To quote Gordon Farquhar at the BBC:
“Armstrong is an accomplished liar. Serious
questions remain over the analysis of blood tests taken from Armstrong in 2009
and 2010 following his comeback. Usada says the tests "built a compelling
argument consistent with blood doping. Armstrong has not yet admitted to doping
on his return to the sport in 2009. Wada president John Fahey doesn't buy Armstrong's
assertion that he was clean during his comeback. Tests were carried out on 38
of his blood samples from that time by the Australian Institute of Sport, which
concluded that the chances of the blood values recorded occurring naturally
were one in a million. The tests were not part of the recognised anti-doping
programme, but Fahey wants Armstrong under oath again on this”.
I couldn’t agree
more. In his Oprah interview Armstrong makes the case that the sport is cleaner
and it is harder to cheat because of out-of-competition testing and the athlete
biological passport. Yet as I have said on this blog before, his blood samples
in 2009 and 2010 were highly suspicious. The data that makes them suspicious
(plasma volume, reticulocyte count) are the same that would raise warnings on
the passport. This is the basis for the “one in a million” comment from Prof.
Chris Gore mentioned in the above quote.
If Lance Armstrong
was not lying to Oprah and he was in fact clean in 2009 and 2010, it calls into
question the whole biological passport system. The only resolution is for
Armstrong to work with UCI, WADA and Chris Gore to explore how such anomalous blood
results could have occurred by chance. As some have suggested on my blog (step
forward “anonymous”) perhaps he used hypoxic tents in Italy? If we don’t know how
this happened there could be a whole lot of “false positive” passport results
and athletes could be accused unnecessarily.
While being absolutely no fan of Lance Armstrong's, indeed I've been convinced of his guilt since well before the USADA case broke this summer, is this not a case of the Prosecutor's Fallacy?
ReplyDeleteWe need to know the probability that he is guilty given his blood values. Not the probability that his blood values occurred given his innocence.
The two are related by Bayes Theorem, to get the probability of innocence given the blood values we take the probability of blood values given innocence (1 in a million) and multiply it by the prior probability of innocence and divide it by the probability that his blood values occurred (irrespective of innocence/guilt).
Now the prior probability of innocence is low (given his admitted crimes) and I'm a mathematician not a doctor, so I can't really comment on the value of the probability of the blood values, but given the failure of the blood passport to catch *anyone* I'm guessing it's low too, perhaps lower than the Probability of Innocence in which case the 1 in a million figure could be much bigger. This is the failure that led to the conviction of Sally Clark for the murder of her 2 children who died of cot death.
But then again, as I said I know nothing about biology and am probably talking rubbish and it's all irrelevant since we know he definitely doped in 1999-2005.
Thanks for this very interesting comment. I will need to think about it in more detail, but my initial thoughts are these.
ReplyDelete1. The blood passport is in fact based on Bayesian statistics
Upon being enrolled in the passport program, the athlete gives blood regularly. At first it is assumed that his blood parameters are the same as the general population. As more tests are taken a statistic is calculated that is biased towards the individual parameters of the specific athlete. The athlete becomes his own control. If his blood deviates from his own normal readings, such as might happen following EPO doping or a blood transfusion, he can fail a doping test even if no specific doping product is found in the blood.
See my blog here and apologies if I have oversimplified the maths:
http://runswimthrowcheat.blogspot.co.uk/2012/05/presbyterian-minister-and-blood-doping.html)
The passport was not in place at the time so I am not sure exactly what calculations Gore used. I guess it would have come out if it went to court.
2. It has in fact caught people. But rather few. Largely for the reasons you outline the statistic that triggers a positive result is set at a high level. The statistic is also used to target people considered "likely" dopers for additional out-of-competition direct testing.
3. Here is the more detailed rationale for Gore's decision (sorry I don't have access to his raw data)
34 blood samples were taken from Armstrong between October 2008 and April 2012. A cluster of five Armstrong samples taken during the 2009 Tour and two samples during the 2010 event contained an unusually low percentage of reticulocytes, or immature red blood cells that are created naturally by the body. The suppressed reticulocyte percentage in Armstrong's 2009 and 2010 Tour de France samples were then compared to the reticulocyte percentage in his other samples. Prof Gore concluded that the approximate likelihood of Armstrong's seven suppressed reticulocyte values during the 2009 and 2010 Tours de France occurring naturally was less than one in a million".
The Bayesian statistics (as I understand it) is what gets you to the one in a million i.e. the Sally Clark fallacy does not hold in the normal blood passport. But I am not 100% sure this is how Gore analysed his data.
See also : http://www.smh.com.au/sport/cycling/ais-professor-triggered-red-alert-on-armstrong-blood-cells-20121011-27fr2.html#ixzz2INXn0LAA
Thank you. that's really interesting, I wasn't aware that the blood passport has actually caught people. My uneducated view was that athletes would simply moderate their doping to avoid suspicious readings in a similar way to how in Armstrong's era they would moderate their doping to keep their haematocrit levels below 50%. But it appears that it's not quite as simple as that.
ReplyDeleteOn the statistics issue perhaps it's a case of imprecise language rather than a problem with the method. I need to get more into the details of how the stats work. The following link from the Swiss anti doping laboratory looks really interesting, http://www.doping.chuv.ch/en/lad_home/lad-prestations-laboratoire/lad-prestations-laboratoire-passeport.htm#abp-10
By the way, I loved your book. I thought you did a really good job of making a really complicated subject accessible to a non-expert reader, a rare skill.
Thanks for your comments about the book. You are quite right about athletes modifying their doping habits to fool the passport. I blogged about this on April 3 "Resignation of a key anti-doping scientist" . And Armstrong in his interview seemed to imply he was using low doses of EPO. But I think there is no way to hide the change in red cell synthesis when you give a blood transfusion and you have samples before and after as seems to be the case for that Gore was analysing in Armstrong's data.
DeleteI have a read an article about doping in tennis and would like to know your opinion.
ReplyDeleteArmstrong in his interview with Oprah said its hard to dope nowadays because of the athlete biological passport but in the other hand Djokovic highlighted that there is lack of blood testing in tennis as he himself has not been tested in 6-7months. Shouldn't all sports introduce athlete biological passport not only cycling?
I've read your blog and I find it really interesting how you negotiate between the topics of level playing field. I am also interested in the field of drug testing so what should be my starting point if I want to pursue a career in this field.
http://www.independent.co.uk/sport/tennis/itf-to-bring-in-biological-passports-to-beat-doping-8462284.html
http://www.dailyrecord.co.uk/sport/tennis/andy-murray-says-there-is-no-doping-1542649
Dear Pranti,
ReplyDeleteThanks for these comments. I agree that sports like tennis should bring in a more aggressive anti doping program. The endurance and power required, particularly in the best of five sets of the modern men's game, clearly places exceptional demands on athletes. Some could be tempted to dope, not necessarily because they would perform better than a clean athlete (the limits to performance are more about skill, psychology and practice than in cycling, where the physiology is more important), but rather to cut corners in training. It is good that it is Djokovich and Murray are amongst the key proponents of more testing. They are frequently described as performing superhuman" feats of endurance and recovery. No doubt we will hear this phrase on Sunday at the Australian Open final (go Andy!). Sensibly they want this to be seen as a compliment, not an accusation.....
Hello Professor Cooper
ReplyDelete"I never... I didn't fail a test," he says. "Some tests were retroactively tested, and yeah I failed those, but there was nothing in the system."
Does the above statement from Armstrong (at 9:57) during the Oprah interview imply that Armstrong is claiming to have no EPO in his system during competition? Perhaps relying only on blood transfusions ?
Who knows what was in his mind? I assume the tests he was referring to our the delayed EPO tests the Paris lab did on the 1999 Tour samples. "Nothing in the system" could be just a further denial that there was a cover up of a positive test.
ReplyDelete