This mathematical approach could be a great addition to the anti-doping arsenal.
This mathematical approach could be a great addition to the anti-doping arsenal. (Photo: Oli Scarff/Getty)
Sweat Science

How to Catch a Blood-Doping Marathoner

A mathematical approach to flagging suspicious race times shows its worth

This mathematical approach could be a great addition to the anti-doping arsenal.
Oli Scarff/Getty(Photo)

Heading out the door? Read this article on the new Outside+ app available now on iOS devices for members! Download the app.

On a blustery, rain-spattered fall day in Frankfurt several years ago, a 32-year-old runner from a small northern European country* pushed through to a personal-best marathon time of 2:17:31. While his performance was unremarkable from an international perspective, it was more than three minutes faster than his previous best—and good enough to earn him selection to the following summer’s European Championships. But it also triggered an alarm.

Unbeknownst to the athlete, his country’s national anti-doping agency had launched a pilot program to test the idea that unexpected jumps in athletic performance could signal a need for targeted drug testing. (I wrote in more detail about the theory behind this approach earlier this year). The racing results of the country’s middle- and long-distance runners were tracked to produce an Athlete’s Performance Passport (APP), an analog of the Athlete’s Biological Passport (ABP) that had recently been introduced to look for suspicious aberrations in blood test results.

No two races are the same, but when conditions are similar, elite endurance runners tend to produce performances that vary from race to race by 1 to 1.4 percent, according to one study. Based on this athlete’s recent results, his run in Frankfurt was 4.22 percent faster than expected. Back in his home country, the anti-doping agency scrambled its targeted testing team. Four days after the race, the athlete was tested; seven days after that, he was tested again.

The math behind the APP relies on a concept called critical power, which determines (in theory, at least) the speed you’re able to maintain for a given distance. The procedure is laid out in a new paper in Frontiers in Physiology by Sergei Iljukov, an anti-doping expert at the Research Institute for Olympic Sports in Finland, and his colleagues.

If you have timed performances at two different distances, you can use critical power to estimate what you should be able to run for any distance. The anti-doping agency had been tracking the athlete for 239 days leading up to the marathon, regularly updating their predictions based on his most recent race results. When he ran a then-best time of 2:20:50 earlier in the year, it was just 0.22 percent faster than the prediction of 2:21:08 based on a previous half marathon and a 12K result. But the 2:17:31 was well beyond expectations.

At this point, you’d be forgiven for being a bit skeptical. A 2:20 guy runs 2:17 and you’re sending in the doping police? Seriously? “Excellent performance itself is not a proof of any wrongdoing or doping,” Ilyukov and his colleagues point out. “However, through longitudinal monitoring, inconsistently excellent performance could be a warning sign that needs further attention from anti-doping authorities.” So the 2:17 triggers a pair of tests. And sure enough, the tests show “clear features of blood doping.”

Below is a picture of what that looks like in the biological passport, assembled from repeated blood tests. There are several parameters you can look at, but the OFF score is a number that depends on your levels of hemoglobin (the oxygen-carrying component of red blood cells) and reticulocytes (immature red blood cells), which has a characteristic response to withdrawing or reinfusing blood or taking blood boosters like EPO.

Normal OFF scores tend to land between about 80 and 110, but in the case of elite athletes, “normal” isn’t a very useful concept. Instead, doping authorities use a Bayesian approach to deduce reasonable upper and lower limits for each individual athlete, based on their previous scores. The thresholds for the first test are simply based on the general population average, but with each test, the athlete’s individual thresholds evolve.

Here are the athlete’s OFF scores during the period of interest (Day 1 was when the APP started; the 2:17 marathon was on Day 239; the two targeted tests were on Days 243 and 250):

(Frontiers in Physiology)

The red lines indicate the upper and lower thresholds, reflecting a one in 1,000 chance of a false positive. The blue line is his measured OFF scores. The key point here is that the blood test taken four days after the marathon exceeds the upper limit, suggesting a reinfusion of stored blood.

In fact, the whole pattern is pretty suspicious, because the initial high values correspond to some good races, and then he has a period of very low OFF scores, down near the lower one-in-1,000 threshold, which would be consistent with extracting blood for later reuse and also happened to correspond to some unusually poor races. Then his OFF score shoots back up and he runs his 2:17:31.

A year and a half after this athlete’s marathon, based on this data, the IAAF formally accused him of doping, which he initially denied. His results from those years were annulled, and he then served a doping ban for another two years. According to the new journal article, he subsequently admitted to doping.

This is a pretty cool example of the performance passport approach working. The big question, for me, is how many false positives the system triggered to produce this one true positive. If 50 percent of the athletes enrolled in the pilot project had an “inconsistently excellent” performance at some point, then it’s not a sustainable or scalable solution. I know I certainly would have triggered the alarm a few times during my running heyday (and no, I didn’t dope). But if the alarm rate is low enough, this could be a great addition to the anti-doping arsenal.

The other thing that strikes me about this is sadder. This guy started doping as a 2:20 marathoner (or perhaps even slower than that, since we don’t know when he started doping). The problem, then, isn’t just about athletes getting corrupted by potential riches, since a 2:20 guy is a long, long way from getting rich. That suggests that, much as I hate to admit it, going back to Roger Bannister–era amateurism wouldn’t eradicate doping from sport. So let’s demand that international sports organizations give people like Iljukov the funding and support they need to keep fighting the good fight.

*Several identifying details about the case have been withheld at the request of the research team to avoid compromising ongoing anti-doping efforts.

My new book, Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance, with a foreword by Malcolm Gladwell, is now available. For more, join me on Twitter and Facebook, and sign up for the Sweat Science email newsletter.

Lead Photo: Oli Scarff/Getty

When you buy something using the retail links in our stories, we may earn a small commission. We do not accept money for editorial gear reviews. Read more about our policy.

Trending on Outside Online