Last week, a number of articles came out discussing the important case of Irish sprinter, Steven Colvert. He was banned for two years after failing a test for EPO.
The background to his case is best explained by others in the following articles:
- The original article in Lab Times, in which Norwegian scientists evaluate the scientific procedures that saw Colvert banned (Jump to page 16 for the Colvert article)
- Prof Roger Pielke adding his insights and interpretations in Newsweek
- Journalist Cathal Dennehy, who wrote two pieces on it:
So, you may think this is a minor case, and hopefully such instances are in the minority, but it’s also a really important one, and so I thought to share a few thoughts in addition to what was quoted in those articles.
False positives and false negatives, and the need for solid science
Conceptually, the point I’d make is that it’s really important for anti-doping to rest on a solid foundation of science, because it’s entire ‘being’ is to protect the clean by catching the cheats. It becomes very difficult to do that with any integrity if the clean are being caught too! Now, in any system, there’s a pay-off in which we accept that there is a CHANCE of an innocent person being wrongly punished, in order to catch all (or almost all) the guilty.
However, with anti-doping, we have two problems. One, we don’t always know what the false positive (and false negative, for that matter) rate is. In other words, we aren’t always aware of the validity of the tests being used to catch dopers, so it is very difficult to evaluate whether we have achieved a satisfactory compromise in this regard. Roger Pielke has written a really good article on this, in which he explains false negative and false positive tests, and talks about some research WADA cited in a previous discussion on this issue.
The second issue in anti-doping is that if a clean athlete protests their innocence and challenges a test as a false positive, then a rigorous review of their case must be available and enacted. This seems to be a problem, highlighted by the Colvert case, because the “follow-up” science seems to fall well below a standard of rigor that it should have.
Here too, Roger wrote a piece talking about one such failing – the failure to keep samples from athletes for more than three months. An athlete can challenge a test result and compel the authorities to store the sample indefinitely, but it’s absurd that a sample, once tested positive, is destroyed so soon in the absence of a challenge. For that matter, even if a sample is negative, it should be kept for ten years for retrospective testing.
The specific case of Steven Colvert
In the Colvert case, the first assessment of whether he’s doping or not was subjective, even for an expert. That’s why you get the WADA expert saying “Positive”, the Norwegian experts saying “Not clearly positive” (note that this is not necessarily the same as “negative”). I found it very difficult to tell which of those samples would be deemed positive for EPO, but I’m certainly not an expert at reading gels, so I wouldn’t put too much stock in my opinion. That there is a conflict between the Norwegian and WADA scientists is the key point.
But then there is a second layer, a second test that was done on Colvert’s sample (called the IEF) which should at the very least support the first. It didn’t. That alone should be grounds to go back, and start again. But they pushed through a guilty verdict in the absence of scientific agreements between their own tests on the same sample.
So whatever “failsafe” existed in round one was ignored or not activated. The subjectivity of the initial decision is less concerning to me than the fact that they brushed aside the inherent contradiction between the two. I think what it shows is that they made the initial call – EPO positive – and then sought only to confirm it.
Now we move on to the B Sample. The same subjective process seems to have been applied to Test 1 of the B sample, but they did not run Test 2 on the B-sample, so we have no way of knowing whether it would have repeated the strange paradox of the A sample. That too is a failure – you can accept a chance of false positives due to subjectivity, even lab error, as long as you have the ability to rerun the tests in their entirety.
So there were multiple procedural failures, starting with the subjectivity of the science (which can be acceptable as a first step, provided it is repeatable and supported by Test 2, which it wasn’t). Then come the procedural issues that appear to have been denied to the athlete, for reasons unknown.
Science and falsifiability
It looks, from my perspective, that the WADA lab have defended their initial position strongly, which is actually “anti-science”. And I think this is a key point – a fundamental requirement of good anti-doping science must be repeatability. Or, from the other direction, it must be open to falsifiability. So, having discovered the initial discrepancy between T1 and T2 on the A sample, good science would invite a re-examination of the sample, looking to disprove what it found the first time around!
If WADA trusted their science, then they would welcome this, because good science is repeatable, and so it would confirm their initial finding and their case against Colvert would be twice as strong. On the other hand, if they don’t trust their science enough to welcome a challenge, a retest and a repeat of procedures, I would ask how we can be confident enough to issue a two-year ban and tarnish an athlete’s reputation based on that science?
As for next steps – if the B sample was not tested fully, and the DNA test could not be performed as requested, then simply procedurally, I’d argue that the case should be dropped (OK, it’s late for that, his ban is completed already). But he should at least be cleared. Does it mean they’d say he didn’t dope? No, because there’s no basis on which to say he didn’t.
But there’s no legal, or scientific basis, to say that he did, either.