Study identifies high-risk factors for track fatalities

LEXINGTON, Ky. – Horses who are at a greater risk of suffering fatal injuries are becoming easier to identify because of the isolation of several high-risk factors gleaned from a six-year-old injury database, according to the epidemiologist analyzing the data for the U.S. racing industry.

Analysis of the Equine Injury Database has led to a national model assessing risk factors for North American horses and the creation of eight additional “track-specific” models, according to Dr. Tim Parkin, an epidemiologist at the University of Glasgow. The national model has turned up several risk factors that can lead to a twofold or threefold jump in the likelihood of a horse suffering a fatal injury, Parkin said.

As a result of the isolation of the high-risk factors, Parkin and several national racing organizations, most notably The Jockey Club, are working on plans to communicate those risk factors to racing offices and trainers in the hopes of preventing higher-risk horses from starting before it is too late. The Jockey Club is expected to announce at its Round Table conference in August the creation of additions to its racing-office software tools that would generate red flags on higher-risk horses at entry time, according to Jockey Club officials.

“We have enough data to be confident in these models,” Parkin said during a presentation at the Safety and Welfare of the Racehorse Summit at Keeneland on Wednesday. Parkin stressed, however, that many local factors can override data trends in national models, and he said the goal would be to produce local models for as many tracks as possible.

Currently, the database contains records of 2.2 million starts over the past six years, representing 94 percent of the starts in North America from 2009-14. Participation in the injury database project is voluntary.

The eight track-specific models were created because the tracks had generated enough data to create statistically significant conclusions, Parkin said. The tracks were not identified.

One high-risk factor to have emerged from the most recent data analysis is whether a horse had suffered an earlier injury that had been entered into the database, Parkin said. With each injury logged into the database, the horse was 30 percent more likely to suffer a fatal injury than the average horse, according to Parkin.

Other risk factors included a horse being relatively new to a trainer’s barn; a layoff of only 60 days, as opposed to 120 days or more; the placement of the horse on a veterinarian’s list at any time in his career; and the age at which a horse starts in its first race, with higher rates of fatalities for older debuting horses.

Statistically speaking, racehorse fatalities are relatively rare events, happening on average in the U.S. at a rate of 2 per 1,000 races. However, Parkin said the data have revealed that a small subset of the racing population, about 5 percent, has a much greater risk of breakdown than the average horse, leading to optimism that horses can be prevented from racing when they fit the high-risk criteria.

A drop in the overall fatality rate has been the ultimate goal for the Equine Injury Database, but the rate has held steady since the project was launched in July 2008. Several states and racetracks, however, have had significant success in lowering their overall rates – most notably Kentucky Downs and Finger Lakes in upstate New York – which has emboldened some to believe that an overall decline is possible through better management of horses and the identification of horses most at risk.

Parkin said that no system can predict the future with absolute certainty, no matter how large the database. However, he noted that the ability to identify risk factors will only sharpen as more data are collected.

“We’re getting there,” Parkin said. “We’re closer and closer.”