Equine Welfare: Can Wearable Sensors Prevent Racehorse Deaths? The Science and the Stakes
|
Following four deaths at the 2026 Cheltenham Festival and 24 equine fatalities in UK competitive racing in the first two and a half months of the year, the question of what technology can do to reduce racehorse deaths is receiving renewed attention. One of the most promising and most scrutinised areas is wearable biometric sensor technology, which is the subject of a major industry-funded research programme that concluded its data-collection phase in December 2025.
The Promise
The central hypothesis of wearable sensors in equine welfare is straightforward: horses show subtle physiological changes before catastrophic injury becomes visible. A horse with a developing stress fracture may show changes in stride symmetry, ground contact time or heart rate that deviates from its established baseline days or weeks before the injury would manifest as lameness or, in the worst cases, a catastrophic failure on the racecourse.
If sensors can reliably detect those deviations in real time, during training gallops or workouts, and if that data can trigger a veterinary examination and a decision to rest or withdraw the horse before the stress becomes catastrophic, the injury and the death can be prevented. The goal, as described by the American Association of Equine Practitioners (AAEP), is “early warning systems for the identification of racehorses at increased risk of musculoskeletal injury.”
This is not a hypothetical proposition. Research led by Professor Chris Whitton at the University of Melbourne, working with Tasmania’s StrideMASTER GPS system, deployed on every horse racing in Tasmania since 2010, has established that monitoring “speed and stride characteristics over time may have potential” for identifying injury risk. The limitation is that research has not yet validated a system to the level of regulatory certainty required to compel intervention.
The AAEP Study
In 2025, ten Thoroughbred industry organisations contributed $785,000 to fund the AAEP’s wearable biometric sensor research project. Six sensor manufacturers, including Garmin, Arioneo, StrideSafe, Alogo Analysis, Equibase/Stable Analytics and Equimetrics, collected data from 200 two-year-old Thoroughbreds wearing their devices during high-speed workouts throughout 2025.
The study’s primary research question was whether these sensors could detect changes in movement patterns that correlate with elevated musculoskeletal injury risk, specifically, whether any of the sensors or combination of sensors could reliably identify horses that subsequently suffered injury. With data collection completed in December 2025, analysis is ongoing. Publication of findings through 2026 is expected to significantly advance (or constrain) the regulatory case for biometric sensor use.
What Sensors Currently Measure
Current equine wearables capture:
Heart rate and HRV(Heart Rate Variability): Changes in HRV are an early indicator of physiological stress. A horse whose post-gallop heart rate recovery is trending slower over weeks may be under greater cumulative stress than its visible behaviour suggests.
Stride symmetry: Any asymmetry between left and right limb loading, detectable in ground contact time measurements, can indicate sub-clinical lameness developing in one limb.
Stride length and frequency: Shortening of stride length at a given speed suggests fatigue or protective shortening related to discomfort.
Acceleration patterns: Abrupt changes in acceleration during a gallop may indicate that a horse is compensating for pain.
The challenge is specificity: many of these parameters vary naturally between horses, between tracks, between weather conditions and between individual training sessions. A sensor that triggers concern every time a horse’s heart rate recovery deviates from its average would generate too many false positives to be usable in practice. The statistical challenge is distinguishing signal from noise in a complex biological system.
Current Adoption in British Racing
Equine wearable technology is increasingly used in training yards across Britain and Europe, but its use in regulatory contexts, where data could compel a horse’s withdrawal from racing, does not yet exist in formal form under BHA Rules.
Industry estimates published by The Plaid Horse in February 2026 suggested that more than 40% of premium racing and breeding operations in Europe and Australia now use some form of biometric monitoring during training. This represents doubling of adoption in under five years, reflecting both falling sensor costs and growing trainer confidence in the data’s utility for optimising conditioning programmes.
The gap between training-tool use and regulatory use is significant. Trainers who use sensor data to inform their own decisions about whether a horse is ready to race are making professional judgments within their existing responsibility. A regulatory framework that used sensor data to override trainer or owner decisions, withdrawing a horse from a major race based on a sensor reading, would require validated statistical thresholds, clear liability frameworks and regulatory authority that does not currently exist.
The Welfare Argument
In the context of four deaths at the 2026 Cheltenham Festival, the welfare case for accelerating sensor research and regulatory frameworks is straightforward. If the technology can identify horses at elevated risk and prevent even a fraction of fatalities, the welfare benefit is substantial. The RSPCA’s call for “urgency” in learning lessons from equine deaths aligns with the direction of travel in sensor research, even if the specific regulatory application remains several years away from implementation.
The AAEP’s Dr Sara Langsam described the sensors’ core promise in terms that are directly relevant to fatality prevention: “Every horse has a distinct, patterned way of travelling. When the horse deviates from their movement ‘footprint,’ biometric tracking will detect even the slightest changes”, potentially before those changes manifest as visible distress or injury on the racecourse.
Whether that promise is realised and whether the racing industry is willing to accept the regulatory and commercial implications of sensor-compelled withdrawals at major meetings, will depend significantly on what the AAEP’s 2025 research findings reveal when they are published.
