Friday, May 18, 2007

Hearing - Friday Ayotte Cross II

We're coming towards opening the session, and attendance is notably lighter. No trouble finding a seat today. Much of the Press is gone. So far, the seats have me, Bill, Hiltzik, and NY Daily News, sent a day late perhaps.
There are some photographers, and a news video.

[MORE]


Paul and Arlene come in; she takes her seat in the well, and Paul sits in the audience. I found them in the hall earlier, shook Paul's hand, and gave Arlene a hug. She smiled and said "God is good everyday". Amber has an armful water bottles for the counsel table, and stumbles a little in her heels. She goes off and returns with another load.

We're told 5 minutes for arbitrators.

Henson came by earlier, and he pointed me to an email he'd sent Rant about the FFF and that it has never paid Geoghegan.

Young, Barnett and Dunn are calmly sitting down. Many have noted that Young looks and has the mannerisms and looks of a Walter Matthau in the late 60's. He comes back and talks to Brenna in the audience.

The seats in front the back of the well have, from left to right, Baker, Henson's seat, and Simon Davis where Geohegan used to be. Then Brent Kay, Paul, Arlene and Amber.

Suh is sitting with his seat pushed back from the table, leaned back, quietly composing himself, now engaged in some banter.

The murmur comes and goes in waves.

McLaren and Brunet arrive, no Campbell. They plug in. Campbell and Botre come in. Botre sits in the jury box in the back row with the guy who keeps time




BRUNET: Morning everyone.

Appearances: USADA 14.3/ Landis 11.6

CROSS BY MR SUH

q: uncertainty at lndd is .8
a: coverage of k2.

q: COFAC accred document had that that number? LNDD0098 EX 26.
a: yes i see.

q: this accreditation is where an exam of the proces,s procedure and verification is done.
a: yes, wheter it meets ISO.

q: the information in the document is important?
a: yes.

q: must be accurate?
a: yes.

q: what is this date on this document?
a: 15 Dec 2006

q: when was the tdf?
a: July aug 2006.

q: LNDD 0086 EX 26, you were not shown this page, yesterday, were you?
a: i know what it is.

q: you weren't shown yesterday.
a: no, but seen it before.

q: this says error is 20%
a: corrected by cofac.

q: correction done after the test was done, correct.
a: if you say so.

q: don't take my word.
a; 1 may 2006.

q: you said it was important their standards are accurate?
a: yes.

BARNETT: OBJECTION cutting off the witness.

BRUNET: accuracy is important.

REREAD.

a: what appears is the list of procedures that fall under the scope of accreditation. that is the way it should appear in the records of cofac. it is important that it is accurate. but if they made a mistake, that doesn't mean there is a mistake in the procedure in the lab.

q: but there is a mistake?
a: in cofac.

q: in the accreditation document?
a: yes.

q: what would a 20% uncertainty be?
a: it's an imprecise unit. you should use the units used to report the measurement. percentage doesn't have much meaning.

CHART OF REPROCESSING REPORT GDC 1350.

q: do you understand what this is?
a: please confirm this is the reprocessing the EDF.

q: yes, [walks through it]
a: yes.

q: you said in your lab, have a pdz eruopa and finnegan delta xp plus, and didn't know what software you had.
a: nope.

q: which is newer.
a: finnegan.

q: when did you get it?
a: 2 years, 1-1/2 years.

q: when did you get the europa
a: 1999

q: about 8 years ago.
a: yes.

q: at your lab you do manual background subtraction.
a: yes.

q: your techs also do manual peak identification.
a: I assume, yes.

q: don't assume.
a: I'm not the operator. They're told to make the necessary correction and integration.

q: you said yesterday, they do manual.
a: they do what they have to do.

q: do what they have to do?
a: yup.

q: both instruments?
a: yes.

q: i want to be clear.
a: I don't know the details of the operations.

q: I thought you said the same things done at LNDD were done at your lab?
a: I believe so, except for details.

BARNETT: OBJECTION

SUH: No more speaking objections Mr. Barnett.

q: now you are saying you don't know exactly what measures are used in your lab?
a: I'm not aware, I'm not the one operationg the machine.

q: this is a technical process.
a: yes.

q: how can you say they are doing what you are doing if you don't know?
a: I know we do reintegrate peaks, baseline, but how done on each instrument, I don't know.

q: so you cannot now say that what you do is the same as what they do technically?
a: digression

a: no, it's not even the same instrument.

BARNET: objections.

SUH: no more walking objections, [ pounds podium ].

q: what did you mean?
a: the actions should be the same. we do not rely on automatic generation of data by software.

q: just looking at the chart in the top, to the top values give you any concern?
a: i have a concern. while i'm satisfied with the manual reprocessing, i have a concern with the change from the e-11k and a-11k; the 5a and 5b look ok.

q: the ones that bother you are 2.01 to .35 and 3.51 to 1.61.
a: yes.

q: you understand what they are being analysed at the same time and same sequence at the 5a and 5b.
a: not same fraction or injection, but same sequence.

[ notes sequence stab, stab, stab, mix cal irms mix cal irms, mixcal irms, mix cal acetate, b3 b, f3 s, f1 bl, f1, f2 bl, b2, mix cal ]

q: this is the sequence.
a: yes.

q: which f has which.
a: f3 11k, f1 andro etio, f2 5a.

q: you sure?
a: maybe not. show me the data, i thought I knew

q: i was writing them down and none of them are correct.

a: f2: et/andro, f3 5a/5b, f1 is 11k's. I will take coffee and be better.

q: in this example of a sequence, in essence the f2 at the end of the sequence is a problem.
a: yes.

q: and the f1 looks odd.
a: yes.

q: but you think the f3 done first is good?
a: I didn't say that the analysis of 2 of 3 fractions are incorrect. I went back and looked... gets technical, and lost.

a: i have not seen how the integration is done.

q: because there's no record.
a: yes.

q: does the fact that 2.02 to .35 change so much cause you to wonder about the manual process by which the data is calculated?
a: If they are subtracted from the same thing. Yes, I am concerned there is something we cannot explain from the manual integraiton.

q: that's right. It was the same technician who performed this process and this one.
a: yes.

q: if you are concerned about how the manual process is done from the original result to the reprocessed, why wouln'dt you also be concerned by the same process done by the same technician on the 5a?

a: if we look at the global picture, there is a good coherent match. If we look at the auto, they agree. The only delta that is different is in the manual integration is the 11k. One difference on one peak doesn't create doubt on the rest of it.

q: because of all the values move approximately together, that's ok.
a: the first and second time, we have comparable data.

q: the same process is being applied to all of the manual originals and manual reproccessing?
a: obviously not, something is different.

q: the process is the same.
a: apparently.

q: same technician?
a: apparently, I was not there.

q: when concerned about an unexplained difference, the process that cause it is the same as used for those two?
a: and it's the same process used on the A to get the same values.

REREAD.

a: yes, and it's the same as used for the A as well.

q: a few more things to show. were you seated when we talked of log files.
a: not that much, I'm afraid.

q: GDC 1066. At your laboratory you delete files when the files dont' matter.
a: we overwrite files that are not useful to be kept.

q: it deletes the data?
a: yes.

q: 11:26:19 commencing mix cal acet. 12:09 saving mix cal ac 01. 12:22 commencing, 13:06 saving again
a: I'm afraid i'm not of much use. I have never seen such a log.

q: we have testimony the first was no good, so it was rerun.
a: i have to read every line.

q: is this the kind of thing you do in your lab?
a: they started the cal mix, stopped, restarted and reinjected, yes.

q: they stopped this because the data was wrong.

OBJECTION: mischaracterization.

q: in your lab, would you do something like this if the data was incorrect?
A: if we know there is something wrong with a cal mix, we'll stop and repeat.

q: would you overwrite the data?
a: yes.

q: which would delete the data?
a: yes.

q: but you don't know exactly what happened here?
a: no, we don't know what happened.

GDC 1073.

q: 13:09:02

[ reading ]

q: in your lab would you overwrite if you thought the data was incorrect.
a: yes. i'm not sure this is what happened. we may have changed the liner and needed to condition the liner. It's just conditioning. This record makes sense, they may have fixed a problem, we don't know.

q: would you be surprised if this wasn't a repair, but just data incorrect.
a: no. it's just the cal mix.

q: overwriting data is ok, and do it in your lab.
a: yes.

[ arelene it twitching ]

q: if you were to go out of sequence, that's ok, and that is done in your lab too.
a: yes, as long as blanks are intermixed with samples.

q: and that manual integration without records is ok, and you do it in your lab.
a: yes sometimes we keep the record when relevant.

q: CoC without interlab transfers is ok, and you do it in your lab?
a: yes.

q: rerunning mixcal acetates over is OK, even if the results disagree?
a: yes.

q: so you can't really that anything was wrong at lndd without criticising your own labs procedures.
a: that's not why I'm here. I'm not under the grill in this case.

q: if you criticse these procedures, you'd criticsie the procedures in your lab.
a: experience. If I say it's ok, it's based on scientific principles and its ok.

RECESS.


14 comments:

daniel m (a/k/a Rant) said...

FYI The comment from Henson's email to me can be found in the this post at Rant Your Head Off, along with Geoghegan's apology.

Anonymous said...

What do you have against the late Walter Matthau? :)

Ferren said...

Are you going to be providing a transcript, TBV? The video is not taking anymore users.

Anonymous said...

This woman with her accent and technical talk is kinda hot.

Anonymous said...

what's the password for today?

Anonymous said...

go to floydlandis.com it is all there


and yes that woman is a hottie

Anonymous said...

Did she really just say "it's ok, because I said so" ?

Anonymous said...

Unless I missed something, Suh just ate that womans lunch and made her pay for it. Wow.

Anonymous said...

Game On! ISO policy and procedures can be your friend or your worst nightmare!

Suh is still in the batter's box, just pumping the balls over the fence!

Forget yesterday's event...Today the truth is shining brightly!

~ Paul

Anonymous said...

Could you be a little more specific about great points Suh is getting in? I don't understand the significance of what he is doing.

Anonymous said...

He's establishing a history of incompetence and "bad science" at the LNDD. They're not implying anything was done intentionally, but that failures, in general, occur with regularity.

Basically, he's getting back to the actual point of the hearing.

Cheryl from Maryland said...

He is also establishing problems with Dr. Ayotte's lab. Dr. Ayotte said under cross that she really doesn't know how the machines operate -- that's for the technicians. So how can she validate the results?

Nor does she have a problem with overwriting the raw data with the "corrected" data. Amazing.

Anonymous said...

Suh is on his game and has just gone along way toward pulling the curtain aside.

Note that she says that her lab keeps a record of manual integration "when it is relevant". Another example of a judgment call by the technician.

Now, lets go back to some prior testimoney. One of the LNDD techs testified that when she was performing as the verifier of test sampling it was acceptable practice to just review the written record and output of the machine. In other words the person who is suposed to verify that the test is done properly signs off without knowing what was actually done and just looks for gross errors. And add in the record of LNDD taking back test results (I remember a post somewhere that indicated that the "hacked" material concerning that was accurate). Suggests that the verifiers were not doing a very good job of scanning the record of a test.

Would you want these folks doing your girl friend's pregnancy test? I really, really , really hope that they are not doing criminal case forensic testing.
pcrosby

Anonymous said...

pcrosby,

if you had a minimum scientific knowledge, you should be aware of such common practices.
About your girl friend pregnancy test, do you even know when it is positive ? Or "declared" positive ?

Sorry for you, no test on earth can be strictly automatic, forensic or not. That's a chance if you think about it.