Wednesday, August 13, 2008

Irregular Report 10

The CyclingNews notes Pat McQuaid's disgust with Spanish cycling as he blasts the federation after the Olympic EPO positive of a Spanish female cyclist. McQuaid then goes into the "way back" machine to dredge up more OP criticism, and calls on WADA's John Fahey to go after the Spanish cycling federation.

The VeloNews has more comment from the UCI about the threat that cycling may be eliminated from the Olympics due to its record of doping violations. Maybe the IOC would also consider ridding the Olympics of track and field for the same reasons? Of course that will never happen no matter what, too much money, too much power in track.

Rant analyzes the Mayo decision noting the inconsistencies in the lab work, and the unfortunate testing pattern that may emerge from the case for the world's athletes:

Perhaps. But the precedent that’s been set is that a lab need only declare results inconclusive in order to keep testing, in the hopes of confirming their initial result. Sometimes that may be true. But it’s also a loophole that’s rife with the possibility of abuse. And, as has been noted recently, the more you test, the more likely it is that you’ll eventually get the result you want. If we’re going to allow multiple testing of the B sample by the anti-doping labs, shouldn’t the athletes also have a right to have the samples tested again and again. What if, in that circumstance, a contradictory result came up. Should the athlete then be acquitted of his/her doping violation?

Steroid Nation compiles a "list of infamy", those athletes banned from the Olympics this year for doping.

Medic Guide
thinks Donald Berry is off base.


Thomas A. Fine said...

If you can't eliminate doping, eliminate all sports with doping.

"Join us now for the 2010 Winter Games. All curling, all the time!"


Larry said...

Tom, I can't resist. From the 2005 archives of the NY Times:

"Seven years after making its debut as a medal sport at the Nagano Olympics, curling has what has become a rite of passage in athletics: its first doping violation."

A Doping Violation in Curling?

("Eightzero") said...

And if you think the speedo swimsuit is "technical doping" what about birth certificate doping?:

Larry said...

On the topic of getting together for a drink ...

... first, it ought to be a corn-based form of alcohol, in honor of Tom Fine. We don't want to deplete our C13 levels.

And yeah! 8-0, I'd be happy to have a drink with you, and d-bob, and anyone else who thinks they can stand the experience. Warning: I'm more wordy in person.

Confession: if we DO ever have that drink, I'm hoping that racejunkie joins us. I want to hear if SHE talks like she writes. Y'know. Like Kerouac on a ten speed after three double lattes, only sexier.

Gary O'Brien said...

Looking at that list I can clearly see why cycling should be eliminated from the Olympics and weightlifting events should be doubled.

Damn dirty cyclists.

strbuk said...

Is racejunkie the only girl invited, or can I have a drink too? :-) Make mine vodka please.


Larry said...

LOL at the thought of "only girl invited". I personally am more of an "only guy invited" kind of person.

strbuk, of COURSE you're invited! But don't count on me remembering what you're drinking, and I'm not responsible for the effect on your metabolites (vodka = potato = not much C13, if I remember).

Come to think of it, you're not only invited, you're our "A" list celebrity. This is something like DeNiro coming to my place for pizza after the Academy Awards.

Thomas A. Fine said...

Oh come ON!

In honor of Floyd, only Amstel Light and Jack Daniels are allowed. Isn't that obvious?


("Eightzero") said...

Larry said: "Warning: I'm more wordy in person."

HAHAHAHA! Well, I sort of had plans to turn this whole thing into a CLE seminar anyway. :-)

And honestly, I'm sort of stunned that an actual *girl* would want me to buy her a drink at a bar. Wow. But yeah, what Larry said - str gets her choice of drinks at the bar, and full service meal as well. My tab.

I'm totally serious. Any and all TBVers in or visiting Seattle need to contact me off-list. I'm buying. That includes you too, Floyd.

whareagle said...

Nuts, 8-0, I was just there! 400 miles in Chelan and about 60 on Vashon. MMMM-mmm! You got good hills, and good beer. And COOL double-rainbows over L. Washington.

Ya'll crack open a Mac & Jacks at Virginia Inn for me. I won't be back 'till January at the earliest.

Oh, and "The Art of Racing in the Rain" is an absolute CLASSIC.

racejunkie said...

Oh my. I read the comments ready to take out any monstrous tool who dares insult Iban Mayo and I end up quite blushing.

Of course I'm game for drinks. Full disclosure: Mr. RJ may want to come along--and at least *he'll* be able to keep up with you guys on a post-bender climbfest!

Rosemary said...

I think a video feed is needed so we can all be there!

marc said...

Very small world indeed, whareagle. I was just in Chelan on my way back to Paris.

strbuk said...

Rats, I am stuck on the east cost, well the eastern Great Lakes to be exact. I hate to miss not only the bar, but a meal as well!!! Lucky for me RJ was kind enough to send beautiful flowers, *that* girl has class!


racejunkie said...

Perhaps we'll have to get the east coast crew together strbk--or get the west coast contingent out this way! Has Floyd sent you that giant case of champagne he owes you yet?

strbuk said...

No RJ sadly Floyd has NOT sent me that case of champers, yet. But I am patiently waiting, HA HA HA!! And yeah an "east coast" get together would be great!!


woody said...

Landis got mentioned on Slashdot.

Mike Solberg said...

While you're all drinking the wells dry you can consider this:

See again the chart of LNDD 0311 (in Ex. 26) - LNDD blank urine CIR values from June, July and August 2006, in which the 5aA has an average value of -1.7! Somehow, in June, July and August of 2006, the negative control urine had an average 5aA of -1.7.

How then I wonder can -3.0 be a legitmate cut-off for a positive test?

I know that the "clean" number for 5aA probably isn't zero, but I would really like to know what "clean" number was used in the WADA study in which they came up with the 3.0 cut-off for all individual metabolites, including 5aA.


Davis Straub said...

Medic Guide seems to be completely clueless.

m said...

I’ve been away, so only now had a chance to read about the Berry opinion piece in Nature. I think it is highly biased and somewhat dishonest in its argument.

1. First, Berry is not a biological scientist but a statistician.

2. Despite this he argues that the doping evidence in the Landis case is a.) “non-probative” of doping, and b.) further agrees with a “fruit of the poisonous tree” type legal argument that since the initial screening test was invalid the confirmation IRMS test never should have been done and impliedly is non-probative also. He makes both arguments pointing to no specific evidence. He makes the legal argument with no supporting reasoning or any facts at all. These conclusory statements show a clear bias since he is just parroting the Landis legal team’s arguments..

3. His basic statistical argument is that for all doping testing we haven’t been told the false positive rates and the false negative rates. From this one might argue that any and all doping tests and results are suspect. There is nothing special about this argument that makes Landis’s results any more suspect than any other doping test. Yet Berry wants us to question Landis’s results in particular. His prosecutor’s fallacy argument has no specific applicability to the Landis case. One might note that his statistical argument here and in 5. below would lead one to supposed that the recent Duenas, Ricco etc. etc. etc. EPO positives all must have had a high probability of a false positive, yet each of these many subjects have admitted they doped.

4. His remedy would basically eliminate all doping tests. He wants blind validation tests on thousands of subjects so that we can determine a true positive and negative rate of from 99.99% to 99.99999% accuracy that even he admits is not possible to achieve. He published an article in Chance magazine in which he outlines this remedy. One might ask how one is going to administer doping substances to subjects blindly, much less use thousands of subjects, or pay the million dollars in costs of such testing.. He claims that such validation accuracy is routinely required in medical testing. However, OMJ who apparently works in the cancer testing field, disputes that, pointing out that medical tests often have low accuracy when the costs of improving the tests outweigh the medical benefits.

5. Berry points out the increased likelihood of a false positive as the number of tests increases. The actual statistical example he uses to highlight this possibility is highly misleading since he does not take into account Landis’s multiple IRMS doping positives on both the B sample and the samples from the other stages. If one takes into account the other positives the chance of all those being false would be only about 5 in a million.

He uses the example of an assumed test with an assumed 95% true positive rate and a corresponding 5% false positive rate. (Granted we don’t know the true positive rate – it could be far higher or lower.) Assuming that Landis had not doped, he correctly observes that performing 8 tests on Landis one might expect a 34% chance of one false positive. He ignores the Landis B test positive or the B test positives from the other stages. Out of those 8 tests Landis tested positive on 4 or 5 (not sure which). Assuming Landis had not doped and a test with 95% accuracy in determining a true negative, the chances of 4 false positives would be 5 in a million. .05*.05*.05*.05*.95*.95*.95*.95 = .000005

The statistical probability that Landis's positive doping results were all false positives is thus very very very low. The fact that he doesn't discuss the B sample tests, suggests ignorance or conscious disingenuousness.

Thomas A. Fine said...

East coast is a good idea. I'm tired of all these darned Californians (and yes, Seattle is part of California, at least according to East Coast geography).

For the record, I'm quieter in person.

Swim, WADA pulled the 3.0 limit out of their collective buts. It's not that there aren't studies, but the only study that actually suggested a limit suggested a higher one. It's as if they looked at these studies and then basically said "3.0 sounds about right".

And yes all the testosterone metabolites consistently average lower than the reference metabolites, to varying degrees.

Davis, I may agree with your assessment. I might have gotten to the root of their misunderstanding, check my latest comment there. Then again, they might just be clueless.


Thomas A. Fine said...

1. He is in fact a biostatistician. He's a leading expert in the statistics of medical testing.

2. The "fruit" argument is a statistical one. The combination of T/E screening for CIR offers statistical protection against the (unknown) false positive rate of the CIR. If WADA is willing to ignore the T/E, this statistical protectioin is gone, exposing the athlete to the raw false positive rate of the CIR (whatever that may be).

3. He does NOT want us to question Floyd's results in particular. He uses Floyd as an example to indict a system, and explicitly says he doesn't know if Floyd is guilty or not.

4. So your saying that two-year penalties and lifetime bans aren't that important so why bother? WADA doesn't even have a rough estimate on the false positive rate.

5. The other "positives" are not real positives. The had no A and B confirmation so in addition to the normal false positive rate, there's also the lab error rate, and possible influences form bias. AND the T/E test was bypassed, see item 2.


Davis Straub said...

Yes Thomas, I did see your comment. It was another comment on Medic's confusion about averaging results that clued me into the fact that they didn't really have any idea of what was going on with drug testing.

m said...


1. "2. The "fruit" argument is a statistical one. The combination of T/E screening for CIR offers statistical protection against the (unknown) false positive rate of the CIR."

Perhaps you could point out where Berry makes this "statistical" argument. He doesn't. All he offers is this conclusory statement:

"In arbitration hearings, the AAA threw out the result of the LNDD's initial screening test because of improper procedures. In my opinion, this should have invalidated the more involved follow-up testing regardless of whether or not sensitivity and specificity had been determined." "IN MY OPINION", yet he cites no argument or evidence to justify his opinion.

2. If the confirmatory T/E ratio test had been performed properly and was positive it would have added weight to the CIR/IRMS positive
B sample results. But remember, both Landis A sample and B sample tested positive using CIR/IRMS.

3. You claim Berry is merely making a general point and not commenting on the Landis case in particular. I agree that this is the best interpretation that one can give to his comments. So why does he make disingenuous and unsupported statements like this: "In my opinion, close scrutiny of quantitative evidence used in Landis's case show it to be non-informative. " "CLOSE SCRUTINY". Yet he points us to no specific evidence.

4. And what's with the argument that the other B sample results were not positive. Of course they were positives for the purpose of Berry's argument. We're talking statstics here, as supposedly was Berry, not legal technicalities. said...
This comment has been removed by the author.