The continuing occurrence of rotational falls in eventing, despite the sport’s implementation of safety devices and measures, has led an engineering student to develop a series of statistical models that better predict the likelihood of such falls.
University of Kentucky College of Engineering 2020 graduate Shannon Wood, MS, said the results from her research could be used to help course designers determine the appropriate safety devices for the fences on a cross country course. The information could also be used by safety device designers to better optimize their effectiveness.
Rotational (or somersault) falls are the leading cause of death and serious injury in the equestrian sport of eventing.
“Horses have always been my passion and increasing horse and rider safety through data-validated means is what motivated me to work on the rotational fall problem,” said Wood, who completed the work under Suzanne Weaver Smith, PhD, professor of mechanical engineering at UK.
The aim of the research was to create a statistical ensemble method¹ for evaluating the physics of potential rotational fall situations, providing an alternative to physical testing dummies for determining indicators of rotation.
Wood was motivated to find solutions because of the continued occurrence of rotational falls, even after the sport implemented safety devices into cross country eventing jumps. She also wanted to rectify the lack of evidence-based methods for safety device testing criteria.
Wood began her trailblazing research in 2016, recognizing that it was important to capture the variety of horses, riders, jump positions and speeds approaching the jumps to better understand the nature of rotational falls in eventing. Previously, physical dummies based on a horse cadaver were used to simulate the effect of the horse hitting the jump.
Based on that existing research, frangible devices currently in place were developed to break under hard contact, resulting in the jump falling, in order to lessen the effects of — or even eliminate — a rotational fall.
“The big thing for my research, instead of using information from just one horse and the physical dummies based on that horse, was being able to use a statistical model that can generate 10,000 horses and riders. There is so much variety in horses, riders and positions, especially in eventing,” she said. “Additionally, current information used to create frangible pins didn’t account for the rider.”
Wood set out to incorporate the variability of these combinations, as well as the situations those horse and rider pairs would face on course, to ultimately create guidelines for the development of more frangible devices.
“Providing design options for course builders and designers is important, as there are only three types of frangible devices in use but over 40 varieties of cross country jumps combined into different ‘questions’ for the horse and rider pair,” she said.
According to Wood’s research, reducing risk to competitors is a multi-faceted process with numerous fronts for possible improvement.
“To provide insight and results for policy decisions and design guidance, each physics-based simulation looks at 10,000 cases of competitors with critical contacts, which is the equivalent of more than 62.5 years of ‘very bad days’. Reducing the 165 occurrences of rotational falls to 19 per year or less is achievable without changing the culture of the sport. This represents one in 1048 starters (0.095%), half the rate of 2015,” Wood says in her thesis.
“Not all rotational falls can be eliminated, though. Simulation results show 2.2% of the critical-contact situations for one-size-fits all cases can’t be mitigated by jump safety devices or designs. Prevention of rider injury in these cases would rely on personal safety protection for three to four per year.”
Her thesis refers to a “Swiss Cheese Model,” or system failure model that has been applied to many high-reliability settings such as medicine, nuclear power and aerospace systems. A severe injury from a rotational fall only occurs if several conditions line up.
According to the model, such an injury is prevented by layers of safety, including training, qualifications, sport rules, course and jump design, among others, to prevent what could lead to the horse and rider ending up in a situation where they make contact with the fence in the critical foreleg region associated with rotational falls. If critical contact with the jump does occur, the mitigation layers reduce the risk of a rotational fall through the action of fence safety devices. Finally, if that fails, individual safety technology such as inflatable vests and helmets react to minimize the consequences of a rotational fall.
Wood points out that her study is meant to help with mitigation rather than prevention. This means that safety devices lessen the consequences of horses and riders in dangerous situations rather than completely keeping competitors from problems.
At the outset of her research, Wood said there was very limited information about the physics of rotational falls and their contributing factors. For instance, there was no known center of gravity (CG) and inertia models for a horse and rider jumping. Wood and Robles Vega, MS, developed those. There were two published epidemiological studies to provide context to root causes of rotational falls for Wood to reference.
The speeds and jumping positions of competitors at various cross country jumps were also unknown. Wood video recorded 218 competitors approaching 10 different jumps on cross country courses in competitions ranging from preliminary to CCI5* level, yielding horse and rider jump configuration angles for different fence types. Through this process, she created a statistical ensemble tool using impulse-momentum methods to identify rotating conditions and design evaluation. This information can help in the design of future jumps and safety devices.
Before her study, it was unknown what force, in actuality force-time impulses, the jump safety equipment should activate for. It was unknown if the device would activate when it “should” and, if by activating, it prevented rotational falls. Wood aimed to answer these questions as part of her research.
For an accurate statistical representation of the horse and rider inertia distributions, Wood’s research began with measurements of 429 training or competing horses for inertia distributions, which included riders’ heights and weights, recognizing that the variety of sizes of the horse and rider combo was important to account for.
In planning her research, Wood took an alternate approach from using the physical dummies based on a horse cadaver that had been done in previous existing studies. Instead, a statistical ensemble model was used to evaluate 10,000 different situations with horse and rider that might potentially lead to rotational falls.
Combining information for these, among 26 total variables, a statistical ensemble simulation using impulse-momentum physics identified conditions for rotation and defined design criteria for future general and situation-specific jumps and safety devices.
A Jump Safety Quality Index was also devised by Wood to represent the benefit of incorporating a safety or frangible fence design for mitigating rotational falls, compared to a regular fixed-fence, as opposed to the detriment and competition penalties of false activation.
All of the situations modeled represented horses contacting the jump with their forelegs. That is a “very bad day” type situation and didn’t include clear jumps or incidental hoof strikes.
Wood says the results of the model created in her study tell the percentage of no-rotation competitors in critical foreleg contact with a fixed fence and for a fence with a prescribed safety device. It is a tool to evaluate the likelihood of a rotational fall and to determine an appropriate safety device and its effectiveness and can be applied to a particular fence.
“For instance, using body positions and speeds from fence 4A, a vertical at the 2018 Land Rover Kentucky Three Day Event, for simulated situations of foreleg contact, the model predicts that 64.9% of horses contacting the fence with their upper foreleg will not rotate. After adding a device that would activate and limit the impulse to 900 N*s within a -10 to 25 degree activation range, 96.4% of competitors would not have a rotational fall,” she said.
“The model also estimates false activations, or when the safety device would activate but no rotational fall would occur. This is important for the culture of the sport and competitors since an 11-point penalty is applied in FEI competitions. These predictions can be done for particular fences or for attempts to group similar jumps together, like with post and rail oxers on similar terrain,” she said.
It was determined that if the center of gravity of the combined horse and rider (approximately near the rider’s knee) crosses the vertical plane above the contact point, a rotational fall will occur. The geometry can be used to understand the results of the physics-based simulations.
Wood feels there are several avenues that should be pursued in future research. One of those should include looking at the hind legs pushing off the ground while jumping and incorporate that with the existing model.
An important future step toward safety would be the implementation of widespread, static (non-panning) video recordings of fences on course for review purposes. Follow-up incident reports should also be completed by riders to add information to safety and accident conditions for future epidemiological studies.
“Over 700,000 fences are attempted in FEI classes, but there’s not a static video camera at the fences. So when something happens, there’s limited reviewable information. Having that record of how each jump was navigated could be beneficial to course designers and other professionals interested in eventing safety,” she said.
“It wouldn’t be terribly expensive. Consider 45 GoPro cameras outfitted, then ‘record’ pressed on remotes by jump judges and turned off when there’s no activity. I estimate it would cost about $15,000 for equipment. However, data processing would cost more. But, so many are calling for transparency on the 15-penalty flag rule and using the cameras to evaluate the rules and safety could be dual purpose,” she said.
Wood also recommended conducting follow-up e-mail “fall” or accident report forms for riders after they are injured to learn more.
Wood’s master’s thesis, titled Safety concepts for every ride: A statistical ensemble simulation to mitigate rotational falls in eventing cross country is provided free via open access by the Mechanical Engineering at UKnowledge. It has been accepted for inclusion in Theses and Dissertations — Mechanical Engineering.
1 Definition of a statistical ensemble: A statistical ensemble is an idealization consisting of a large number of virtual copies (sometimes infinitely many) of a system, considered all at once, each of which represents a possible state that the real system might be in.
Article courtesy University of Kentucky College of Agriculture, Food and Environment.