentucky Derby. A report by the Horseracing Integrity and Safety Authority found no single cause for the deaths, but made several recommendations, including expanding the use of wearable technology to prevent injuries. Researchers at Washington State University have been at the forefront of this work. A team there has been testing a wearable sensor that they say can reveal problems before they become truly serious.
Denise McSweeney is the lead author of a new study about this. She is an equine surgery resident at WSU’s College of Veterinary Medicine and she joins us now. It’s great to have you on Think Out Loud.
Denise McSweeney: Thanks very much, Dave. It’s great to be here. Thanks for having me.
Miller: So I mentioned the really shocking numbers from that one famous track, maybe the most storied track in American horse racing. But broadly, how common are serious horse racing injuries?
McSweeney: A record is kept of serious injuries which occurred to thoroughbred racehorses by The Jockey Club, and that’s called the Equine Injury Database. That record looks at any horse that suffered a fatality during racing. And that includes musculoskeletal injuries and cardiac deaths. For the year 2023, that number was 1.3 in every 1,000 starts. Those records started being collected in 2009. At the time, the number was around 2 in every 1,000 starts. So there have been serious improvements over the last decade or so.
Miller: And when we say serious injuries, does that usually mean that the horse has to be put down?
McSweeney: Yes. So that 1.3 in 1,000 starts is in relation to non-musculoskeletal injuries, so horses that may have suffered a cardiac event. And then it’s also including horses that suffered what we call a catastrophic musculoskeletal injury. That means that that horse has suffered a traumatic injury, to usually a limb, that makes it necessary for that horse to be humanely euthanized.
Miller: Why is it that that has to happen? Why can’t they get some kind of treatment, and then I don’t know, just literally go out to pastures, eat a lot of hay and retire lying down?
McSweeney: Yeah, that’s the dream. There are some fractures that we do repair. So it’s not the case that every single horse that breaks a leg has to be euthanized. There are some candidates that are good candidates for surgery.
The main problem, and the big reason why horses are not the best patients for surgical repair of fractures, is that the horse is standing on four legs. And immediately after he has fractured a leg, if we go to repair this leg, he needs to immediately be weight bearing. The reason is that when you look at a horse’s foot and you look at that hoof, what’s actually inside in that half is one bone, and that’s the coffin bone. That’s the equivalent to the bone that’s in the tip of our finger. That’s being supported by structures inside in the hoof caps called the laminae. And those are very sensitive to pressure and changes such as being overloaded. So if a horse, for example, were to fracture his right front leg, then he would actually have to redistribute all the weight from that right front leg to his left front leg, while he can’t use the right front one. And that causes a lot of damage to the laminae. And it actually results in that bone that’s sitting in the hoof capsule sinking. And that’s what we refer to as laminitis. That’s extremely painful for a horse.
So they need to be able to put all of their weight back on the fractured leg immediately after a surgical repair to avoid laminitis, as laminitis itself is enough to warrant a horse’s euthanasia due to welfare concerns. So the bones can be repaired on occasion. However, it depends on how well we can make them comfortable enough to stand on that opposite leg.
Miller: You used to work at an equine hospital in Kentucky. How did that experience shape your interest in racehorses?
McSweeney: So I was fortunate enough to complete a surgical internship at Hagyard Equine Medical Institute in Lexington. And I will say that thoroughbreds are like no other, because in that industry and in that location, you’re able to follow these awesome creatures. They may come in as a mare who’s having difficulty delivering a foal. So let’s say you take that mare to surgery and perform a Cesarean and you bring that baby into the world. And then you get to follow them as they grow up. These horses go through sales. So you’ll see these horses go from being a little wet newborn after being pulled out of his mother’s uterus to a lovely shiny yearling heading to the sales ring. You might watch him make a million bucks, and then you get to follow him on to a racing career and see how he did. It’s really wonderful to be able to watch the evolution of something that you, yourself, might have been hands on with from the very first minute that that horse stepped foot into this world. It’s really a really fantastic industry, and there’s an amazing group of people in it. So I feel quite passionate about the thoroughbred industry.
Miller: Well, let’s turn to this newish tool that ideally could make this industry a lot safer. What is StrideSAFE?
McSweeney: StrideSAFE is a company set up by Dr David Lambert. It essentially consists of a small sensor, smaller than the size of an iPhone, that’s worn in the saddle cloths of race horses when they race. It sits in the saddle cloth and it measures the horse’s movements in three directions. It’s going to look at dorsal/ventral, which is up and down; mediolateral, which is side to side; and longitudinal, back to forth movement of the horse as it races. It collects about 2,400 data points every second.
So the horse’s movement then is plotted against time, and it actually produces a graph which illustrates the horse’s stride as they move. Based on that horse’s stride, we can see if it deviates significantly from what we see as the ideal stride, and then identify any horses whose stride appears to be dramatically different to a good stride, and therefore putting them at risk of injury.
Miller: How was it trained to know what a good or healthy stride feels like, charts out like?
McSweeney: Initially, when this data was being collected and looked at, it was being used to identify horses which may actually be high performing athletes. So it was being used to look at performance prediction. And in doing so, it was recognized that the very top quality horses were able to produce a very similar stride, and do so consistently through their race. So actually, the ideal stride that we use for this algorithm is based on the stride pattern of 37 graded stakes winning horses. That’s what we call our ideal stride.
Miller: Is it that they shared similarities with each other, or that they were all impressively uniform compared to themselves?
McSweeney: Both. All of those horses were able to produce very similar strides to one another, and consistently do so throughout their race. There wasn’t a huge amount of variation between their strides.
Miller: And what has the system learned about the gaits that, on average, are more likely to proceed a catastrophic injury?
McSweeney: I guess the best way to put this is that our ideal stride is those 37 great stakes winning horses. So we have the ideal stride. And from putting these sensors on thousands and thousands of horses, we began to see a change in horses before they actually suffered a catastrophic injury. So using machine learning, using AI, the algorithm was able to see common things that occurred in a horse’s stride on the lead up to a catastrophic injury. So initially, we were just looking at standard deviations from the mean, looking at how far from the ideal stride a horse’s stride deviated to see how much at risk they were of potentially suffering a catastrophic injury. But with more and more data and more information on the horses, we were now able to see actually how far did it deviate from the ideal? And how similar was it to a horse who was, after, suffering a catastrophic injury?
Miller: It seems like it’s one thing to give an algorithm data from horses that were either healthy or injury prone, or had an injury later. It’s another then to use that algorithm as a predictive tool. How do you verify that it’s accurate?
McSweeney: The thing is that this system, it’s not a diagnostic tool. It’s a screening test. The reality is that 1.3 in 1,000 starts, we know that the incidence of catastrophic injuries, that’s pretty low. It’s not like we’re having horses break every single day. So we had from a cohort of horses … let’s say it was 6,600 starts. Out of that group, if we categorize the horses 1 through 5 –1 being at the lowest risk and 5 the highest risk – we had about 70% come in as category 1, and then 3.5% come in as category 5. I think we had 18 musculoskeletal fatalities out of that particular group.
And we know that there’s going to be horses that get triggered as having had a high risk run. But there’s also interpretation here that has to come in from the veterinarian and the individual horse. On occasion, when these horses are running, something can happen during the race. They can be ridden in a certain manner which can cause them to have some irregular strides. This is usually if a horse needs to be pulled up sharply to stop it from grabbing the heels of a horse in front of it, if it’s pulled out wide to avoid a horse. So there are things that can happen that can trigger a high risk warning. So we know that these horses are at an increased risk. But there’s definitely some interpretation to be done with the trainer and the veterinarian who’s attending the horse.
Miller: What can you tell us about some of those horses that came up with a 5 score – the more worrisome end of this – where it seems like it was appropriate for the system to flag them?
McSweeney: So the cases that we reported on in this case series, one of those horses, he was wearing the sensor for the first time actually. And we actually had data on this horse from training and racing. And his stride was quite erratic. When he went out and raced and he came back with a high risk warning, we reviewed the race footage and didn’t see anything in the race that could have caused this, so then went and spoke with the trainer and their veterinarian and they didn’t really find anything clinically wrong with the horse. He seemed to be looking OK.
But we did have funding for advanced diagnostic imaging, which they availed of. And then they sent this horse for a standing PET scan. PET scan, it’s a nuclear medicine that’s used typically in cancer medicine in humans. However, it’s been recently developed to be used in horses for orthopedic issues. So they get injected with a radiotracer. And what happens is that they will go and get taken up at an area of high bone turnover, an area where you have bone being resorbed and laid down. And you have to remember that bone is a living thing. So when you think about yourself, if you have a fracture, especially if you were a kid, got a fracture in your arm, took several weeks in a cast and it got better. Well, if you were to have a PET scan at that time, your fracture would light up like a Christmas tree.
So that’s what this horse did, he went and had a PET scan. And it lit up in an area on the distal part of his leg, down low. And that lit up in a spot that we know to be a recurring site of injury in racehorses. So that put this horse at an increased risk of what we call a condylar fracture. That horse was then removed from training and racing. He was given I think 130 days off out in the paddock, out in the pasture. And then he came back to training slowly. He’s since raced twice, and on both occasions, he’s had a low risk category in his races.
So that’s one that we saw as a high risk course, he went and had the appropriate imaging, and potentially avoided a catastrophic injury.
Miller: It seems significant that in that case the jockey didn’t notice anything, the trainer didn’t, the human experts didn’t pick up something that the sensors did?
McSweeney: Yeah. And that’s where these come in. Every horse before it goes out and trains in the morning, it gets examined by his trainer and his veterinarian. Every horse before it races gets examined by the regulatory of the racetrack. And for the most part, it’s not that somebody’s missing something obvious on these horses. The problem is that they have pre-existing pathology. Quite a lot of the time these horses have something that’s going on. And it will be in the shape of some micro crack, some resorption, a little bit of callous formation. And that’s to an extent normal, until it’s not. So there’s a period of time where the resorption of the bone has occurred and the new bone being laid down hasn’t finished. And that puts that bone then in a very weak state.
However, when you’re jogging a horse, when they’re just trotting along, they don’t always show the signs that they’re going to show at the gallop. The forces on their legs are different at the trot than they are at the gallop. And horses are examined at the trot. So they’re not always subjectively lame when you trot them, they don’t always react if you flex on their legs or squeeze on them. It’s not that these horses are out there jogging up and down the footpath with their head bobbing up and down. Most of them look perfectly fine. So unfortunately, they are getting missed and they are getting into races.
What’s very interesting is a lot of these horses don’t even show a decrease in performance. Horses can go out, they can win a race, and they can actually fracture a leg in that race that they have gone and won. So it’s quite challenging to identify these horses.
Miller: In general, are the veterinarians you’re talking to, the trainers, the owners, interested in new technology like this? Or are they wary of an AI, an algorithm coming and telling them things that they don’t trust?
McSweeney: It’s like anything else, it’s 50-50. You have some people who are really interested in new technology, they’re the guys with all the gadgets. And then you have some people who are a little bit more wary. They’re afraid that this is going to take over from just good genuine horsemanship.
Miller: I have an idea of some kind of cigar chomping, old-fashioned guy right now, who says “what are you doing with your sensors?”
McSweeney: You’d be surprised, actually. You’ve got some of the guys who’ve been in this a long, long time, and they’re enthusiastic about whatever can come in and make the sport safer. I don’t want to stereotype those guys, because honestly, everybody really just wants things to be better.
You certainly have people who may feel a touch offended, as if they’re being told that they can’t see if there’s something wrong with their horse. But I think you have that in every walk of life. For the most part, people just want to always make sure that the industry is safer.
Miller: Can you imagine a time when this sensor, or something like it, were required by horseracing associations?
McSweeney: Yes, absolutely. Right now, StrideSAFE has been used on over 12 or 13 tracks in the United States. So it’s already been used very, very frequently. However, I think we’re never going to replace good horsemanship, and never going to replace a thorough veterinary exam. But I do think this is something that’s going to become widespread on race tracks across the US.
Miller: Denise McSweeney, thanks very much.
McSweeney: Thank you very much for having me.
Miller: Denise McSweeney is an equine surgery resident and veterinarian at Washington State University’s College of Veterinary Medicine.
Contact “Think Out Loud®”
If you’d like to comment on any of the topics in this show or suggest a topic of your own, please get in touch with us on Facebook, send an email to thinkoutloud@opb.org, or you can leave a voicemail for us at 503-293-1983. The call-in phone number during the noon hour is 888-665-5865.