It’s the London Marathon this week. It will almost certainly be won by an athlete from east Africa. In both the men’s and women’s races. After all, it’s been 16 years since a non-east African won the men’s race, and nine years on the women’s side. Other major big city marathons show a similar concentration of dominance. Even if Mo Farah wins in London on Sunday, causing the BBC and large swathes of England to self-combust with smugness and congratulatory cognitive dissonance, it’ll count, at least to the physiologist in me, as a victory for the “predisposition” of east African runners for distance running success.
That dominance of east African runners is unparalleled, and it is one of the most fascinating topics to discuss for a physiologist or performance analyst. Countless papers have been written trying to describe it, hundreds of research studies conducted to find its origins, and thousands of athletes have visited Kenya and Ethiopia to ‘absorb it’ in training camps (and for a few other reasons, as we’ll see).
I’ve even written a few myself. One example, published in 2015, is what I thought was a very interesting paper looking at the Kenyan Distance Running Phenomenon, where my colleagues Vincent Onywera, Jordan Santos and I looked at the contribution made not just by east Africans, or Kenyans, but specific Kenyan tribes and sub-tribes, to the global distance running picture over four decades. This was the paper:
This was one of our key graphs, showing how Kenyans emerged in both track and marathon running to dominate the last decade, and that this dominance was really concentrated in the Kalenjin tribe specifically:
And finally, one of the key points in the paper, one that I thought was really interesting:
We published a few other papers on Kenyan runners, including a batch of studies where we looked at brain oxygenation, biomechanics, and neuromuscular function in elite Kenyan runners. So yeah, I think it’s fair to say that I really love this topic. I’m fascinated by the physiology, the mechanics, the system of Talent ID (which doesn’t look like much if you view it through a Western paradigm, but it’s there), the ‘soft’ and hard factors explaining their successes.
I also believe them to be true – I think THERE IS a physiological basis for the concentration of east African/Kenyan/Kalenjin/Nandi runners. I believe that THERE ARE legitimate biomechanical advantages that are more likely to be found in these populations than elsewhere, and which explain their over-representation. In turn, I believe that there are principles and concepts that studying east African runners can teach the world about being better runners.
“Take nothing away from them, but…
But there’s a confounder that you simply cannot ignore unless you’re in total denial – doping. Some days, especially lately, where the drip-feed of doping accusations and confirmations (the latest being Asbel Kiprop) is intensifying into a steady stream, I find myself thinking ‘Shit, I wonder if all those explanations and arguments are nothing more than evidence of systemic unregulated doping, and have I naively (or knowingly, even) contributed to that fairytale?’.
Some would say that the presence of doped athletes takes nothing away from the performance of others. This is disingenuous, and wants to pretend there is an alternate reality (you see it in cycling, especially among Team Sky fans, a lot).
Of course it does! If a handful of athletes, especially from one training group, have been popped for EPO every few months for the last five years, but are still running slower than Athlete X, then the performances of Athlete X are absolutely less trustworthy than they were without the knowledge of ‘linked dopers’ or people known to facilitate doping (doctors, managers, team directors etc)!
That’s not to say that X can be declared guilty by association – doping is a bit more complex than catching a cold from someone just because you share a house with them. But the performances of all athletes are tainted, open to doubt, and thus diminished by the doping of a few. This, in my opinion, should be the call to action to try to fix the problem, not to ignore it.
Science of course never goes there. Too speculative for “us” to suggest such a thing.
So you won’t find many scientific publications that include in their conclusions “We cannot guarantee that the findings of this study and the documented dominance of east African distance runners are not the result of widespread systematic doping that has to date been relatively undetected because of the inaccessibility of remote training locations and a sophisticated system of doping used by doctors and agents working with athletes in these communities”, that’s for sure!
Yet when you look at the history of sport, then the mix of elements that is required for doping most certainly does exist in east Africa. Allied to the late emergence of talent from these countries (think how many athletes come out of nowhere late in life), large improvements late in their careers, you have a team environment facilitated by central authorities often with dodgy pasts, along with sufficient economic incentives to ‘escape’ a set of circumstances. It’s all there. Add the fact that altitude makes detection of doping difficult using the ABP, plus micro-dosing that helps avoid all testing methods, and you actually have good reason to make that paragraph above part of every discussion.
It’s nothing more than pattern recognition. If you see A, you might leap forward to Z. But that’s a big leap. But if you see A, then B, then C, then D, and so on, then even if the order is a little different, or if five are missing and 15 are present, well, it’s rather more reasonable than Z comes at the end of the sequence, no? Whether you eventually see it or not. That’s why people doubt the integrity of performances in all sport, including these.
The known unknown – scale of the problem
So that’s why we need to talk. Well, I’ll write, then you/we can talk. I have no answers on this issue, and I’m not going to reach a conclusion now, but the reality is that doping confounds every physiological, genetic, biomechanical, evolutionary, socio-cultural or economic reason ever put forward to explain why east Africans are the best runners, and so when an athlete from east Africa breaks the tape in London on Sunday, in a time of 2:03:47 or 2:17:13, or whatever, what are we to think? Are we seeing another athlete who has hidden away at the right times, for long enough to dope themselves two minutes faster, or we seeing a legitimate performance? Beats me.
But there does come a point where one can’t simply say “It’s only a few isolated cases”. And that point was passed some time ago. It’s in the 138 athletes from Kenya banned since 2004. It’s in the Rita Jeptoos, Matthew Kisorios, Jemima Sumgongs and most recently, Asbel Kiprops. And those are just the highest profile names who I could think of off the top of my head without doing any searches. They share a handful of major marathon titles, Olympic and World gold medals, and plenty of historic performances among them. There are hundreds more ‘next-tier’ athletes who have also been caught, and probably many more who have not (such is the reality of anti-doping – it has a low sensitivity).
Ethiopia, I have no doubt, will be similar. The difference may be scrutiny. If the spotlight is bright enough, then even small blemishes may appear, let alone the large ugly scars in the performance history. Ethiopia remains, relatively speaking, in the anti-doping darkness, whereas Kenya has been subject to an intensified targeted effort to deter doping in recent years. This is why athletes who used to disappear into Kenya for training now go to Ethiopia – the appeal is not simply the altitude, the company, the training environment, it’s the isolation and separation from anti-doping authorities.
And here, to repeat a point I made in my most recent article about Kiprop, the stark reality is that it’s not really that difficult to dope even when you’re about to be tested. Look at what an anti-doping expert said at the Kiprop hearing:
So it’s possible to take a banned drug and have no fear about failing a test within about 3 days. That’s the window. Make no mistake, the relative isolation of training camps, which drastically reduces the possibility of being tested in that narrow window, is a big deal for doping. This is especially true if you know the testing is coming (ahem, Ethiopia…)
Any reason for confidence?
On what basis might one have some confidence? Well, the Athletics Integrity Unit (AIU, who I really think are doing some very good stuff) have intensified their focus on “intelligence gathering” as a means to better target test athletes, and that’s helpful. Their approach is to identify a pool of 150 elite runners, and then to split them into high risk and low risk groups, based on things like:
- Their biological passport profile
Those are the ones mentioned by the AIU, along with “other factors”. It must be acknowledged that this is not a new approach – remember the UCI’s suspicion index? That was basically the same thing, though perhaps a little less refined by the performance analysis. A couple points on these:
I think the ABP can be (and probably has been, though we don’t always know when) a very useful guide to target testing. It would flag up patterns of blood values that are not quite outrageous enough to pursue a doping case, but suspicious enough to warrant intensive testing. The problem is that even intensive testing might not be enough to actually find the drug, if you go back to that “window of opportunity” concept mentioned above, where the window might be “open” for only three days. At the extreme, if an athlete was tested every single day by a trustworthy source, then you could have confidence. But at 3 days, your confidence in a “clean sample means clean athlete” paradigm has dropped considerably, maybe even entirely.
What is it about performance that suggests doubts?
The performance profiling bit is interesting. Since 2007, that’s what I’ve (along with a few others) tried to do in cycling, when I first looked at the Contador climb of Verbier (well, I did this until it got boring and circular because no authorities in cycling were willing to take the steps required to generate trustworthy data to allow long-term comparisons, or to actually do anything about it).
The same concepts can be applied for running. The challenge is separating the signal from the noise – there are so many factors affecting performance that actually understanding what a time means can be tricky. It’s actually easier for cycing, where you can measure power output, but for a marathon, how do you compare a 2:08 in New York to a 2:12 in Boston on a windy, cold day, to a 2:03 in Berlin on a perfect day?
Then add in the relative rarity of high quality races per athlete per year (2, maybe three, compared to maybe 10 cycling performances in just two GTs a season) means you have too few data points among elite marathon runners to be confident in their meaning, especially given the other confounders. Ultimately, performance in running is quite a lot more difficult to interpret than in cycling.
Nevertheless, I reckon three high-level attributes of performance would contribute to the classification of an athlete as high risk or low risk, irrespective of sport:
- Anyone who runs a time in the top 50 in history, or who comes top 5 in a big city marathon, is automatically high risk. This is simply what you learn from history, and it’s logical. It’s pattern recognition, once again. Problem is, it’s non-specific, and hardly filters out any athletes – basically every one of those 150 elite athletes could be in the high-risk group within two years.
- Any major improvement in performance, with a huge increase in suspicion if that improvement happens more than about three years into the athlete’s career, must be viewed as highly suspicious. If I see a 28 year old with an 8-year career suggesting the capability run a sub-2:04 in their seventh marathon, after a bumpy or smooth progression of a minute or two over five or six marathons in three years, that’s going to be a lot less suspicious than a 28 year old who has been stuck at 2:10 in six marathons over three years, but who then runs 2:04. Similarly, a 27 year old woman who runs a 2:18 or a 66:00 half marathon, with no previous notable performances is actually less suspicious than a 27 year old woman who runs 2:18 or 2:19 having run for the previous five years and never broken 2:23 or 68 min. It’s the change that counts most, and the age at which change happens. Here again, the presence of this performance doesn’t prove doping, but wow, it’s a huge red flag, given how physiology and performance normally progress.
- Erratic performance. Just as a sudden improvement is a flag, especially in a runner who is older and who has been around a trapped at a level for year, so too is a regression in performance. Some runners appear, then disappear, and that may be as indicative of doping as only appearing. Indeed, one of the hypotheses you’d make when a more comprehensive anti-doping programme is introduced in a certain territory is that performances will decline as a result of the spotlight. Russian athletes provide a good illustration of this. It is also what papers by researchers such as Sergei Iljukov and James Hopker have recognized would happen – introducing stronger anti-doping “interventions” creates an opportunity to assess how influential doping might have been in hindsight (much like the current spike of measles cases around the world strongly suggests that a) vaccines work, and b) anti-vaxxers are total bellends. Anyway, I digress. Back to sport, one way to avoid being erratic, by the way, is to race only three or four times a year. Because that way, you can remove yourself from the window, and do your doping thing without concern, and appear only when it is safe. So erratic performances and very rare appearances COULD mean a similar thing.
So those three things are the performance criteria that I reckon should put an athlete in the ‘high risk’ group. It can probably be sub-divided even further to add to resolution, but you have to be careful not to make it so “HD” that you never pick up the signal because you measure too much noise. And to repeat, none prove doping, but if I see an athlete winning London who meets criteria 2 or 3 above (they automatically meet Criteria 1, by definition, which is the problem with that one!), then I’m less trusting than if they at least have a history and some “longitudinal credibility”.
But what are the other factors that might put an athlete into high or low risk?
Other factors informing “intelligence”
Here, you look for the afore-mentioned pattern recognition. What patterns or behaviours or attributes are known to be associated with doping? An athlete who associates with a coach implicated of or known to have doped athletes is a massive red flag. An athlete who denies knowing that coach and who pretends a relationship does not exist makes that flag even larger (doing something is bad, hiding it is worse).
An athlete who fails whereabouts tests, or who misses samples, that’s a big problem, because of what I wrote above about the window of opportunity for doping control and detection. Trouble is, we don’t know this until it is leaked, which is where the lack of transparency once again hurts credibility (though it protects potential cheaters, and therein is the problem).
An athlete who is represented by an agent who has had numerous run ins with doping, including many cases, is a huge reason for doubt. The “entourage” argument is always tricky, because as mentioned, guilt by association is not fair. But skepticism by association is very normal, and probably necessary. In cycling, for instance, the continued involvement of doctors and managers who have been known to have doped themselves, or have doped other riders, makes it rather more difficult to believe that the present “new generation” are clean.
And finally, informants. An athlete about who multiple stories exist of doping, or running away from doping control officers and jumping a wall to escape, for tip-offs about a supplier who provides EPO to athletes in an area where the drug is not exactly abundant, is always going to be the primary source of information. This too creates challenges – smoke does not always equal fire, so what authorities need to do is set up a system that filters out the least reliable tips, but which really pours more energy into the decent and good ones.
This is why the future of anti-doping will likely rely less on testing and more on investigation. Right now, I sometimes feel we are in the middle of two models or ‘paradigms’. The problem this creates is that we are asked to trust the testing paradigm, but this has been shown to be worth only a very small amount of trust.
But the thing we MIGHT one day trust (investigation by credible agencies, including criminal ones – just think how many major doping revelations have been triggered by police, including the recent Winter sport raids in Austria) doesn’t exist yet. We are in a sort of doping confidence purgatory.
So anyway, I’m out of thoughts, and I feel like any more will simply be repetition. I started out wondering how I should evaluate the winners of the London Marathon. I’m no closer to knowing, other than it will be with a significant dose of skepticism, but also ‘professional interest’ because whoever it is still comes from a group that I think is physiologically, biomechanically, genetically different and fascinating for the way that biology has met socio-cultural, environmental and economic incentives and strategies. There’s a certain pleasure in that, despite the necessary eyebrow raised at some (and it’s higher for some than others, as I hope this post has made you realize!).