At a press conference on April 21, the American Association of Equine Practitioners (AAEP) shared results from a first-of-its-kind study evaluating wearable biometric sensors for musculoskeletal injury prevention in racehorses, which could shape how trainers and veterinarians monitor horses in real time. 

Speaking on behalf of the research team, Sara Langsam, VMD, chair of the AAEP’s Racing Committee and sensor project lead coordinator, emphasized this was the first project to test these technologies prospectively, in a real-time setting. They focused on 2-year-old racehorses because they were all at a similar point in their careers. The goal was to determine whether wearable sensors could reliably identify horses at risk of injury before clinical signs appeared. 

To be considered viable, Langsam explained, the devices needed to meet several criteria—they had to transmit data efficiently, be scalable for widespread industry use, and remain user-friendly enough for adoption across entire racing populations. Six companies were ultimately selected to participate. However, incomplete datasets led to the exclusion of two companies. The researchers anonymized all data used in the analysis. 

Peta Hitchens, BAppSci, MVPHMgt, MFTV, PhD, associate professor at the University of Melbourne, in Australia, outlined the study’s structure and findings. Each horse was required to wear a sensor during every officially recorded breeze, as tracked by Equibase. If a horse missed two recorded breezes, the researchers required a veterinary exam and logged outcomes—ranging from green (no concern) to red (high concern)—within 48 hours. 

The dataset ultimately included 4,252 breezes from 561 horses, with a median of 11 breezes per horse. Researchers also analyzed 2,543 veterinary exam reports alongside detailed training and racing records, including speed, distance, timing between breezes, track conditions, and prior injury history. 

Across the study period the researchers identified 221 distinct injuries across 181 horses, including both bone and soft tissue injuries. Notably, injuries most often followed an unofficial (not recorded) breeze (77%), compared to official breezes (17%) or races (7%), and typically occurred about 10 weeks into training. 

The data revealed clear risk patterns. Horses were more likely to sustain injury if they had a prior injury, received a yellow or red flag in their most recent sensor reading, accumulated multiple red flags over a 90-day period, or were early in training. Unofficial breezes also carried greater risk than official ones. Hitchens said this is likely another indicator of being early on in their training, not that breezing in an unofficial breeze is inherently riskier than in an official breeze. 

Hitchens noted that sensor performance was relatively consistent across companies, suggesting the technology itself holds promise. However, limitations remain, such as delays in data reporting, the need for dedicated personnel to track horses, and incomplete injury records. 

Hitchens said the take-home message is significant: Wearable biometric sensors might offer a practical, scalable tool for identifying at-risk horses earlier, potentially allowing for intervention before musculoskeletal injuries occur. In the future of this project, researchers hope to further refine the algorithms to ensure early detection rather than late-stage prediction and validate individual biometric sensors to reach sufficient sensitivity and specificity for implementation by the racing industry.