The Flawed Methodolgy
I think this got confused, but we can explore this over time.
I believe what cross of Dr. Brenna was getting to, but not quite closing upon, was that the peak identification methodology of the IRMS testing at LNDD is fatally flawed by design.
[more]
Cross of Brenna's rebuttal seemed very close to catching him in an outright lie, but Suh let him off the hook, perhaps from being short on time or distracted by the upcoming close. It was by far the most aggressive questioning of the hearing.
None of the accreditation inspections, internal review of the process, or Dr. Aquilerra and Dr. Botre's supervision of the retesting and reprocessing detected this methodological flaw.
Well, perhaps Aquilerra and Botre noticed, but they didn't say anything about it. Which is a more charitable interpretation?
It has to do with the correct use of relative retention times, but the simple explanation still escapes me. I believe this is the heart of Herr Doktor's testimony, and that Brenna's rebuttal tried to sidestep the issue by diversion on a non-relevant point, which was done correctly.
If the more technically adept of our readers can try to explain, we'd all be indebted.
[UPDATE:] Clearest answer so far:
You3 said...The testimony indicated that by ISL every peak needed to have a retention time that was sufficiently close (1% or 0.2min) to a corresponding peak in a reference sample. This check insures the right peak (elute) is chosen, the gc column is behaving correctly and there are no contaminants or interferences. But, what reference is to be used for IRMS peaks? If you use the GC/MS reference results, the elution times don't match. It turns out that there isn't even a good linear fit to the relationship according to my calculations. In cross, USADA addressed this issue by pointing out that the calibrations samples (MixCal?) could be used for reference times. But, Davis pointed out that some of the metabolites are not included in those samples, specifically Andro, 5a-diol and P-diol, in other words all the ones that matter for Landis. So, there is no way of establishing that these 3 peaks elute at a time sufficiently close to corresponding peaks in a reference sample.
I think that everyone believes that these peaks are mainly composed of the expected metabolite, but a key ISL check is missing with regard to proper GC operation, contaminants and interferences.
I have previously observed that it would be good to include both 5b-diol and 5a-diol in the calibration to validate that there are no unexpected interferences between the two peaks.
Herr Doktor Professor Wolfram Meier-Augenstein also pointed out they have a cal mixture with the right references in the lab, but choose not to use it.
89 comments:
That's what I understood as the testimony was unfolding but I am by no means a scientist. Hopefully someone else can crystalize it for us.
Simply put, each type of molecule has its own specific mass and structure, each of which give it its own unique "signature". Thus golf, basket and baseballs are all round but of different sizes and densities. So at a starting point of zero you can calibrate such a machine to know, from trial runs, that the golf balls come out with a peak at 60 seconds, baseballs at 90 sec and basketball at 120 sec. If you therefore take an unknown set of balls and toss them in and the peak comes out at 90 seconds you can assume they are all baseballs. Of course to know that, the flow rate must be held constant at all times such that the relative retention rates can be counted upon. Apparently holding such flow rates constant was too difficult for the french to do: too many cafe breaks no doubt.
But then to know they are truly baseballs you need to collect that peak at 90 seconds and toss that in the mass spec to analyze the components of the baseballs and determine that, yes, they have the characteristic chemical compounds of baseballs that allows them to be differentiated from golf balls and basketballs. But again, the MS has to be calibrated (preferably without a frigging hunk of iron stuck to the top of the magnet) such that the chemical signatures can be truly defined and believed, a probelm that the french also had a problem with. Garbage in, garbage out...
No wonder the French Open has chosen to do their drug testing elsewhere. And it is a crying shame that the Canadians and UCLA and SLC and the other WADA doping labs don't stand up together and tell the world they think the french lab sucks. This is a cone of silence is unforgivable in this situation.
Based on my reading TBV and watching the hearings, I believe Floyd is the rightful winner of the Tour de France.
TBV has done a remarkable job in helping those who witnessed the public hearing and those who could watch the courtroom session to better understand Floyd's situation and the potential problems future cyclists will have if USADA and WADA do not clean house, before leaking to the world some one else has stepped over the line with the public's confidence.
Its ashame about the LeMond Affair, because it takes away from the science.
With that statement, I will be contributing once again to the FFF, and hopefully others who have observed the hearings will do like wise.
~Paul
Bryn Mawr, PA
Continuing the baseball analogy, when the data on the balls is put in the mass spec, is this when the retention times have to match those from the first machine? This was my understanding of Herr Doktors point. If my understanding is correct, then Brenna was either blowing smoke in his rebuttal or doesn't have a clue (something I find hard to believe). I think he was comparing two separate runs through two different machines, but measuring the same thing.
Davis appears to agree Dr. M-A based on this exchange with Young in his cross examination...
"q: so when you run the EDF's with auto BG, that takes out consideration of manual selection.
a: yes.
q: you asked that be done.
a: correct.
q: and they still came out 5a positive?
a: they aren't even the right peaks, the numbers are meaningless."
At risk of being made to look stupid (i.e. Tyler Hamilton), I share Pual's conclusion. In addition to the obvious lab issues, I also look at it this way: The TDF "Puerto" guys were out. We were all excited seeing a "more human" race. Landis started pulling ahead, but not with the arrogance of the "Puerto (and Armstong-no editorial comment intended)" types-i.e. the break that was allowed to go. Landis raced humanly; shown additionally by his collapse. The "Puerto types" never showed those collapses. I may be dreaming, but I just may echo Floyd's coaches comment-we should be talking about what an amazing ride FL's comeback was. And also point out that the other teams screwed up in not chasing him, or he wouldn't have got enough time to win anyway. Okay, time to watch Giro instead of TBV, which was actually more exciting until today.
AnonDT
I think what Herr Doktor was saying was a bit more complex. Using the baseball analogy above, what we really see is three peaks separated by 30 seconds. Yes, they come out at 60, 90, and 120, but the difference is what is critical. This creates a sort of "signature".
What I heard in the testimony from Dr. Davis was that the techs at LNDD were using the software to move the integration lines to match up with peaks of the signature based not on when the compounds should have come out(retention time) but, when all the compounds actually came out(relative retention times) as per the signature.
Without knowing the actual retention times for the compounds, you are just guessing(or using your experience) to match the compounds to the retention times.
The one thing I will always remember when ever I see TDF pictures of Floyd dumping water on his head is how effective he used his team car to keep his body temperature and pulse rate down by his use of the water that was readly available to him, which the riders were not able to receive in that quanity.
Dr. Lim has discussed how going through 60 bottles helped Floyd to win.
This point is very rarely mentioned, but it was critical to his success.
~ Paul
Bryn Mawr, PA
As a related aside, will this year's TdF riders allow themselves to be tested by LNDD? Maybe they'll get their act together and refuse to ride unless their testing is done elsewhere. Why take a chance on your career being ruined by a "lab that can't shoot straight."
As others, I am not an expert.
However, I followed everything presented here and was able to get on the video stream several times.
I was primarly surprised that I was able to follow the scientific discussion, but what surprised me most is that the defense closing did not hammer home the fact that the USADA experts were claiming positive results on faulty data.
Perhaps the panel got the point, but I think they could have stressed a little harder and made it more clear that when the entire dataset is bad then no conclusion can be drawn - either positive or negative.
I also don't understand why the TUE Cortisone was _NEVER_ mentioned. If this was NOT specifially detected, isn't it possible that its presence was skewing the results - especially considering the improper practices that were in place.
What comes to mind is the constant reprocessing. If they were getting peaks for the Cortisone and then MOVING POINTS AND ALTERING BACKGROUND until those points more resembled what they expected to see in a screen positive for T, that might explain the gaps in time, the non-linearity of the results, and the overall haphazard nature of the testing.
All of this resulted from a faulty T/E screen - which USADA never touched. If they had clear proof that the T/E was really 11:1 then they would have hammered on that point all day long. But they didn't.
To me, that proves the rest of this witchhunt - which started because of bad results from a bad test - is nothing more than grasping at straws. Straws which, as the defense experts made clear, weren't even there in the first place.
Whatever Landis did or didn't do (and I am now inclined to say "didn't") Anonymous 8:56 is absolutely correct. The riders need to exert some control over this. They have been slapped into submission by WADA and are stuck with a proceedure that is frankly offensive to our system of justice. I know we (the US) are only a small section of the sport, but the errors identified in this lab and the proceedural hurdles any accused athelete has to face are simply unfair, and need to be changed.
Anon 8:47, Lets see if I have this right now. So the balls are going through the machine and will be coming out at a set interval of say, 30 sec. When they did their manual corrections, they actually read the peaks at 31, 59, 92, 115 sec? This is the time that should have been within 1% or .2 min. (of 30 secs)? Am I getting closer? I am not a scientist and I don't play one on TV!
9:06 sacky
good post. I've wondered about the cortisone myself. When the news about Landis's hip and the cortisone came out during the TdeF, my thoughts were no good is going to come out of this.
As I understand it (and I do have a PhD in Biochemistry although I'm not a mass spectrometrist), in the GCMS the compounds in the sample are separated (this is the chromatography, GC = Gas Chromatography) and then the MS (Mass Spectrometer) is used to identify what's in the peak. Essentially it tells you with an error of less than one atom how big it is and what it's charge is (the 1 ion versus 3 ions argument centers on this issue).
The IRMS is a second instrument that also uses the GC column to separate the compounds. However, in this case you don't get analysis of what the compounds are. Instead the instrument tells you how much carbon-12 and how much carbon-13 there is in each peak (the 44/45 trace mentioned). It also tells you about Deuterium and Hydrogen (the 2/1 trace).
What LNDD is doing is using the GCMS to see a pattern of peaks with the ones of interest being say the 3rd and 5th. With the GCMS they know exactly what these are. They then look at the pattern coming from the IRMS and look at the isotope ratios of the 3rd and 5th peak.
The problem pointed out by Her Doktor is that they are matching patterns, NOT times. The Landis camp says (and apparently the ISL) says that they should be looking at the peak from IRMS that appears at the same time as the peak from the GCMS. Landis says same time = same compound for the 2 tests. LNDD says same pattern = same compound.
I hope this explanation helps.
timm and others, yeah you have the basic concepts right.
but the question is maybe someone got a whiffle ball in the mix. you didn't see any baseballs at 90 sec but you did see a peak at 92 sec. so that the hell is THAT? could be a whiffle ball. so the 92 sec peak needs to go to the mass spec and analysis done to know if it has the same signature as a baseball, or a whiffle ball, or who knows what. my understanding of the french lab is that they did not try to figure out if they were indeed whiffle balls but instead just adjusted the lines to the extent that the 92 sec peak now looks like a 90 sec peak so they MUST be baseballs. a very unscientific (and damned lazy) approach.
as far as LIVE WRONG above goes, I don't disagree with his (easy to know it ain't a girl: too much soy based T for that) basic sentiment about everyone doping thus Floyd must be guilty. If USADA really wanted to know that....I'll repeat that in case he reads slow.....IF USADA REALLY WANTED TO KNOW if Floyd had doped or not, then his B samples and other day samples should have been analyzed at UCLA which is orders of magnitude mo betta than the french lab. The fact they did NOT press for this testing tells me they did NOT want to know, and thus i can't even begin to assume FL is guilty.
It has been a number of years since I worked in a lab, and I did TLC and HPLC not GC, but here's my take.
I think that they were saying was that they had numerous reference peaks that correlated with the peaks in the Landis sample in the first machine.
Then when they run the second machine, IRMS?, the so called positive control (mix cal acetate) has only a single component (1 peak). The retention time for this one peak was not the same as any of the peaks in the Landis sample, but they assumed that the peak matched the closest peak in the landis sample. They then integrated (quantified that peak) to get a number. That number may or may not be accurate for the sunstance in question, depending on weather they guessed right or not on which peak it was.
I could be wrong, because I haven't been able to match the chromatagrams together with the testimony. This is just what I think the discussion may have been about. That is why Suh asked where the satndards were for the other substances.
I have 55 degrees* here in the Great Pacific Northwest, and even if I had 0 degrees I would repeatedly ascertain LNDD is a mess. And will join in the chorus who say the riders should say non, non, non monsieur not by that lab.
That lab has done more harm to the Tour than all the riders ever could.
* fahrenheit
Anon 9:20
that helped a lot! thank you.
bamalaw said:
(Anon9:20 beat me to this and has a better explanation, but I spent a lot of time on this so here it is)
My (limited, weak, non-scientific) understanding is that the relative retention times measure across instruments by identifying the IS (internal standard) and making sure the identified ions come out at proportionally indentifiable times. If the GCMS takes 15:00 and the IS comes out at 3:00, with golfballs, baseballs and besketballs at 6, 9, and 12 respectively; then if IRMS takes 00:15, then IS should show up at 00:03, with the balls at :06, :09, :12, to be identifiable. Or am I completely off base here? Dr. M-A testified on cross:
q: if the purpose of relative times is to compare two instruments, in most analysis using GCMS, why use relative.
a: to get absolute certainty on a run to run basis.
q: when doing simple gcms they use relative retention?
a: they can, also use absolute, also use mass spectra to tell if it goes off. in this instance the purpose is two fold, the GCMS, and require that reference to identify the peak on the IRMS, because you don't have the second level of identification. Need it go QC the instruments and identify, and this has clearly not happened here, it's 6% out.
q: if you were to apply this criteria to the GCMS instrument,
a: including the mass spectrum.
q: did you also talk about CIR of the internal standard, this morning?
a: yes. i did.
q: you know they don't use the IS for quantification.
a: they should use the internal standard in the matrix for QC.
q: if you hear they testify they only use it for identifcation... The delta differences you pointed out as outside their criteria, would only be outside if they used the standard for comparative delta values.
a: the problem is how they have identified the IS, is the time marker, but it's not even close to the time of the GCMS. How on earth did they identify which one out of 5 or 6 is that it. So did they use the delta value to say that's it, but it's outside their limit, on what ground do they say that's the IS?
q: where in their SOP do they say they need to do that?
a: If they don't, I'm baffled by how they identify the peak.
q: where is it in the spec it matters if it's outside the spec?
a: they must have some criteria, how do they choose? Divine intervention? I'm amazed. You're left with the GCMS, which has mass spectra, then you get to retention time, and it doesn't match in the IRMS. How do you identify one unknown peak among 5 unknown peaks? I don't know how they do it.
q: where does it say they quantify the IS?
a: how do they know it is the internal standard?
q: you've already told us your not an expert in steroid metabolism, so we'll skip
Is there anyone here with some explanation of how the test results could be correct? How the apparent lab errors wouldn't affect a conclusion that the results were positive? Certainly there appear to be explanations of why the results aren't valid, but I'm just wondering if there's a coherent argument otherwise, using the science involved. Thanks.
Anon 9:20 Thanks! That is what I originally thought it was. Now, does Brenna get it?
Chaz: to furhter your troll discussion there is the following.
According to the prosecution Landis had a win at all cost mentality (thus the Motherf****r comment and lemond et al.) which lead him to cheat.
The USADA according to Travis also seems to have a win at all costs mentality so does this not also mean that the USADA cheats?
The issue is not really one of cheating per se since cheating is rare in science. The issue is one of proper methods that allow one to make valid and reliable conclusions.
We may nevr know who cheated, if anyone. We can know if the conclusions are valid given the methods.
Sephen
Brenna "gets" 1.3 million dollars.
Bill
anon 9:35, I think the point is that these results can't determine anything... neither a positive or a negative. Basically, if I'm correct (so not a scientist), it's a little bit like saying I'm thinking of a number between 1-100... hey, you could guess 17, but it's purely a guess (however, given enough guesses, you are more likely to guess the right number, right?). That's how accurate this result is... there is absolutely no way to know one way or the other if Floyd was clean... the tests were so improperly done that all data from them is useless... So the results don't show that Floyd doped, but they also don't show that he didn't. They show absolutely nothing.
This brings up another interesting topic, the grant game. I saw another example of this in governement just yesterday. Independent University researchers get grant money to do research from the government. The government hopes the research will support there proposed legislation. When it doesn't government officials strong arm the researcher to not present the study during the public meetings on the legislation by the indirect threat against further grant money.
Remind me again how long we can expect to wait before a decision is reached by the panel?
How would you feel about LNDD doing you're anual cholesterol test. If they get certain values for LDL and HDL (both absolute and ratio are important) then your doctor is going to put you on Meds. Expensive, life altering meds. Given what you've seen, would you want a second opinion on the lab results? I sure would!
Bill or others,
When the panel allowed additional samples to be tested, why didn't they use another lab? It seems like this one ruling would have made this entire process so much more enlightening for everybody. I know UCLA made a lame excuse about the machine being down. Maybe they were afraid they would prove that LNDD was doing shoddy work? Surely, there had to be another lab they could have used to make the process much more fair.
Here's hoping Botre got it, and will admit it. With all of us out here not exactly having a handle on it, you have to figure that the panel has three completly different views on it too. That means they're going to ask Botre to explain it to them.
I suggest a diet high in asparagus to any athlete having their samples tested at LNDD... ya it might be juvenille, but there would be at least some satisfaction in it...
Bill, re: what Brenna "gets,"
TOUCHE' !
Camille
1½ years ago...
http://www.cbc.ca/sports/story/2005/09/09/lance_armstrong20050909.html
...and no action.
If UCI is committed to change, they must do something in light of testimony during these Landis hearings.
Quoting Jim...
Here's hoping Botre got it, and will admit it. With all of us out here not exactly having a handle on it, you have to figure that the panel has three completly different views on it too. That means they're going to ask Botre to explain it to them.
But Botre is the head of the Rome lab and we know he can't say anything critical of another WADA lab. The independent expert is anything but...
It was also very telling in Young's closing arguments (and before) that he avoided even using the term LNDD, knowing the letters already have a POOR connotation. Instead he would say "Paris lab" or "Paris technicians" (occasionally "French lab" but that also has negative connotations for many). But we all love Paris, non? It's okay to say "PARIS... lab" ... emphasis on the Paris part for all those romantic implications that word carries. Young's closing was full of equivalent word games. What else did he have? "Win at all costs" indeed.
Camille
So LNDD is content to use the SWAG method. Uh!
You3 said...
The testimony indicated that by ISL every peak needed to have a retention time that was sufficiently close (1% or 0.2min) to a corresponding peak in a reference sample. This check insures the right peak (elute) is chosen, the gc column is behaving correctly and there are no contaminants or interferences. But, what reference is to be used for IRMS peaks? If you use the GC/MS reference results, the elution times don't match. It turns out that there isn't even a good linear fit to the relationship according to my calculations. In cross, USADA addressed this issue by pointing out that the calibrations samples (MixCal?) could be used for reference times. But, Davis pointed out that some of the metabolites are not included in those samples, specifically Andro, 5a-diol and P-diol, in other words all the ones that matter for Landis. So, there is no way of establishing that these 3 peaks elute at a time sufficiently close to corresponding peaks in a reference sample.
I think that everyone believes that these peaks are mainly composed of the expected metabolite, but a key ISL check is missing with regard to proper GC operation, contaminants and interferences.
I have previously observed that it would be good to include both 5b-diol and 5a-diol in the calibration to validate that there are no unexpected interferences between the two peaks.
Given the clear problems with LNDD, I'd like to see the decision go Floyd's way, and I'd like to see the USADA so shell-shocked and fearful at the prospect of their crumbling world that they do not appeal to the CAS. Then Floyd needs to go to civil court to recoup a lifetime of lost earnings--in the tens of millions. . .I can dream, can't I. . .
After all is said and done, I would ask the anti Floyds this question: Looking at all of the testimony, and using a logical thought process intertwined with honesty, if your career was on the line, would you trust the results and how they came to be?
Anon 10:28,
Judge Hue has already pointed out that USADA would probably leave it as is, if they lose. Unfortunatly WADA and others can still appeal it.
I think that JPSMP gave a good explanation, so let me ask some questions.
The ISL has some rules, and then allows adoption of Technical Documents that have some more rules, have you read those? They are at the WADA site as part of their downloads?
Technicially, an athlete can't attack the science per se, ie, you can't say the test doesn't measure what you say it does, but some limited ways to attack it occur, such as can 5 alpha vary independently of 5 beta or the nandr0-19 issues of it can be produced endogenously.
So the procedure requires that the athlete show divergence of practise from the standards. In that case the lab then proves that the divergence did not cause the AAF (the burden shifts discussed by Judge Hue).
To get to first base under the rules, FL's team must show a divergrence from the standard. The debate between Brenna and WM-A and Davis matters in one respect as I read the TD2003IDCR, in that the only specific mention of retention time is in the section on CG. Retention time in the "tandem" method sections does not speak to the retention times as far as I can tell. Does anyone see a different interpretation than that? http://www.wada-ama.org/rtecontent/document/criteria_1_2.pdf
The three ion vs. one ion may have a better chance of being an argument that requires explanation. As best I can tell, the wada types didn't really have an argument for that except to say identification is not a problem because the pattern of the peaks match. As I read the td, the pattern of peaks may be sufficient match if one is looking for concentration, but cannot speak to identification of the peak to begin with. Am I missing the boat on that?
Lastly, IRMS is really not particularly covered in the discussion in relevant ways to me. Hence WM-A's question, if you don't have a rule to identify your results in IRMS what is telling you that you have the right peak identification, divine intervention.
But the catch 22 is, if there isn't a rule to break, there isn't a burden shift so the lab has nothing to defend. If you can't attack the science, ie, the whole thing is built on sand, its a house of cards because you have no rules to identify substances, you can measure them, but not identify them, and it is the right substance because I say so.
Again, I am not a scientist, I will be happy to hear explanations.
jr.
Didn't Davis say that Botre supervised the B sample testing? If so, did he not understand what he was doing? Scary thought if he is the one advising the arbs.
Thanks 10:32. I didn't catch where Suh clearly identified the specific violations of ISL rules; did he?
Are there ISL's for manual subtraction, reprocessing, and recordkeeping?
Jr.,
If I read the ISL, section 5.4, it appears to me that there are numerous rules that could be demonstrated not to have been followed, by LNDD's own admission, starting with section 5.4.4.4 on control of data (p. 34), and the requirement for the software to mandate adequate tracking of operator intervention. I think the whole question of method validation (Section 5.4.4.3.1: Uncertainty in identification) is brought into play with the retention time and coelution issues. Certainly, if the lab can't identify a particular peak in the chromatogram with certainty, and in compliance with their technical docs (TD2003IDCR), than I would expect that also to be an ISL violation. So I think the ISL is written broadly enough that there are any number of issues brought up in the testimony that would apply.
One thing that doesn't seem to be mentioned much is that the parties will be submitting Proposed Findings of Fact and Conclusions of Law. In this document, the parties have the opportunity to tell the panel what facts they think they established in the proceedings, highlighting the supporting evidence, and then show how these facts meet the legal criteria for the result they want. That means that Landis' team will have the opportunity to tie the testimony of their experts to the specific ISL violation they want to establish. I look forward to reading these documents, particularly the one submitted by Landis' team.
Rocket,
Before the hearing I WAS of the belief that Floyd had cheated. Now, I don't believe I saw a single shred of SCIENTIFIC evidence that couldn't be questioned because of shoddy, inept, biased work done by the Paris crew. I also feel that NO athlete will get a fair shake when it comes to the Paris lab and all other WADA accredited labs as long as the "Code of Ethics" is adhered to.
Brenna said this in the rebuttal cross-examination...
"on the gcms side we see a pattern, and then [stuff] and you see the same pattern here, and they have approximately the same look from this distance. So, I'd start with that peak, anchored on the internal standard, and can identify the last peak based on pattern, which is one of the ways that I'd identify peaks comparing these."
I assume that when he says "you see the same pattern here" he means in the IRMS plot. He seems to be saying matching patterns between the two plots is easy. Drs. M-A and Davis don't agree.
Can somebody explain in a fairly simple way what's so difficult about matching patterns in the two plots (assuming the time scales can't be matched up)?
Some specific questions would be...
Are there a whole bunch of peak patterns that could "have approximately the same look"? Wouldn't that depend also on the zoom factor? For a sequence of peaks are the relative amplitudes (from one peak to another) pretty much the same in both plots?
Thanks,
~ Cub (too lazy to do the research for myself)
Not a scientist, but I think the post discussing the LNDD's use of patterns to identify the peaks is correct -- Dr. M-A testified that even though the lab periodically injected the Mix-Cal acetate during the test to establish control values, because there was a sloping baseline, the control itself could be off. The lab did no further testing to confirm the control.
Also, during Dr. M-A's testimony, he referred to the lab's acceptance of error percentages that exceded WADA standards, not just ISL.
From DPF, and the best analysis I've seen yet:
The scientific testimony we saw during this hearing may confuse many people on this forum, leaving them wondering who to believe. I think much of this testimony, though, is ultimately irrelevant. Floyd’s lawyers, as to be expected, tried to point out as many flaws as possible in the case, to cast doubt on the credibility of LNDD. But even if LNDD has been sloppy in its procedures at times, the question is whether this sloppiness affects its conclusions, and in particular just one conclusion on which the entire case hinges.
One of Floyd’s metabolites, 5aA, has a highly negative delta value (relative to the reference pregnane), less than – 6, on both A and B. This value is so far below the criterion of – 3.0 that it makes irrelevant the question of whether an error value should be added to this criterion, as Floyd’s team has argued. Even if an error value of 0.8, or even of 1.35 as suggested by deBoer, is added, the 5aA value is still far below - 3.0. Furthermore, while only this metabolite has a value this low, directors of labs other than LNDD have indicated that they think a single metabolite is sufficient to base a doping violation on. And USADA has cited precedent that a doping charge can be upheld on the basis of IRMS alone, regardless of T/E results.
Thus Floyd's case comes down to two questions:
1) Could a value of – 6 result from chance, as a false positive?
2) Could this value result from some error in the analysis?
1) There has been a lot of discussion in this forum about the odds that a positive is a false positive. It has been pointed out that if most riders don’t dope, or if many riders who do dope successfully evade the test, then the proportion of positives that are false is relatively high. This is the trendy argument based on Bayesian statistics. For example, if 1% of the riders dope and test positive, and the false positive rate (% of non-dopers who test positive) is also 1%, then about half of all positives are actually false positive. So even though the odds of an innocent rider’s testing positive are quite low, the odds that a positive is false can be quite high.
The Puerto revelations suggest that far more than 1% of the peloton dopes, but it is unclear not only what % of dopers there really is, as well as what proportion of them test positive. For the sake of this discussion, I will use the assumptions made in an article on Bayesian statistics linked here recently. The authors assumed that 10% of the riders dope, but only half of them are caught. The % of riders who are caught, in this case 50%, is the sensitivity of the test, and the inverse of that (100% - 50% = 50%) is the false negative rate. They also assumed that the 99% of the non-dopers tested negative. This is the specificity of the test, and its inverse, 1% in this case, is the rate of false positives.
Using these assumptions, out of a population of 1000 riders, there would be 100 dopers, 50 of whom would test positive (true positive), and 900 non-dopers, 9 of whom would also test positive (false positives). The % of all positives that are true, known as the positive predictive value (PPV) or posterior probability, is 50/(50 + 9), or about 85%. The odds that a positive is false is the inverse of this, about 15%, which is considered unacceptably high.
However—and this is my main point here--the IRMS test adds a great deal more PPV. Suppose we test all the T/E positives, true and false, by IRMS. Suppose we again assume 50% sensitivity (though the IRMS is almost certainly much better than this), and 99% specificity, or 1% false positives. Out of 100 T/E positives, 85 are true positives, and 42.5 will test positive on IRMS. In the same group of 100, there are 15 T/E false positives, and 0.15 will also test as false positives on IRMS. The PPV is now 42.35/42.50 = 99.65%. This means that about 1 out of 300 riders who fail both tests are false positives. Even if the false positive rate were 2%, the PPV would be 99.3%, meaning less than one double positive out of 100 is false. If we assume the IRMS is actually more sensitive, that 90% of dopers test positive on it, then with a false positive rate of 1%, the PPV is 99.80%, meaning that one in 500 riders who fail both tests are clean. With a 2% false positive rate, the PPV is 99.6%. Even with a false positive rate of 5%, the PPV is 99.0%.
Now it does have to be emphasized that the weak link in this reasoning is that we have no idea how many dopers successfully evade the tests. But the point is, the IRMS test acts as a second filter to weed out false positives. Unless one makes the extremely pessimistic assumption that most or all dopers don't get caught--in which case the tests truly are useless--the IRMS test should greatly increase the PPV, that is, the likelihood that a positive is a true positive.
Based on these assumptions, let’s ask what the false positive rate on the IRMS is likely to be. On an earlier thread Duckstrap provided some estimates, by asking what the probability is that the criterion of – 3.0 delta units is met for one or more metabolites. While this is a valid and relevant approach in evaluating the fairness of the WADA standards, it is misleading to apply it to Floyd’s tests, because of that one metabolite with a delta value of – 6. Duck’s analysis assumes that -6 counts for no more than -3, when in fact it accounts for much, much more.
How probable is a -6? We can start by noting that no non-doping subject has ever been reported to have a T metabolite with a delta value this low. One of these studies, by Ueki and Okano (1999), tested more than 350 subjects. One of the most detailed studies, the article by Catlin’s group that has been extensively discussed in this forum, reported a mean delta value (difference between 5aA and the reference pregnane) of -2.09 and a SD of 0.63. The lowest reported delta value, - 3.72, was about 2.6 SDs below the mean.
Floyd’s mean stage 17 delta value for the same metabolite comparison was – 6.26. The control in this analysis had a value of – 1.76. If we assume a standard deviation of 0.63, as reported in Catlin’s study, Floyd’s value was more than 7 SDs below the mean. Even if we use a somewhat larger SD of about 1.0, as suggested by Maitre’s study, his value is more than 4 SDs below the mean. This is considerably less than one in a thousand, and indicates that we could throw out the T/E results entirely—act as if Floyd had never been subjected to this test—and still confidently convict him based solely on the basis of the IRMS.
This discussion makes several critical points. First, Floyd’s 5aA value is so low as to rule out, beyond a reasonable doubt, the possibility that it is a false positive. We don't even have to worry about how many dopers there are, and how many are caught, to draw this conclusion. Duck's analysis takes into account that each of the four metabolites could have an independent probability of being a false positive, so the probability that one of the four might be a false positive is substantially higher than the probability that any one in particular would be. However, the 5aA value is clearly so low that even making this additional assumption cannot account for it.
Second, because this value can’t be a false positive, it confirms the judgment of WADA scientists that a single metabolite with a value this low is a sufficient basis on which to conclude that doping took place. This judgment has recently been supported by studies, such as those at Cologne, indicating that it is possible to administer exogenous T and have only 5aA meet the -3.0 criterion. But the point is that even without these data, the – 6 value should stand by itself. If a single metabolite has a value so low it can’t be a false positive, then ipso facto it must be indicative of doping. The only way around this conclusion I know of is to argue for some unusual physiological effect, such as alcohol. But beyond the lack of strong evidence for this, the fact that this metabolite had a consistently low value on most other stages that were tested makes this kind of argument very difficult.
Moreover, a strong argument can be made that not one but three of Floyd’s metabolite values are indicative of doping. As I showed above, if a rider fails the T/E test, then additional failure of the IRMS test strongly indicates doping even if the false positive rate is very high, as high as 2-5%. A value of 2.2% corresponds to 2 SDs beyond the mean. So if a rider fails the T/E test and in addition has delta values of just 2 SDs beyond the mean, this is highly indicative of doping. Statistics suggest that less than one in a hundred clean riders would test this way by chance. In addition to 5aA, two other of Floyd’s metabolites, 5bA and A, had fairly high negative values. 5bA – P was about 2.0 delta units more negative than the control 5bA – P value, while A – 11K was about 2.8 units more negative than the corresponding control. Based on Catlin’s or Maitre’s data, both of these delta values were 2 SDs or more negative than the control.
This conclusion is not quite as strong as the one for 5a, because it depends on the assumptions I made earlier about the number of dopers and those caught. If the number of dopers is very low, or if most dopers are not caught, then the T/E test is largely useless, and the IRMS results must be evaluated strictly on their own. But if we assume the T/E test has the level of sensitivity that I suggested, then even moderately negative delta values, if in conjunction with a positive T/E, should be, from a scientific if not a legal point of view, sufficient evidence to convict.
2) If Floyd’s 5aA value is proof beyond a reasonable doubt of doping, the only other challenge his team has is to argue that this value was obtained as a result of some error in the analysis. This error must be systematic. That is, since a similar value was observed for the A and B samples, and highly negative values ( - 4 to – 5) were found for other stages, the value cannot be attributed to a single poor chromatogram, for example, nor to poor judgment in determining the portion of a peak to be measured. It has to be a regular procedure, carried out intentionally and presumably in good faith, on all the analyses. Moreover, just because Floyd’s other metabolite values are much less negative, this systematic error must be a procedure that was only applied to 5aA analysis, and not to the other peaks. And finally, since the 5aA values for the controls were much less negative, this systematic error must be related to something distinct about Floyd’s urine, for example, resulting from the presence of some substance that was not in the control urine.
It should be apparent that this is a very tall order. While several scientists testified on Floyd’s behalf that there are significant errors in the analyses, no one to my knowledge (I couldn't access the courtroom, so I followed the case through TBV and on the forum here) has provided a strong reason to believe that these errors are of the kind needed to account for the highly negative 5aA values. The most likely candidate, as we have discussed here before, would be a contaminant that co-elutes with 5aA and skews its value in the negative direction. We do know from the chromatograms that Floyd’s urine does contain substances not found in the control urine, so this is certainly a reasonable possibility. It is not by any means a frivolous argument.
As we have discussed in this forum, the complete mass spectra should be able to make a definitive ruling on this possibility. Last I heard, these spectra had neither been demanded nor provided, so I don't know where this matter stands. I doubt that they would account for the 5aA value, but it would certainly be nice to settle this issue once and for all. Assuming it were settled, and contamination could not account for the 5aA value, there should be no problem in concluding a doping violation.
I just niminated TBV for a blogging award. In two categories, actually. Best sports blog and best blog of all time. Go vote.
http://www.bloggerschoiceawards.com/blogs/show/16327
Nominated, even.
I won't post the thread, but it was noted on a DPF thread that Dr. M.A. was clearly wrong in some of his testimony, related to why he thought the test results were suspicious. He was using incorrect numbers.
Anon 12:17 informs us that a DPF thread comment stated that Dr. M.A. was "clearly wrong" in some of his testimony. Now if that were true, don't you think Young and his gaggle of WADA experts would have cross-examined him on it or put on a rebuttal witness to bring out those errors? Let's see, should I believe a DPF comment (probably anonymous) or the testimony of an expert that was subject to cross-examination and rebuttal? Pretty hard choice to make.
Anon 12:17
That does not surprise me. Although I was "born" on DPF, I will not likely got back too often as there are advocates there who resort to personal and warrantless attacks on people rather than on ideas.
The DPF argument is that Brenna spoke to the truth better than the doctor. Brenna is the 1.3 million dollar man. Dr M-A flew in on a private jet, at a cost exceeding $20, 000. You have to take what anyone says with a certain skepticism. DPF advocates vigorously contend in support of their world view in a context that seeks to preserve that world view. I suggest you try to synthesize the evidence yourself and come to your own conclusion. Should you conclude in Landis' favor, and choose to go there, be prepared, gird your loins and enter into battle.
If you can't decide, stay informed, consider the final written decision and see if you can come to your own conclusions that way.
I was just being lazy before. Here's the thread, and it wasn't an anonymous post:
However, the point made was that a discrepancy between metabolites was (almost?) never more than 2 per mil even among those who doped, and that was just one of the major reasons for questioning the lab results. That is equally as unlikely as one of them being at -6 by chance. Both are only remotely possible with normal physiology."
That point was indeed made by Dr. Augenstein, and he was wrong. The 5a values were -3 to - 3.5 more negative than two of the other metabolite values throughout the latter third or half of the study.
Someone at DPF who is commenting on the science in favor of USADA/WADAs position uses their own name? I do when there. Thomas A Fine, Carlton Reed (who else?) do and Floyd and Will, to their detriment, did too. As far as I know, no member of the "anti-Landis" camp posts at DPF under their own name, but I could be wrong. Any enlightenment would be appreciated.
People! Please stop pretending to be lab chemists! The problem here is FRAUD, not chemistry. LNDD submitted an apparently falsified document dated 2007, then scrathed out and redated 2006, pretending that the document had existed all along. LNDD claims to be supplying "scientific" analysis but they don't even know how to operate their lab machines correctly! LNDD claims to be supplying "scientific" analysis but it takes a courtroom interogation to get data out of them and they abide by a WADA code of silence ("ethics") that prevents any WADA representative from questioning what's going on!
This is not a problem of science. This is problem of fraud and incompetence.
dfp,
I agree with what you say and were this a court of law, the Landis case might not have survived a directed verdict or, if it went to fact finding, would not pass the requisit burden of proof.
HOWEVER,WADA didn't set this sytem up to do that. It set the system up to convict dopers. As a result, with the adverse analtical finding by a WADA certified lab (don't kill the messenger)Landis is PRESUMED GUILTY. The hearing was his attempt to be the first athlete in USADA history to sucessfully argue and "prove" that International Lab Standards were violated. But EVEN IF he is the first athlete ever to do that, all USADA has to prove to the Panel is that those violations DID NOT CAUSE the adverse analytical finding, that the fact tat Landis doped, did. So some examination of whether even 10,000 violations "caused" the positive RATHER THAN the fact that he doped. This is why I say that if the Panel determines him to be guilty, the WADA Code and WADA system may compel the finding.And that is also why the people here want to study and understand the scientific arguments. Garbage in....Garbage out works in life but not in a WADA disciplinary hearing.
Judge Bill Hue
I for one am grateful for all the scientific insights provided by those with some background. And of course many thanks to Judge Hue for keeping the focus on the actual legal procedure and standards that are being followed in this hearing. It's refreshing to sort through facts, rather than try to justify a position.
Michael
Is it correct that if there is an ISL violation with regard to preserving data of operator intervention in the manual reprocessing, that USADA cannot prove that the manual reprocessing was not the cause, because they did not save the data? Isn't that what Davis meant when he said they don't even know what they did?
damn, I wish I were not a math/science-tard!
bamalaw
I'm still chasing, trying to catch up. Did a long comment on the prior post on some elemtns of the case that I think are important.
Here I would like to admit that I am far from being a scientist and just want to ask two questions.
What is the likely effect of the lifting rings being left attached to the magnet. High School physics tells me that the magnetic field(s) would have been altered in what was supposed to be a highly sesitive machine?
Second, is there anyone out there who is willing to provide a testing protocol that they believe is resistant to the criticisms that the techno types are running at us? Or to generate a list of the documents/data which they believe would be appropriate to make a test result packet reliably reflect what was done with the sample to arrive at the test result? e.g. Should mass spec accompany IRMS results to establish what element/compound registered as a peak on the IRMS?
pcrosby
Here is a way to easily understand the problem with the IRMS relative retention time issue. Lets say you know when you drive your car, that when the speedometer points to 20mph, you are safe driving in a school zone and won't get a speeding ticket, that when it points at 35 mph, you are safe driving on a side street, and when it reads 55 mph, you are within the legal limit on the highway. Now black out all of your windows so you can't see any other frame of reference and let somebody else who can see the road drive, and replace the speedometer with one that looks exactly like your old one, except that it reads from 764 to 1751 and has no units on it. If you are told that 821 is safe in the school zone, what numbers would you assume are safe on the other two roads?
That is an example of a one point calibration. You don't know for sure that the 764 is equivalent to the 0 on your old speedometer. You also don't know for sure that the 1751 is equivalent to the 120 mph your old speedometer maxed out at. And finally, you don't have any basis to assume that the scale on the speedometer has a linear relationship to speed, because you don't know the units. All you know is that 821 = 20 mph. Anything you assume about any other speed on the dial is just that...an assumption. You have no basis in fact or knowledge to say one way or the other which speed on the dial represents 35mph or if 35mph even falls within the scale reflected on the speedometer.
That is what happened at LNDD. They used a single point to calibrate the relative retention times from the GCMS to the IRMS. Their retention times on the IRMS did not even come close to matching the GCMS. And they had ample evidence (supported by Davis' testimony in court) that the machine they were using in fact had problems with linearity. So they looked at their one point (the mix-cal acetate) and then picked the pattern of peaks that looked similar to what they saw on the GCMS, with no valid basis for assuming that the pattern would be the same on the IRMS. Let me repeat that. Without calibrating to more than a single reference peak, they have no basis from which to assume that the pattern will look the same on the IRMS.
In our example above, 821 represents 20 mph. What if 822 represented 35 mph and 55mph was out at 2000? That is a non linear relationship. What if 821=20, 1764=22, and 10000=35? Then we can't know if we are going too fast once we go beyond 22mph, because our speedometer doesn't go that high.
In the case of the IRMS, the ISL standard is that you have to have three points of reference. That not only tells you what the scale of the retention times represents, but also tells you if the relationship is linear. So if LNDD only knew what one peak was, and had no way of knowing if their scale was correct or if the relationship was linear, and the retention times were so far out of whack (minutes off), then regardless of what the pattern may or may not look like, they have no idea if the peak they see is one that corresponds to testosterone or asparagus in the sample. And for sure, they have no way of quantifying what they see, because the background has a slope to it that they have no way to quantify with only one known data point.
"damn, I wish I were not a math/science-tard!"
bamalaw,
I wish I were. This whole hearing [as marvelously reported and scrutinized as it was] served two purposes, for me:
It pulled the curtain on the ineptitude and broad-ranging failures of the LNDD.
It also remind me of my own weaknesses in math, biology and physics during my school years.
I think that OMJ and DS at DPF are still debating the numbers and how wrong WM-A was, given that the number to dispute this was the conference report of a non-peer reviewed in the normal sense of peer reviewed article not yet published. The same issues were discussed here on the day of the testimony about is a sample of 1 statisitcally significant, is it proof of how exoT actually metabolizes.
Both Shakelton and the other doctors really can't say its established science that the 5aA and 5bA move in the pattern that would prove FL is doping, one person is not enough. It is some evidence. And again, usually 2/mil except for one person in all people studied so far is not exactly overwhelming. And again, the literature hasn't been 'published' yet, who knows what will be the official word.
And out of all the hours the WM-A testified, you would disbelieve everything else he said about methodolgy because of one unpublished number?
PS DS, I will look at the lab standards again. On the coc issue, I think they will buy Ayotte's explanation that if in a secure area, no recording of handling is required. As for the manual changes, the argument was about process vs. once data was recorded as conclusion. I'll look at that.
jr.
Anon1:54, thank you, that was brilliantly explained!
bamalaw
As a cycling fan, but not an expert in lab science, not law, I find the most interesting outcome of this case to be the division it creates among followers of cycling: both passionate for different reasons, but I would argue that both sides would welcome some improvements in the tests.
I think the cycling fans side with the USADA/Young decision, and genuinely want to clean up the sport more. Landis supporters are interested in the intellectual chess battle of the science and legal matter and are less focussed on the sport as a whole. Pro-WADA cycling supports are upset that drugs are killing their sport, and pro-Landis people are challenging doping existence or ability to accurately measure that existence.
It is obvious, however, with all the doping mea-culpas in the professional cycling world (Eric Zabel this morning... many other prominent names of late) that this case is an anomoly, and definitely hard to accept for the fans. It is hard for people to believe that one guy, who was officially cited as having evidence of PEDs in his body, is different and actually not a doper, while a great many others are facing up and apologizing.
I do think this case could be a point of reconciliation. We all know there is a problem in the cycling world. Maybe people need to talk about it more, make better verifications so both sides (Scientists and fans) are happy.
It is regrettable that start athletes cannot come out publicly in support of their sport, and against doping more often (like Jason Giambi has recently done in baseball). When this happens, the Athelete is treated like an enemy, a monkey (?) or something else.
Unfortunately I think what people may take away from this case is: the system is flawed, you can't trust a government institution to do things right, and if you fight enough, you can find enough flaws to get cleared, so you may as well take the risk to dope, hope you don't get caught.
C Lafoy
5.4.4.4.1.3 The software shall prevent the changing of results
unless there is a system to document the person doing
the editing and that editing can be limited to users with
proper level of access.
5.4.4.4.1.4 All data entry, recording of reporting processes and all
changes to reported data shall be recorded with an
audit trail. This shall include the date and time, the
information that was changed, and the individual
performing the task.
DS, looking back, I think the first requirement is met so long as who did the work is identified and they have appropriate clearance, I think LNDD probably met that requirement.
As for Young's closing argument that was bothering me, he mentioned changes in results, and distinguished the manual reprocessing on that basis, but given the nature of the changes made by Frelot and Mongongu, I think that data entry would describe their actions and each change entered should have been documented. It was a list of three actions that required an audit trail, he distinguished one but not the others.
What do you think of 5.4.7.3 as another potential ISL violation?
jr.
CLafoy,
How about this. I support Landis, not because I believe he is or isn't clean, but because I would rather let a doper go than sanction somebody who is clean. If the lab work is flawed (and there is no question that it is), we should accept that, fix it, and move on. I don't think that all of the WADA/USADA system is flawed...I think this case would never have made it to arbitration if UCLA had run the samples. Then again, I think the system is flawed, because the burden of proof is on the athlete, yet the athlete's access to information and data is SEVERELY limited. I am a cycling fan, and I want a clean sport. But more than that, I want the sport to become clean while secure in the knowledge that I didn't wrongly sanction a clean rider in the name of trying to show zero tolerance to dirty riders.
Jr,
The problem is that the changes made by the techs were not documented as a part of the audit trail. I think that point will turn on the definition of "reported data", because reading that section very strictly would indicate that they can change the machine output all they want without documentation, as long as they leave the "reported" result alone once they are done.
jr.,
I still think that 5.4.4.4.1.4 is violated when the operators were changing the baselines and adding "data points" (I still can't believe that!) without documentation that they were doing so, let alone why they were doing that. "Using my experience" which is first discovered upon cross examination is not a reason. I might agree that 5.4.4.4.1.3 is not violated so long as there is password protection to get access to the system.
I agree with you that 5.4.7.3 would be another likely ISL violation insofar as they didn't show complete mass spectra to properly control for possible coelution. Additionally, they run appropriate positive controls contemporaneously with the analytical runs. I would think that 5.4.4.3.1, Uncertainty of identification also applies, with the lack of control for coelution problems. I am reading things in a cursory fashion, but I am sure their are plenty of specific things the Landis folks will find.
Jr,
OK - no debate here. I just think there is an absence of the bigger picture when you get too deep in the science. Restoring confidence in the tests and LNDD/WADA will be a challenge, but so will re-building the credibility of cyclists, even those not formally convicted. It would be great to hear more cyclists voice themselves on realities of doping evidence they have seen rather than keeping a code of silence. The the sport can rebuild too.
anon 2:54 and ds
I was bothered about the recorded data part of 5.4.4.4.1.4 when I heard Young's argument, but after more carefully rereading it today, I realized that "All data entry, recording of reporting processes and all changes to reported data" is a typical list separated by the conjunctive "and" so that all three separate and distinct terms in the list must be met if we are relying on typical rules of construction. So I think we have another Youngian technique which tries to frame the argument narrower than the language of the ISL. All data entry is part of processing, putting in baselines, setting limits of integration on the peaks are data points used by the software to do calculations. I don't see how you cut that out. So I am agreeing with DS that I think they can't change the baseline, background, limits of integration across peaks, etc. without an audit trail. And clearly, without the audit trail, we don't know if the original S17 as automatically calculated came up -6. something or another. And as I understand it, nothing was saved that could recreate that original value for S17. But I could be wrong on that last point. Anyone else positive about it?
anon 3:09
I think that while I haven't posted here much, pretty much everybody at DPF know I think FL doped. I can and do separate what I think about the system from what I think an individual did. I do on belief think its certainly more probable than not he doped. But I can also believe for different reasons that USADA didn't prove to my comfortable satisfaction that under the rules his test is entitled to a presumption of correctness. I can also believe that the system has some built in inequities and doesn't meet my personal standards of due process for a penal hearing. All of that is removed from believing many if not most cyclists dope. That more cyclists have to change the 'SOP' of cycling which has been doping for decades. That whole teams need to change their emphasis. I can also believe that will be hard as long as money is in the system, sponsors want exposure, riders want to make it big. You can berate the sport for being dirty, but you have to prove an individual is guilty to take away his livlihood. Those things are very separate in my mind. But I'm a lawyer so living in alternate realities at the same time comes easily to me.
jr.
jr.
Anon.12:01 makes a pretty convincing argument that the 5aA values alone are sufficient proof of guilt, particularly since similar 5aA values were shown to be present in samples from multiple stages. Assuming Floyd succeeded in shifting the burden back to the accuser (and that seems abundantly clear), would some tech-type reply to anon.12:01 to explain why the 5aA values alone wouldn't meet USADA's renewed burden? The multitude of LNDD's apparent sins is astounding, but doesn't USADA still get past St. Peter if it answers just one question correctly, and is that answser 5aA?
To Anon 1.49:
Here's a comment from the trade-only forum on BikeBiz.com...
just asked someone who knows a lot about these things:
"I've never heard of magnets being used to lift something as delicate as a mass spec (ours usually have lugs on the heavy bits for mechanical lifting), but even if they were...
(a) They would probably be electromagnets and, even left in place, would be switched off.
(b) Mass specs are carefully shielded against external fields, and any stray field effects would be picked up in the routine calibration, controls, etc...
(c) Sounds like someone clutching at straws?"
anon 3:57:
The 5aA numbers generated by LNDD do not match from day to day, and more importantly they are not even remotely reliable due to so many serious errors. One way to avoid the issue is by make assumptions that numbers generated by LNDD are accurate and then start a debate based on those "results." Goldberger, M-H, and Davis showed how the tremendous number of errors / problems at LNDD caused any number from any of the urine tests to be unreliable. Any debate that doesn't look at each of the errors / problems is simply a way to fool people who didn't see the hearing that the LNDD numbers are the "results."
Anon 4:43:
The metal lifters were left right next to a magnet crucial for the accuracy of the equipment.
Jr. 3:25
I agree with our comments on the sport and the pressures. One of the problems is - as I understand it - a doping cyclist who is caught is sanctioned, but no one else is. Human nature is all too often untrustworthy.
Changes need to be made to improve and assure that the science is good, the tests accurately and reliably carried out and the full record of a sample is available from the last drop into the cup until the final test results are printed. I see that as an easier task than dealing with human nature.
One way may be to impose sanctions on the manager of a team found to be employing a doper. As it stands now I think that the cyclist runs the risk, the manager gets to claim the podium with him/her, but can kick them to the curb if they dope. The managers are or should be able to track the development of their stable and identify those who are performaing or improving at an unanticipated rate. They are in a position to monitor what is going on and, if they are placed at risk, then the level of effective self-policing may increase.
This could induce mass paranoia among managers, but that is not necessarily bad. For some reason this does not offend my sense of due process as much as I expected. If a manager can put himself on the podium with someone he should reasonably suspect is doping, then I have less trouble with him going down with the doper.
For the sport to become cleaner will require some degree of self-policing. Experience tells us that fear of consequences are a strong motivator.
for Who Else?: My impression is that the lifting rings are not magnets themselves, but attached to the magnet in order to move the equipment into place and then removed. You may have answered my question anyway; maybe the process of preparing the machine for use inn testing would compensate for any effect. But, the designer indicated that there was a problem with the unit maintaining linearity even when it was properly set up and operated.
pcrosby
to anon 4:43-
The magent is a functional part of the Spectrometer. The metal "Rings" are what is attached to facilitate the lifting in place of the magnet. Leaving the rings in place can discrupt the field of the magnet and are explicity to be removed, or the mahince will not function properly, per the manufacturer.
The FACT that they were still in place goes to the staff's knowledge of the machine, thier training, and the reliability of the ensuing results.
If I were an athlete who tested POS at LNDD by CIR, and it was on the machine with the "Rings", I'd contest those results for sure.
Also, I'm not sure your friend truly "knows these things".
Annon 5:23.
Thanks, that is what I thought might be the case.
pcrosby
pcrosby,
The direct testimony on this issue is at 109.20-112.55 on 05/22/07PM
Regards,
ps: I am not completely obsessed, just had to find it to defend my recollection on DPF.
Off topic question for anybody who might know the answer - Now that the hearing is over is the gag order lifted? Can Floyd now post all documents and data received prior to the hearing?
Anon. 3:57,
Those comments are quoted from One-Mint-Julich, one of the more reasoned proponents of the "Landis doped" crowd on DPF. He and I have, and still do, disagree (respectfully for the most part), as to the strength of the evidence against Landis. The problem I see with his argument is based primarily on the physiological points brought up by Dr. Amory. In the prior literature to date, the metabolites 5aA and 5bA are strongly correlated, i.e. when one goes down in CIR, so does the other one. Generally, the largest differences are on the order of 2 o/oo even in people who doped. However, in the Landis samples, they appear to be uncoupled, and not by just a little, but by 4 o/oo, and this is not just on one sample but in all of B samples as well. The only evidence that this has ever been observed is that from a single individual in a paper (with peer review not yet complete) from the lab director in Cologne. I regard the occurence of this kind of physiological anomoly just as unlikely as One-Mint believes that a single metabolite can have a CIR of -6 without doping. His point presupposes that the analysis is conducted correctly, and I further believe that Landis' team showed numerous problems with the analysis, many of which are direct violations of the ISL.
Anon 6:23
The hearing is not over. It has been adjourned. The hearing will be over ~ when the decision is made public. The gag order is not due to be lifted until then. Expect a month to six weeks.
Jeff from Newark, DE
Regarding the "metal rings", I can't tell if it's clear from the above posts, but my understanding is that Davis testified clearly that the machine with the rings was NOT a machine used for any of the Landis tests.
The machine with the micky mouse ears was used on the re-tested alternate B samples.
Davis was quite clear about this in his original testimony - he was never offering it in direct relation to the S17 tests, but as an indication how little LNDD understood the machines.
Young had him repeat this under cross, which as reported by some outlets as "Davis admitted..."
TBV
Parting thought... because I think this is the end (until 4 weeks).
Judge Hue is a great guy and adept in making the complicated easy to understand... a rare skill indeed.
I like jr, he/she is very open to argument and even though I am on the side of (I think FL doped) he is reasonable guy...
Thanks for your coverage of this trial and I really think that it's great to have a bunch of people like yourselves supporting someone who is in trouble. I hope you continue to do so for the rest of Floyd's career.
C. Lafoy
PS: about the comments about being anti-French: Please try to be more reasonable people... The French in general love Americans... I am married to one
Post a Comment