This week blood has been a big issue. In particular the blood of Paula Radcliffe, Britain’s athletics superstar and marathon world record holder. I imagine most of you have seen some of the coverage.
This has been a fairly fast moving story today, I have had to scrap the first draft and start again! The latest is that Radcliffe has released three values she claims were those flagged as suspicious. Her reported ‘OFF-scores’ were 114.86, 109.86 and 109.3. ‘Off-scores’ are calculated with a formula from the concentration of haemoglobin and proportion of new red cells (reticulocytes) [details of that calculation here for those interested]. Her explanations for these fluctuations include training at altitude and less than two hours from exercise to sample collection. The altitude training may have an effect on increasing haemoglobin, which could increase the off-score and according to the World Anti Doping Agency collection guidelines should be recorded on the sample collection form. The two hour stand down after exercise is now mandatory, but was not prior, in the above WADA guidelines. But can she ever clear her name, if innocent?
These readings don’t fluctuate much do they?
Not a great deal, but there are standard numbers used to flag suspicious samples, following altitude training this is reported to be 111.7 for females, which means only one of Radcliffe’ samples was above that threshold. So – only one case to answer not three? Well not really, as I described in my intro to the passport one test isn’t taken on it’s own. Each tests is judged against all prior tests and all future ones also. The case of Jonathan Tiernan-Locke (cyclist) showed that a sample can be flagged as suspicious a year later after more data is collected. So, to truly test these numbers you would need a longitudinal history of her ‘off-scores’ to judge how these three fit with the others.
Does that prove anything?
No, the data she released shows one suspicious sample, which she is convinced she has an innocent explanation for. If the rest of the values in her profile are all 110 or so she has no real problems from what I can see. But what do the individual components of her blood show in those samples? Low reticulocytes can mean the body has turned off it’s natural production of red cells due to blood doping or EPO use. Or they could well be innocently explained. But can she convince everyone it is the latter?
Does the passport work and is it fit for purpose?
Maybe. Maybe not. There are obviously issues with it. Apparently it is sensitive enough to measure changes through dehydration or altitude, but it has been shown not to be sensitive enough to detect ‘microdosing’ (well illustrated by the Panorama investigation into doping earlier this year and previous work by ‘so-called expert’ Michael Ashenden ((c) Seb Coe 2015)). Microdosing means taking small doses of a drug or blood product more frequently than the traditional larger once or twice weekly doses.
Is it then not fit for purpose? That will depend on what you think it’s purpose is – to eliminate doping or to limit the extremes? Does it eliminate doping? No, it can be worked around as proven, therefor it is not fit for that purpose. Does it eliminate extremes of doping? Possibly so, certainly microdosing blood or EPO is not as dramatic as full doses, but can still have considerable effect.
False positives
In the forensic world it is often said that it is better for a guilty person to go free (what we can call a false negative) than an innocent person to be found guilty (false positive). To this end the passport may have a problem. If the fluctuations may be such that an innocent person has values that can be read as suspicious you risk getting false positives. An innocent sportsperson could face a career ending ban unnecessarily, either because the panel of experts do not agree with their explanation of the fluctuations or because they do not have a money to fight an often lengthy and expensive legal battle. In forensic terms this is akin to those innocently executed in the US through hair evidence found to be unreliable years later.
So why has she changed her tune on transparency?
We know how Paula was very clear on transparency and antidoping in the past, now she is saying she shouldn’t have to release her data. A colleague put it to me today:
@WeAreForensic She was happy to release data when it was presumed all blood data, even post-comp, would be stable. We now know it isn't.
— TheOtherDrX (@TheOtherDrX) September 10, 2015
I can accept that – she now knows that the tests do not tell the whole picture and wants her data evaluated in private not through public scrutiny. When the passport was first talked of over ten years ago I think there was an expectation that it would be extremely sensitive and selective to doping that it would be the definitive test, as she herself campaigned for. In my opinion she hasn’t handled this case (or her reaction to it) brilliantly but again, that does not make her guilty.
To me, if it was shown that Paula Radcliffe had doped, it would rank as a bigger shock (by far) than that of Lance Armstrong because there has been no rumour, innuendo or no ex-employee coming out against her. By a few years into Lance’s career there were already all of these things and people refused to believe them, whereas with Paula nothing. I know it would shock most British athletics fans and would have a massive negative impact in the sport. Let’s hope that it is cleared up though, whichever way it falls. If innocent the last thing Paula deserves is an asterix indicating suspicion hanging over her marathon record from now on.
Very interesting. Initially I thought this was just another allegation but I agree with you that it is actually seismic whatever the truth may be. You raise important issues about both the biological passport and transparency. I have two broad questions.
Noting your comments on false positives, what are the pros and cons of test data for elite athletes in all sports being made routinely available for public scrutiny? Paula seems to have made matters worse as you suggest by releasing partial data.
The other question concerns the populations used to get the ‘standard numbers’ for suspicious samples. Who are they? My concern is maybe naïve but I would expect even non-doping world champions and record breakers to show extremes in tests. In the Guardian today Paula’s ‘physical therapist’ for 17 years was adamant that she was clean. The third of my two questions now occurs to me. Can extreme and dedicated training have a similar effect to doping and if so could we ever tell the difference?
Thank you, very good questions. The more I think about it it does show some considerable flaws in the methodology which I don’t think have been answered. In answers to you questions:
1. it has been proposed fairly often by some that all data from all tests be released. I think in cases such as this (see similarities to Chris Froome in the Tour de France this year), it is the selective release that makes things worse. The athlete is then accused of covering up therefore being a cheat. If you put it all out there the argument goes that it will let everyone see the complexity. However, in a legal case the evidence is first presented in detail in court, though often you know that for instance fingerprints were found in the press prior. There is merit it keeping things ‘hidden’. A forensic example would be that your fingerprints were found on a laptop known to be stolen from your friends house. If this is published in the press it may be thought that you had stolen it, rather than simply having had legitimate access to it to last time you met your friend for coffee! In that case (as in the altitude excuse) the circumstances are most important. A rather long way of saying I am not sure which is better, either all published or none!
2. Standard data is only one of the criteria for that very reason, elite athletes have an elite physiology which is not always identical to the general population. This is supposed to be the beauty of the passport though, you are compared with your own values longer term. We need those longer term values in each case to make a judgement.
C. Training does have an effect on the values. Tests have been done on cyclists riding three weeks races and the blood values have been seen to change significantly (not always predictably but there are patterns). Unpredictable results mean that results hold less evidential power..
Thanks for your questions!
Hi Tom. A few thoughts: I certainly believe Paula Radcliffe has never doped. She has been made to look slightly foolish by changing her stance on whether testing data should be made public or not (for others under suspicion in the past yes, but for her now, no).
Each of these 3 samples would have been looked at in the context of her previous samples. For a sample to flag as suspicious, 3 of the parameters measured have to be outside the calculated normal range for that athlete at a probability level of 99.9%. Then the data may get referred to the expert panel, who would view the profile and take into account all the extra information – medications/altitude training/medical intervention (legal)/transport temperature/lab analysis etc – could any of that explain the one high value? Having ruled all these out, then the onus is on the athlete to come up with an explanation – if they can’t then they may face a sanction. I assume if Paula’s data was scrutinised the panel would have taken her time at altitude into account and decided this did provide an adequate explanation.
As for fitness for purpose, the passport has been shown to have reduced extreme values in the peloton (see 2010 paper by UCI’s Mario Zorzoli, himself later investigated). Maybe microdosing could be detected if the probability thresholds were lowered say to 99% – meaning more samples would flag as suspicious – but at greater risk of the false positives you mentioned.
In answer to Si’s 3rd question, the effects of prolonged training at altitude are identical to the effects of taking EPO, i.e. increase in HGB = greater O2 carrying capacity = great benefit to endurance athletes. One is legal and one is cheating!
Thanks very much for those comments, very helpful. Let’s not forget the reason we are here in the first place is that it was alleged the IAAF had not followed correct protocol and had not investigated any of these ‘suspicious’ samples. Yes it should be that her samples were investigated and cleared but have they been? This is unclear currently. Her release of one parameter for each test does not help, as you say it is three outside values needed. Therefore can we guess if these three tests have been flagged then at least two more parameters for each test were also not normal? Should she release them all? How do you feel about that?
The false positive negative thing to me is very troubling: relax the criteria to exclude false positives and you risk missing loads of real positives or tighten up and you are getting lots of innocent people drawn into suspicion.
Is the ABP (haematological module) really no more than a screening tool and should it only be used as such?
Thanks again!
No surprise really. It’s all in one old good saying: “…The thief shouts loudest…”.