Friday, December 26, 2008

The Winnowing: One Mint Julich

One Mint Julich is a frequent contributor on the Daily Peloton Forums, and heavily participated in the productive scientific discussion there. At our request, he summarizes it all from his point of view as follows.

[Back to the Introduction]

I’m a scientist who is fairly familiar with, though not expert in, the methodology used to test Floyd. I was fairly impressed with the evidence against him in the IRMS test, including:

1)the size of the delta values for one of the metabolites;

2)that this metabolite also exhibited a large delta value in Floyd’s samples after several other stages of the TDF;

3)that a second metabolite also had a sizeable delta value (after both stage 17 and as I recall, several other stages), above or very near the 3.0 criterion (depending on how one treated the standard error for the procedure)

4)that the delta values for the remaining two metabolites, though not reaching the 3.0 criterion, were still substantially negative;

I did not feel that the various technical errors pointed out during the trial were that damning against WADA’s case. For example, the evidence was very compelling that the metabolites were identified correctly, and while contamination was possible (which could skew the delta value of a correctly identified metabolite), it seemed to me unlikely to exist to the extent needed to account for the high delta values, and there was no evidence raised during the trial to indicate there might have been such contamination. Certainly the amount of sloppiness should raise some concern, since if the lab consistently made this many errors, sooner or later they would likely significantly affect the conclusions of some doping test. But I did not think they did in this particular case.

I also believe that the conclusion was strengthened by the initial positive T/E test. Granted, there were problems with that test, too, and in the end it was actually thrown out by the panel. But calling results inadmissible does not mean the conclusion they point to is false, any more than failing to read a defendant his/her rights means the defendant didn’t commit the crime. Floyd did have a fairly high T/E value, I think that’s very clear, and given that the T/E test is an independent indicator of synthetic testosterone, I think it provides greater reason to believe in the IRMS results.

Having said all this, I never felt the case against Floyd was slamdunk, and had this been a capital crime, where guilt must pass the beyond reasonable doubt test, I would have been somewhat conflicted. I felt the biggest problem was the lack of consistency among the delta values for different metabolites. The literature generally indicates that the delta values of these metabolites follow each other closely, and given the origin of these values, that is what one would expect. WADA did produce some evidence that this would not be the case following transdermal application of testosterone, but this was preliminary data, based on only a few cases, has not yet to my knowledge been published, and still lacks, I believe, a strong conceptual basis.

The bottom line for me, though, was that one very high delta value. Barring contamination, I don’t see how it could indicate anything but exogenous testosterone.

2 comments:

Ali said...

OMJ, It's nice to hear an honest, unbiased opinion on this matter.

However, I'll say right now that I feel you fail to realise the true significance of the practicalities of performing the IRMS test. It's not just my opinion, Dr Simon Davis has expressed a similar view that the "result" you get is very sensitive to how you process the data. It is easily possble to shift the results from a negative to a positive with but a nudge of the mouse. That's a fact which nobody wants to accept, but remains the truth.

I really don't think that fact was ever accepted by most observers ... maybe it just seems too unlikely that "science" could be so subjective. Well, let history judge this situation, not me. I know the facts. They're pretty easy to follow. It's a no brainer (as you Americans say). It all boils down to mathematics, and the numbers don't lie !

Alasdair

Larry said...

OMJ, it is good to see you appear here, even if it is at the last minute! You have many fans on this site, me among them.

I had always hoped to get your reaction to my last magnum opus here, at Bourbon and Beer.

I have read you to say that the LNDD testing and the WADA delta-delta standard are fair topics for criticism, but that it is hard to get around a delta-delta reading of -6. That's a good argument, but it's based on the assumption that a delta-delta of -6 falls well outside of three standard deviations beyond the mean. The point of my last magnum opus is that the scientific studies fail to agree either on the mean or on the measurement of a delta-delta standard deviation.

I'm no scientist, but I can read a scientific study, and I can compare one study to the other. At least I think I can.

In any event ... I have learned a lot from you. Any reaction to my magnum opus would be appreciated. Thanks for all you've written, and best of luck to you in your future endeavors.