A reporter asked me recently what I thought the next breakthrough in sports medicine would be. Maybe it’s the fact that I majored in economics in college instead of biology, chemistry, or physics, but my answer might surprise athletes and sports fans. I believe that statistical models and injury analytics are the next frontier for professional sports teams and their physicians.
In 2003, Moneyball, the best-selling novel by Michael Lewis, drew both immediate praise and ire. Lewis described the adoption of statistical analysis to scout players by Oakland A’s general manager Billy Beane and his colleagues. Scouts using terms like “five-tool player” and “level swing” were replaced with statistics such as on-base percentage and slugging percentage.
There is much debate as to the success of these methods of analysis to shape team rosters, as baseball purists quickly note. After all, the A’s have not won a World Series during Beane’s tenure as general manager (Their payroll is only a fraction of that of many large-market teams, to be fair). There is no question that this idea is catching on, though, as teams throughout baseball now employ data analysts, and these approaches seem to be entering the NBA and English Premier League.
What do these models have to do with sports medicine? Everything, potentially. Consider a statistic from a recent article in ESPN the Magazine by Molly Knight. Between 2007 and 2011, the total of all the salaries paid to Major League Baseball players reached $13.5 billion. $2.1 billion of that amount, or approximately 15%, was paid to players on the DL at the time.
Part of the challenge for team general managers in signing players who can produce on the field lies in guessing which of those players will actually remain on the field. Talent scouts can try to predict on-field success, but the team relies on their team doctors and athletic trainers for a judgment about whether the player is currently healthy but also a prediction that the player will stay healthy for the length of the contract.
As team physicians, we review medical records, ask players about current problems, thoroughly examine their bodies, and review x-rays and MRIs. Then we make the best decisions possible. Even with all of that information, predicting who will stay healthy and who will get hurt is uncertain.
Statistical models and injury analytics in professional sports?
Podcast: Play in new window | Download
Click here for full episodes or subscribe on iTunes.
For instance, why has Amar’e Stoudemire played many successful NBA seasons after his microfracture knee surgery in 2005, while 2007 #1 NBA draft pick Greg Oden has only played in 82 games in his entire career after the same surgery?
2006 NL Cy Young Winner and 2008 runner up Brandon Webb was considered one of baseball’s most dominant pitchers heading into the 2009 season. Despite no sign of shoulder trouble throughout his career, he has pitched only four innings since the start of that 2009 season, having reportedly undergone two rotator cuff shoulder surgeries.
I know as well as anyone that injuries in sports can often be simply the result of bad luck. Errant foul balls, dangerous tackles, and so many other freak incidents can injure star players. Injuries can and do happen all the time. If we could better predict who is more likely to get hurt and who is more likely to struggle to overcome injury or get reinjured, the teams, and the sports in general, win.
For instance, we know that football players are often cleared to return to play six months after ACL surgery. However, only about 70% of college players actually return. If we had models that could analyze data and better predict whether the injured player will be in the successful 70% or the struggling 30%, the team could better prepare for the future.
Is using statistical analysis to improve our medical decisions regarding signing players unrealistic? It would require a tremendous amount of information, such as age, prior injuries, position played, practices and games missed, and so much more. But leagues are already starting to collect this data. Maybe I’m overly optimistic. But even more than any new surgical technique, cell therapy, or rehab protocol, I think injury analytics could become the future of sports medicine.
Tweet about injury statistics.
I want to hear what you think? Do you think statistical analysis of player injuries will become a reality for professional sports teams? Share your thoughts!
Note: This post appears in my sports medicine column in the March 15, 2012 issue of The Post and Courier.
