Sunday, October 21, 2012

Armstrong, EPO, and living in the real world

The whole world seems ready to pile onto Lance Armstrong this week, after the US anti-doping agency (USADA) released an extensive report supporting its claims that Armstrong and his team conspired to use the drug EPO and blood doping to "ultimately gain an unfair competitive advantage through superior doping practices." The narrative of the investigation has even inspired a potboiler article in the NY Times detailing the story of how one teammate after another testified to the culture of doping within the US Postal team, with all fingers pointing to Armstrong.

What's lost in this discussion, of course, is context. It seems clear from the testimony that Armstrong and US Postal systematically used EPO and blood boosting to increase the oxygen-carrying capacity of their bloodstreams. What isn't being reported, however, is the way many of Armstrong's competitors of the era had used the same techniques. At a deeper level, the laser-like focus of the USADA on Armstrong ignores an uncomfortable truth: banning a practice that can't be reliably detected puts competitive athletes in an impossible situation.

In the late '90s, cycling was rife with use of an artificial copy of the human protein EPO, which increases the number of red blood cells in the blood. Because we all have EPO circulating in our blood, detection of recombinant (artificial) EPO is difficult. When Armstrong was winning his first Tours de France, there was no test in use for EPO, and even when a urine test for EPO was introduced, it could only pick up recombinant  EPO for about four days after injection, while the effects lasted much longer. In short, there was no reliable way to catch cyclists who were using EPO. As a result, EPO use was rampant, as post-retirement confessions have revealed (for instance, Tour winner Bjarne Riis). The US Postal case further illustrates the pattern; riders on the team during Armstrong's winning run in the Tour were not caught in in-competition testing, but rather have confessed EPO use and doping under intense pressure long after the fact. Still not convinced? Here's a telling figure ("Tour riders with a doping past" - thanks to Alexey Merz) showing who Armstrong was competing against. Most of these riders were also not caught in competition but after later revelations.

In such an environment, an athlete faces a stark choice. If a banned practice cannot be reliably enforced, the athlete can reasonably assume that his or her competitors are using the practice. The athlete must either forsake the banned practice, putting him or herself at a disadvantage relative to the competition, or must take advantage of the banned practice to level the de facto playing field. In a sport like competitive cycling, a factor like the oxygen-carrying capacity of the blood is crucial, and a disadvantage there could literally separate winners from "also-rans." The justification for using the banned practice isn't just that "everyone else is doing it," but rather, that other competitors are doing it and not getting caught.

One point that has not been emphasized enough in the Armstrong case is that while there was no solid test for long-term EPO use during his career, the UCI had instituted another control that was easily measured, a maximum hematocrit of 50. In other words, there was a cap on how many red blood cells could be circulating in a rider's blood, whatever the cause. Suspensions were not punitive, but they prevented cyclists from competing with hematocrits above the threshold. The tests are straightforward, and the result was that cyclists boosted their blood to levels just under the threshold by a combination of training and, in all likelihood, EPO use and transfusions. While this is far from a perfect solution, in that it did not punish those that used EPO, it did have the effect of limiting the potential advantages of blood boosting by effectively putting all riders at roughly the maximum  hematocrit. The USADA statement that Armstrong was able to "gain an unfair competitive advantage" should be re-examined in that context.

The situation may be improving with the introduction of the UCI's Biological Passport, a long-term profile of cyclists' hematocrits and the levels of both "young" and "old" red blood cells (comparing those levels may reveal artificial disruptions from boosting or EPO). That program was begun in 2007, two years after Armstrong's last Tour win. If the Passport proves to be a reliable way to detect artificial blood-boosting, then athletes can compete without worrying that their rivals are doping; if there is an EPO-free future to cycling, it lies with reliable and effective testing, not with extensive investigations into past conduct focused exclusively on winning cyclists or teams. We can vent our indignation at Armstrong, and perhaps that makes us feel better, but it is not fair to expect a competitive athlete to behave differently given the situation of the early 2000's. We could not detect the cheaters, with the result that most of the others decided to cheat as well. To expect a different result is simply unrealistic.

UPDATE (7/23/13): This week the Economist pointed out a further wrinkle in the doping environment based on game-theory analysis. Buechel, Emrich, and Pohlkamp modeled the atheletes, the sanctioning bodies, and the fans and sponsors as agents, finding that doing only occasional testing as a kind of PR exercise (and reporting only the positive tests) created a public perception of a few "bad apples" doping in an otherwise clean sport, when in fact most competitors could be doping as a rational response to the limited-testing environment. The authors suggest that more frequent testing, with public disclosure of all results (both negative and positive) would dispel a false perception that doping was limited.

The results of the game-theory analysis are remarkably consistent with actual events, such as scorn heaped on particular athletes who test positive (Armstrong). It also jibes with the impressions of many that race organizers and sanctioning bodies would promote scapegoating narratives about those particular athletes to deflect criticism from their own practices (limited private testing), and the suspicion that those sanctioning bodies are not really interested in actually eliminating doping from the sport for rationally selfish reasons. The conclusions also predict the withdrawals of sponsors in the face of explicit doping, including the withdrawal of Rabobank from the image of its team in 2013 after allegations against Armstrong, with the Dutch bank continuing to honor contracts and pay expenses, but withdrawing its name, logo and colors from team jerseys, which were re-styled as "Blanco Pro". The T-mobile team, riddled with doping allegations, was likewise purged of riders who had tested positive and re-introduced as HTC-High Road to shed the stain of doping on the organization.

Perhaps sponsors and fans can push the UCI and race organizers to be more persistent, dogged, and transparent about testing regimes; Buechel's analysis suggests that is the only way forward that aligns the rational interests of racers with a doping-free race environment. After all, the existence of pro cycling depends completely on fans and the sponsors who want to look good for those fans.