As the New York State Gaming Commission's Equine Medical Director, Dr. Scott Palmer is responsible, more than anyone else in the state, for horse safety. It's a responsibility Palmer takes very seriously, and with new wearable biometric technology called StrideSafe, Palmer and his team are doing revolutionary work in detecting potential musculoskeletal injuries, which lead to the majority of horse fatalities, in their earliest stages. Tuesday, Palmer joined the TDN Writers' Room presented by Keeneland as the Green Group Guest of the Week to explain the technology and process of a program that, if adopted nationwide, could help get the U.S. Thoroughbred fatality rate as close to zero as possible.
“As we know from our 20 years or so of research in this area,
85% of the horses that break down have preexisting musculoskeletal disease or damage, and if we can identify that damage early, we can intervene and take care of it,” Palmer said. “You've probably heard a lot about PET scans and CAT scans and advanced imaging. These are great diagnostic tools, but they're not great screening tools because first of all, they're not readily available. Second of all, they're not inexpensive. Third of all is that they are not necessarily accurate as to what we're measuring. So they're terrific, but we need a screening device to identify horses that need to go get that imaging, because we can't do a PET scan on every Thoroughbred racehorse. It's just not practical.
“[StrideSafe] is like a check engine light in your car. When you're driving down the road and the check engine light comes on, that doesn't mean you have to stop the car immediately, but it means you're going to get this thing looked at because something's going on here. If you don't do that, something bad is going to happen. So that's what we've got here is a check engine light, and with that kind of information, that's going to help us to identify these horses at risk of injury, because we can see lameness before a human being can see it or before the jockey can feel it.”
The StrideSafe device is about as big as a cell phone and is put in the saddle cloth of racehorses at New York tracks when they run. It detects any deviation in each facet of a horse's stride and labels that horse in the colors of a traffic light–green, yellow and red. From their pilot study last year, Palmer and his team found that horses labeled red were much less likely to run back in the same amount of time as green or yellow horses.
“If a horse is a red-flag horse, it means this horse has got something significantly abnormal about his gait. And that is meaningful,” Palmer said. “That's a danger sign if we see these red signals. A horse can have a different degree of variation from normal, and we're not too worried about the yellow horses. The yellow means okay, caution light, slow down. Take a look. The red light means you really have to get this horse looked at because something's going on here. And maybe it's not obvious to the human eye right away. The horses that were red [in our study], only 40% of those horses made it back to train or race within four months. Almost 80% of the greens and yellows did. So in other words, what this means is that I can accurately tell a trainer, if you get a red alert, you've got a 40% chance of making it back to race in four months. That's a really bad business model, if nothing else, and it also means that your horse is likely at risk of injury. So we're going to record every horse in every race at Saratoga, and we've been working with Joe Appelbaum and NYTHA, where for all the red reports, I will notify Joe and he's going to send an email to the trainer of that horse, saying this is what it means, this is what it doesn't mean, this is what you need to do.”
Last week at Royal Ascot, The Ridler (Brazen Beau) sprang a surprise in the G2 Norfolk S. when out-performing his odds of 50-1. That wasn't the only odd spell The Ridler cast during the race.
Into the final furlong, The Ridler drifted markedly left across the field, hampering the chances of several other runners. Controversially for some, The Ridler kept the race after a steward's inquiry.
Duff Gordon presented a chart showing the acceleration and speed of a horse during the race, and how it was impeded by the winner.
“You can see how that horse in blue, how much its velocity was cut off,” said Duff Gordon, pointing to the marked deceleration of a beaten runner at the time of The Ridler's antics.
“That's a much better way of telling the story rather than running the replay 700 times, which is what the TV companies have done to date,” said Duff Gordon. “The stewards can't yet use that information,” he added, “but hopefully they will soon.”
TPD's chart shows the deceleration of the impeded horses in the Norfolk
Duff Gordon was part of a panel looking at some of the data collection technologies forging a new path in equine welfare and performance during racing and training, and racecourse customer experience.
“This is a new frontier,” said Scott Palmer, equine medical director for the New York Gaming Commission, about the StrideSAFE sensor, which TDN has written about here.
Nevertheless, the panelists emphasized how, in many regards, the stamp from data collection on the racing world is still very much a fresh one.
Greater accuracy comes from sheer depth in numbers, and pick-up of these technologies among industry stakeholders at large has been finicky at best.
Correct interpretation of the data is also key. Racing is hardly awash with number crunchers, data analysts and epidemiologists. That's why, said Palmer, “this is baby steps right now.”
Duff Gordon agreed.
“We create in-running odds. That's making 10,000 calculations per-horse, per-second to predict who's going to win the race. We can refine that all the time,” he said.
“The last few years has been about acquiring the data, getting onto as many racecourses as possible. The present and the future's all about producing front ends and mining that data, so we're hiring huge numbers of data scientists,” Duff Gordon added. “Any budding data scientists, please do get in touch with me. We can never get enough of them.”
Presided over by New York Thoroughbred Horseman's Association president, Joe Appelbaum, the panel also included Valentin Rapin, managing director and co-founder of Arioneo, a horse performance company.
The technologies that Arioneo produces are geared around morning training, including heart-rate monitoring, locomotion and stride data, and GPS tracking of speed, distance and acceleration.
Broadly speaking, these discreet, easy-to-use technologies can be utilized to gauge things like the overall fitness level of the horse, it's optimum speed, suitable distances and ground preferences.
There's a welfare element, too, explained Rapin, as the heart-rate monitor gives trainers the ability to check for cardiac abnormalities.
“Let's say the horse is about 125 beats a minute every day. If one day you see the heart rate go up to 140, 150, it will probably mean there is something wrong,” said Rapin.
“You can contact your vet directly or your vet can also have access to the data via distance to monitor this for you,” Rapin added.
“It's really in the strike-rate,” interjected Duff Gordon.
“The trainers who have that huge, extra level of insight, their horses are ready to rumble and you have trainers with less horses punching above their weight,” Duff Gordon said, pointing towards the likes of George Boughey, a young UK-based trainer with an English Classic win to his name into just his fourth year with a license.
“He's got less than 50 horses, and [has] no right to be winning [like he is],” said Duff Gordon, adding how the likes of Boughey have harnessed the use of performance monitoring technologies with shrewd purchases at the sales.
StrideSAFE is a technology that fits into a horse's saddle towel with the ability to detect at high speeds lameness invisible to the naked eye. Since last summer, it has been used on thousands of starters across Saratoga, Belmont and Aqueduct.
“What were able to do with this device is actually detect lameness in early stages to provide for timely intervention,” said Palmer.
“We can work with the trainer and say, 'Your horse is not lame today, but the warning signs are clear. You need to get this horse evaluated by a veterinarian. It needs a good diagnostic workout because something's going on here,'” Palmer added.
Hundreds of times a second, this wireless iPhone-shaped device takes an assortment of measurements to capture in minute detail the movement of the horse at high speed.
These measurements include the horse's acceleration and deceleration, the up and down concussive movement of the horse, and its medial-lateral motion–what is, in other words, the horse's movement from side to side.
It works like a traffic light signal, providing a green for all-clear, an amber for possible caution, and a red for possible danger. These ratings are calculated by how many standard deviations the horse is from the norm.
“If they exceed three standard deviations from the mean we have what we call red alert, and that means there's something happening in this race that needs to be investigated,” said Palmer. “There's a change here.”
Palmer told the story of early on into the trial of a horse that finished second in a maiden special weight at Saratoga.
Immediately after the race, the horse appeared healthy and sound. But when Palmer received the StrideSAFE read-out of the race the following day, it showed worrying abnormalities in the horse's balance.
“When I got this data, I didn't know exactly what it meant,” said Palmer, who said that he called the trainer, asked about the horse.
“The trainer said, 'Well, this horse finished the race great.' Jockey had hopped off. Trainer was there in the unsaddling area. Everybody was happy, finished second in a maiden special weight for 2-year-olds at Saratoga,” said Palmer.
By the time the horse had walked back to the test barn and began to cool-out, it became lame–so much so, the horse had to be shipped back to the barn. A subsequent radiograph showed the horse had suffered an acute fracture of the third carpal bone.
“It was a very dramatic example for me that we were measuring something important,” said Palmer.
While researchers are still analyzing the data from the longer study, Palmer shared some numbers showing attrition rates among a small cohort of horses studied at Saratoga: 15 that had received red alerts, 25 with a yellow alert, and 91 with green alerts.
Broadly speaking, the horses with a red alert generally did not make it back to race for a “significant amount of time” as compared to the yellow and green horses, said Palmer.
The red horses didn't compete in as many races as the yellow and green alerted horses during a four-month follow-up period, nor complete as many high-speed workouts, added Palmer.
“Interestingly, only 40% of the horses that had a red classification were able to race at all over the next four months after the analyzed race compared to almost 80% of the greens and the yellows,” said Palmer. “I thought that was powerfully significant.”
In his “zealot's pitch” at the end of the panel discussion, Appelbaum implored the industry to accelerate its adoption of these kinds of technologies.
“There's not a human football player or soccer player at the professional level that's not using a catapult vest. In F1, the drivers all use gloves that track their heart rate and perspiration. Weightlifters are all using push-bands to understand their load,” said Appelbaum.
“Humans sports are about 10 years ahead of us,” Appelbaum added. “But we can catch up and we should catch up. It's not just for the benefit of the betting public, but it's really for the benefit of the horses.”
By the beginning of April, there had been two fatal musculoskeletal injuries during the current Santa Anita meet. Wind the clock back to an identical window in 2019, there had been 22.
During Aqueduct's benighted 2011-2012 Winter meet, 21 horses died, 18 of which were fractures sustained during racing. Over the next seven years, New York's four racetracks saw a 50% reduction in racing fatalities.
What is the common denominator? Huge advances in identifying those horses at higher risk of sustaining fatal injuries in racing and training, and much tighter safety nets to filter these horses out of the racing pipeline before a catastrophic event occurs.
Racing's North Star is to reduce the number of musculoskeletal injuries to a single duck egg. But there remains a stubborn impediment: The ability to easily and accurately identify those emerging and subtle injuries that can't be detected with the naked eye but have the potential to devolve into a sickening fracture.
How prevalent are these sorts of issues? Well, 21 one of the 22 horses that died during Santa Anita's 2018-2019 Winter meet showed pre-existing pathology at the site of their fatal injuries.
Enter stage left the 21st century, and a collision of new technologies that bring an objective, mathematical approach to pin-pointing these hitherto elusive and barely perceptible problems.
The New York Racing Association (NYRA) is deep into a project that began last summer to trial a sophisticated biometric sensor mechanism which fits snugly into the horse's saddle cloth and can detect minute changes in a horse's gait at high speed. Called StrideSAFE, the sensor has been proven to detect problems in a horse's stride sometimes weeks in advance of a catastrophic event occurring.
Over on the opposite side of the country, The Stronach Group, under its 1/ST banner, is gearing up to unveil its own system which uses high-definition cameras to create detailed skeletal movement maps of horses as they navigate the racetrack. Company officials believe the technology has the potential to similarly red-flag horses at the very earliest warning stages.
What's more, these new kids on the block converge with an emerging generation of imaging modalities-think PET, MRI and CT-more capable than their diagnostic ancestors of providing a clear yes or no answer to the presence of subtle pre-existing problems.
And now, some of the industry's most pragmatic, unflappable leaders are making the argument that, given further development and understanding, these biometric systems hold the key to reducing fatal musculoskeletal injuries from the sport even further-potentially altogether.
“This is probably one of the most important contributions to the Thoroughbred horse industry that has ever been made,” said Scott Palmer, equine medical director for the New York State Gaming Commission, about the StrideSAFE sensor. “I do have a big stake in saving horses' lives, and so, in some respects this has been a holy grail.”
The Sensor Is The Ultimate Jockey
When David Lambert studies horse racing through a technology lens, he compares it to a ladder propped against a house going into the third-floor bedroom window.
Currently, racing's hitherto timid embrace of cutting-edge technologies-even basic training monitors, for example, used liberally for decades in human sports-have kept the industry rooted to the lawn.
“Me trying to get this [StrideSAFE sensor] introduced is the first rung of the ladder,” Lambert explained. “Keep that up over time, slowly but surely, you'll get right up to the third floor and walk into the bedroom.”
Lambert is a loquacious and affable 73-year-old veterinarian hailing from the North of England, who has dabbled in the odd bit of race-riding and training. He's lived Stateside for more than 50 years, during which time, he's spent decades digging through the mathematical arcana of performance prediction, and the role that sensors play.
Mickael Holmstroem, a Swedish PhD with expertise in equine conformation and locomotion, and Kevin Donoghue, professor of electrical and computer engineering at the University of Kentucky, have been Lambert's business and academic partners for much of this journey.
In 2019, they all formally embarked upon a mission to design a system that StrideSAFE looks like today. In all, Lambert reckons he has poured $1 million of his own money into its development.
It's no surprise, then, that he talks about it with the enthusiasm of a teenager with the keys to his first car.
“The sensor is the ultimate jockey,” Lambert said. “It's the best rider that ever rode a horse, never forgets anything, and picks up data 800 times a second in each of three directions for a total of 2,400 data points a second from the horse.”
So, how does it work?
This wireless iPhone-shaped device fits snugly into the saddle towel, and eight-hundred times a second, it takes an assortment of measurements to capture in minute detail the movement of the horse at high speed.
These measurements include the horse's acceleration and deceleration, its up and down concussive movement, and its medial-lateral motion-in other words, the horse's movement from side to side.
Ultimately, the sensors capture the sorts of high-speed lameness invisible to the naked eye but significant enough to cause major musculoskeletal failures at some point down the line unless someone intervenes on the horse's behalf.
“A horse can basically stand to race. Their bones are strong enough, their ligaments are strong enough,” said Lambert. “But what they can't stand is imperfection over and over and over again. They're going to break something.”
When explaining the equine biomechanics underpinning the success of the StrideSAFE technology, Lambert first compares the horse to an antipodean cousin-the kangaroo.
“People don't get that,” he said, of the comparison. “Sixty or seventy percent of the energy it produces to go fast is from spring or elastic recoil.” He then breaks a single stride into three stages.
In the first phase of the gallop, the hindlimbs load and propel the horse forward, kangaroo-like. In the second, the horse shifts its weight to the front, its forelimbs acting like shock-absorbers. This is followed by a period of suspension, the horse entirely airborne, a time for it to catch its breath.
But if the horse suffers a physical problem, however, it cannot adjust its body to compensate when its feet are on the ground. It can only do this midair, rotating its spine and pelvis in preparation for landing.
“The horse does all kinds of things in the air, twisting and shaking and moving,” Lambert explained. Imagine a race-car hurtling along at high speed, one of its bolts working loose.
What's more, these midair adjustments are infinitesimal, occurring within a 1/100th of a second window imperceptible to even the jockey-but not a sensor.
“It tries its best to re-align itself and repeats it all over again,” Lambert added, of the horse. “Then six, eight, ten weeks in, that front leg is going to feel it. You're going to get a joint or a knee. You start to see the obvious lameness.”
Coady Photography
StrideSAFE works like a traffic light signal, providing a green for all-clear, a yellow for caution, and a red for possible danger. In mathematical terms, a red means that the horse's gait abnormalities are beyond two standard deviations of the norm.
It's important to note that a red-light doesn't necessarily indicate a brewing issue. It could simply mean the horse is slow or that it doesn't try.
Nevertheless, from trials at tracks in Tasmania, and at Emerald Downs and Kentucky Downs here in the U.S., the sensor has repeatedly proven effective at highlighting gait abnormalities sometimes weeks in advance of a fatal breakdown.
Lambert shared data slides showing the five-race progression of a 6-year-old mare. During her first three runs, the mare's way of going-what Lambert calls her “fingerprint”-showed no abnormalities.
Her fourth race garnered a red-flag, however, despite finishing an encouraging second that delighted connections. A few weeks later in her fifth race, the mare fatally broke-down.
“In a normal day of racing just one or two horses will get a red flag warning and this small group contains a significant proportion of the horses in danger of catastrophic injury,” Lambert said.
Early injury detection can help to not only avoid painful, expensive and time-consuming treatments but expedite convalescence time, too-a fillip for an industry grappling with the consequences of a dwindling horse inventory.
“Attrition of racehorses is an enormous problem,” said Scott Palmer. “This device gives the opportunity not just to identify horses with a gait abnormality before it becomes evident normally to a human being, it allows regulatory veterinarians and racing officials to work together with trainers early in the process to help keep horses in training.”
Last summer, NYRA used StrideSAFE on every horse in one race per day at Saratoga, allotting each the requisite red, yellow or green label. By the end of the meet, 3% of the horses measured had been red-flagged.
All participants were then tracked over the subsequent four months, to see if and when they returned to race-day competition.
Of the green horses, 78% were able to race back within four months. Of the yellow horses, 72% raced within four months. But only 40% of the red-flagged horses returned to race within four months.
“If you have horses that don't run back regularly, don't race on a regular basis, there can be a number of reasons for that, but the most common one is lameness,” said Palmer.
So impressed has Palmer been with the technology, NYRA has used it on every horse to race during the most recent Belmont Park and Aqueduct meets, for a number totaling roughly 6,000 recordings.
That data is now being evaluated. Plans are also afoot to trial the technology on horses during training. But Palmer already imagines a future where this technology is a more permanent part of the NYRA furniture.
“My vision about this is that when I get notified of a red-alert, I can just send an email to the trainer that says, 'trainer, your horse just got a red-alert today. What does it mean, what does it not mean, and what are your next steps,'” said Palmer. “That horse is going to get extra scrutiny, and that's the name of the game right now.”
And it reveals much when Palmer, chief veterinarian at one of global racing's highest profile jurisdictions, admits that StrideSAFE has evolved his understanding of the equine athlete.
He said, “I will never look at the horses in the same way.”
While StrideSAFE utilizes motion sensors that attach to the horse, 1/ST is readying for later this year the launch of a three-year, multi-million-dollar effort to design a biometric system with multiple uses, including the ability to create detailed skeletal movement maps of horses using high-definition cameras.
“We're at the beginning of the journey in terms of 'how far can we take it?'” said Paul Williams, who heads up technology at 1/ST. “But we're beyond the beginning of 'very excited about what it can do.'”
The basic building blocks of this system consisted of creating virtual 3D models of each 1/ST track which were then then mapped against the position, angle and zoom of the TV cameras already dotted around these facilities.
It took a year, said Williams, to be able to pin-point a horse on the track to within 13 centimeters of its actual location. Since then, he and his crew have whittled that down to a six or seven-centimeter range of accuracy.
“That gives us a level of accuracy where we can track the physical horses and people and vehicles and weather, weirdly, of the locations from the TV camera,” said Williams.
And with 85 million historical race clips already plugged into the system, “that's a nice place to be because you have such a large data lake to start to test and to infer theories,” he said.
Among the information the system collects includes acceleration and deceleration and horse stride length. What's more, the system, said Williams, can “effectively replace” and in some instances vastly improve upon a host of commonly used industry practices and technologies, like start-stop time, race order, race-speed, top-speed, and the number of times the jockey uses the whip.
“Even down to pseudo-jockey aero dynamics, based on where they're positioning their weight on the equine athlete,” Williams said, pointing out that some of the derivative data could be packaged and geared towards gamblers.
And from this model, Williams and Dionne Benson, 1/ST's chief veterinary officer, are in the process of adapting it to identify patterns of horse lameness not visible to the naked eye.
The basic principle is fairly simple: High-definition TV cameras will pick up QR codes-think restaurant menu barcodes-attached to a horse's saddle towel during training, sending back in real-time a rich pool of highly detailed, high frame-rate data.
Over time, this system can accrue historical skeletal movement maps for each horse at all gaits, from walk to high-speed workouts.
Though this part of the system is still being beta-tested, the range of motion captured by the cameras is so sensitive, said Benson, they can pick up fractional changes to the fetlock, for example.
“At the trot, fetlock drop is very informative,” Benson explained. “If we know a fetlock is dropping more than it had been, that is a potential indicator of a problem going on-maybe because the ligament is weakening.”
A key to accurately pin-pointing horses with early brewing problems is a matter of proportion-in other words, having a comprehensive enough database of races, workouts and training days for the computer algorithm to identify outliers.
An outlier could be a horse that displays troubling gait changes over a period, for example. Or the system could red-flag certain horses using proven patterns of movement abnormality among horses in general. And so, there's another crucial component to this venture.
“Right now, you can't see a lot of the stuff that's picked up by the computer with the naked eye. So, there's going to be a period of time where we're going to be looking at issues that we may think is something but isn't,” said Benson.
The system's ultimate success, therefore, also rests upon what protocols are in place to siphon red-flagged horses towards diagnostic technologies-modalities like the PET and MRI units already housed at Santa Anita-capable of detecting those minor bone adaptations that can turn ugly.
Given the system's performance to date, both Williams and Benson appear noticeably sanguine about its promise to screen out early the horses who illustrate the sorts of subtle problems that could prove catastrophic.
“I'm pretty confident, because of the quality of data that keeps coming out of this thing, that we will get very close,” said Williams.
The system is expected to be launched and live at all 1/ST locations later this year, including training locations.
“If we get positive feedback there, I think we'll look to extending it beyond our tracks to our partners,” said Williams. “Horse populations move around the country, and to have this be a useful benefit for the industry, it's got to track a horse all the time.”
What stands out from discussions with proponents of these biometric technologies is the potential for adaptation, using them to compare the relative strengths and weaknesses of different jockeys, for example, a horse's performance on different tracks, and how hard it works.
Lambert tells the story of a horse fatally injured during a race. He said the post-race read-out showed noticeable gait abnormalities a full 70 seconds before the fatal event-from the minute the horse exited the gate, in fact.
And so, theoretically, such systems could also eventually be used in real-time, opening the door to preventing catastrophic injuries from occurring during a race or workout.
For the ambitious trainer looking for an edge other than through pharmaceutical intervention, therefore, technologies like StrideSAFE hold the key, said Lambert.
“You get to know your horse really, really well. You get to be the best horseman you can be by having the right kind of data to care for your horse,” he said. “It's the future.”
