DRUG TESTING ADVISORY BOARD

OPEN SESSION

March 11, 2003

Agenda Item: Welcome

MR. STEPHENSON (Chair): Good morning. I would like to open this session of the Drug Testing Advisory Board. This is the public session. We have some new members to introduce from the Board, and there will be an opportunity for public comments at the end of this open session. At the break, if anyone wishes to make a public comment, please let us know so we can schedule the appropriate amount of time.

For those of you who are new to the group, we had pondered how to begin this session. We thought about going with the oldest of the new members first, we thought about going with the wisest. We gave up on both of those ideas, so we thought we would start with John Irving.

MR. IRVING (DTAB member): My name is John Irving. This is my second time on the Board. I was on the original Board. I spent 21 years in the Navy, mostly with the drug program. Started in 1971, so I probably go back -- I am the oldest as far as drug testing is concerned. Set up some of the Navy laboratories, ran the Navy program, and was in the advantageous position of being in Washington when Dr. Walsh needed somebody here to help with the guidelines initially. I was involved in setting up the guidelines. Donna Bush is the one who kicked me out, replaced me.

I was with NIDA during the initial setup of the program. Since then I've worked for Roche Biomedical Laboratories, Aegis Laboratories, Psychemedics, and now I've set up a new lab in Memphis called Sure-Test.
What I bring to the table is a fair amount of history about the program, doing urine testing for a long period of time, a couple of years of hair testing, and someone who is not very quiet at meetings.

DR. REID (DTAB member): I am Ferguson Reed. I practiced surgery in Richmond, Virginia, for 25 years. Then I went with the Department of State as one of their medical officers for 13 years, while there became trained in addiction medicine, became a member of ASAM, certified as an addictionologist. After the State Department, I started working with the employee health programs with Dr. Ian MacDonald. I've been there for nine years.

It's a pleasure to be here. I think that I can relate with the people who have problems or the ones who test positive for drug tests. We are talking to them every day, so I can bring a different view to the table.

DR. FOCHTMAN (DTAB member): Fred Fochtman. I have about 20 years experience with drug testing and analytical toxicology. I have been a responsible person in a laboratory. I have worked for a small laboratory in Pittsburgh that developed drug testing back in the late seventies and early eighties; and then when the program started, served as a responsible person; but since then have taken a position in academics as an associate professor of pharmacology andtoxicology. For the last 4 years, I have served as chief toxicologist in a coroner's medical examiner's system. I have had a number of years in analytical toxicology and I have served as an NLCP inspector. I bring that experience to the table, and I'm anxious to see what is involved with this Board and to be helpful as much as I can.

MR. STEPHENSON: Thank you. We welcome the continued excellent caliber of individuals that have come to the Board to participate in these discussions and to help provide essential advice on the development of technologies and applications.

Agenda Item: HHS Update

MR. STEPHENSON: We recently participated in a meeting that was hosted on behalf of the Department of Transportation and the FAA. It's one of the few positive kind of collaborations that I have seen that's been in good weather. Normally we always go north in the winter and south in the summer, but this time we actually wound up in Tampa in February. I don't know how that happened, but we were thankful for it because it gave us a chance to convene in a place that had good atmosphere and very good discussion over a number of things that were critically important for specimen validity testing for specialized populations.

At the conclusion of that meeting and as a consequence of it, we made a decision to withdraw the imminent signing of the final rule for SVT testing applications for Federal employees. We did that in the spirit of trying to better review and understand all the elements that were presented at the conference and to go forward with a consolidated plan to the best of our ability that would reflect what we knew as a result of that conference as well as other things.
As a part of that process we have now incorporated the SVT elements into our comprehensive rewrites of the testing guidelines that include alternative specimens. Those have been forwarded to our Office of General Counsel for initial review and we expect comments as expeditiously as we can and processing those as soon as we can.
There is not a final answer that we have on SVT and we are working with DOT to look at alternative interim procedures and those will be discussed later.
Second, we are in the process of establishing a new contract for the National Laboratory Certification Program. We are in the process of looking for sources and interested parties. There will be a forthcoming Request for Proposals (RFP) that will be released in the near future. Under this new contract, we are looking at quadrupling the work of the NLCP efforts by tasking and relations of activities because we are going to incorporate all of the other elements needed for alternative specimen identification, testing, quality assurance, and so forth. If you are interested in looking at this, maybe not as an offeror but just to know what is going on, pay attention to what is coming out. The timing for all of this is very critical. We are at a point where the funds that have been used to support alternative specimen testing and SVT work and the emergency inspections that we had taken on over the last couple of years related to SVT have generated us right up against the firewall of funding. We are at a point where we know that we can get to the end of the existing contract, but we are going to have to be very careful. The work that we are doing has to be very carefully organized. We are not likely to take on any new exotic projects between now and the middle of August, when we expect to have the new contract in place. Having said that, we will always respond to urgent things that are out there and we will worryabout how to pay for it later. That has been our history.
Lastly, I want to address the group in public session about the continued blatant marketing, manufacture, distribution, and use of adulterants in this country and the process that we have had to go through as a Board and as a national program to respond to the deliberate efforts of those who would wish to suborn the testing. I do not have any answers for you, except that I am proud that at least 9 states are taking the charge on for themselves as individual states to address the issues at different times to look at this and hopefully that we can collectively identify opportunities to block some of those enthusiastic efforts that seem to be continuing. They evolve, and folks continue to say that it doesn't make any difference what specimen you use and it doesn't make any difference what test you choose to identify; we'll simply work through the process and make changes and stay ahead of you. What that means, I think, is a part of what we do every day. It's a challenge that we face every day.

Agenda Item: DOT Update

MR. EDGELL (DOT): Since the last time that we met as the Drug Testing Advisory Board there have been a few things occurring in DOT that I'd like to update you on. We had issued a notice of proposed rulemaking on the 30th of September, with the comment period closing on the 14th of November 2002, regarding the DOT drug and alcohol management system, information system reporting.

We proposed to streamline the annual reporting of drug and alcohol data to the operating administrations through the use of a one-page MIS data collection form. Our efforts in this area were meant to standardize across the OA's the information collected and to reduce the amount of data reported by transportation employers.
The comments that we received -- I believe we had 21 -- the majority were in support of the rule for the reasons stated in our original purpose. A few were opposed to it. One was opposed to the timing as far as the release, when it would be implemented if we went ahead with the final rule, and I think a couple were opposed to the change because it required additional or new programming software changes in the industry. We are still in the middle of rulemaking on this subject. We have not issued a final rule. That is imminent.
Next, on October 1, 2002, we issued a final rule regarding the addition of a breath tube that would be used for non-evidential alcohol screening. At the time we issued our changes to Part 40 in 1994 and then again in 1995, we said at that time that if the National Highway Traffic Safety Administration approved devices using different technologies, but still capable of detecting the presence of alcohol at .02 or greater, then they would be added to the DOT rule. There was a breath tube that was added to the NHTS conforming products list for alcohol screening devices, so this is just a screening device used on the first test, not an evidential device. And we crafted language to allow use of the breath tube at the DOT workplace. The instructions paralleled those for the saliva device in that both devices prohibit use after certain expiration dates and both have procedures for conducting additional tests if the proper procedures for that device cannot be followed. In other words, you drop back to an evidential breath device. That was released as a final rule on October the 1st, 2002.
An update on the DOT Advisory Committee. We published a notice in the Federal Register this past Friday (March 7), the intent to hold the second meeting of what we predict willbe 3 total for the Department of Transportation Electronic Transmission and Storage of Drug Testing Information Federal Advisory Committee. This public meeting will be held April 7 and 8 at the Embassy Suites Hotel, Crystal City. The purpose of this committee is to recommend to the Department the type and level of electronic security that should be used for the transmission and storage of drug testing information and to assess the type and format and methodology that would be appropriate and recommend those, make those recommendations along with electronic signature technology to the Department for input into a subsequent notice of proposed rulemaking.
Mr. Stephenson mentioned the task that was assigned to the Department of Transportation by the Senate Appropriations Committee. It was a drug and alcohol validity testing study. I'll just read the committee language: "The committee was concerned over reports that some categories of transportation employees could inadvertently fail to meet the current validity standards due to treatments for certain health-related issues, working conditions, or dietary habits."
We convened a meeting in Tampa, Florida, from the 4th to the 6th of February 2003. We convened a meeting with a number of forensic toxicologists, clinical chemists, nephrologists, nutritionists, occupational and environmental medicine specialists. The Federal air surgeon was there, several attorneys, and program specialists, looking at the issues around specimen validity testing. There were specific speakers in the areas of working conditions, dietary habits, medical conditions, and treatments. We probably had a total of 20 to 25 speakers. It was an excellent conference and the outcomes of that conference have created some changes, both at DOT and HHS, as Mr. Stephenson explained, and the report from this meeting is in preparation now, with our goal of before the end of March to get it back to the Appropriations Committee subcommittee.
Last item, as of March 1st, the U.S. Coast Guard and the Transportation Security Administration are now members of the Department of Homeland Security. The TSA was only with us for one year. The U.S. Coast Guard has been with us for many years and very entwined with the workings of DOT.

Agenda Item: NRC Update

DR. WEST (NRC): I am with the U.S. Nuclear Regulatory Commission and part of a relatively new office that's been formed, the Office of Nuclear Security and Incident Response. I have responsibility as the program manager for our drug and alcohol testing program relative to the licensees, specifically the nuclear power plant licensees of the NRC.

I just want to give you a brief update since the last meeting on our rulemaking effort with regard to our Regulation,10 CFR Part 26, also known as the Fitness For Duty rule. It seems that each time I give you an update, it's essentially a status. We have recently, towards the end of last year, completed almost monthly meetings with various stakeholders on the draft rule language.
Those meetings came to an end around October of last year and since that time we have been moving forward with the proposed rulemaking package. There are a lot of pieces to a rulemaking package and a very tedious process, but we're moving forward with regard to having a proposed rule to our Office of the Executive Director, the Executive Director of Operations, by June of this year for a proposed rule.
I speak we are at the juncture of pulling together all the pieces of the package to get it out for what we refer to as our office review and concurrence. We get it out to the various offices within the NRC, Office of General Counsel and so on. That's the specific point that we are at now. We are on schedule. We are trying to stay on schedule. It's a challenge, with many other things that are competing for staff time and resources. But again, we're expecting to have a proposed rule for drug and alcohol testing for our licensees, principally the nuclear power plant licensees, to our Executive Director of Operations by June of this year.

Agenda Item: Reporting Table (Lab A - Reporting Result(s) on a Single or Bottle A Specimen)

DR. VOGL (HHS): I hope everyone picked up a copy of the reporting table. The title is "Lab A - Reporting Result(s) on a Single or Bottle A Specimen." "Lab A" refers to the laboratory that receives the single specimen (which is being used by many of our Federal agencies) or both bottles from a split specimen collection (i.e., Bottle A and Bottle B) from the collection site.

I would like to give a little background as to why this table was developed. In September 1998, we issued Program Document 35. This document provided guidance to the laboratories on reporting specimen validity test results. It was based on the Federal custody and control form (Federal CCF) that was being used at that time. That Federal CCF had limited spaces and boxes to report the results. There was also a DOT memo to MROs that accompanied the program document. The memo described how the MRO was to interpret the specimen validity test results reported by the laboratory.
On July 28, 1999, we issued Program Document 37, which provided additional guidance in the form of questions and answers for the laboratories on how to record results using the old Federal CCF.
After that time, we started working on a "new" Federal CCF. We published a Federal Register notice on June 23, 2000, which contained the new 5-part Federal CCF. We had revised the laboratory results section. We included additional boxes for the different types of results that laboratories were reporting out. Unfortunately, the Federal Register notice did not provide any guidance on how to report the results. It had the new CCF and you could see the information that went on the CCF, but it did not go into detail on how to report results.
After that time, DOT in its revised Part 40, issued December 2000, required the use of the new 5-part CCF and implemented a policy which permitted reporting multiple results. So we have a new CCF, we have a requirement to report multiple results if you find multiple results on a specimen, but no specific guidance on how to fill out the CCF.
In 2002, laboratories were also starting to use electronic reports. We needed a way to ensure that the information on the electronic report was consistent with the information on the Federal CCF.
We handed out a draft of this reporting table at the SOFT meeting in Detroit in early October 2002 and received comments from individuals at that meeting. We continued working on it and were successful in coming up with the version that you have. We sent it out on January 14 by email to all the laboratories and inspectors. The table reflects -- and this is important to keep in mind -- it reflects current policy.
Of course, when you put out anything in a final form, you immediately receive questions. The following week we received a dozen or so questions from several laboratory directors and, rather than just answering each question and sending it directly to each person, we created a series of questions and answers as a file and sent it back to all the laboratories and to the inspectors. That seemed to work quite nicely. Essentially, it must have answered all the questions about the reporting table because we have not received any additional questions, or they are afraid to ask.
If you look at the reporting table, the first thing is it covers every possible combination of results - whether it can happen or not. We just wanted to preempt the possibility that somewhere some crazy combination comes up and we do not have it on the reporting table. We did make it all inclusive from everything we could think of that could possibly happen.
We are trying to standardize the comments that the laboratory is required to put either on the electronic report or on the Federal CCF. We have a few explanatory notes which we felt were important.
Lastly, on an invalid result, we are expecting the laboratory to call the MRO to discuss the result and the comment that is on the report. This will ensure that the MRO fully understands why the specimen was reported as an invalid result.
The top half shows the boxes that are checked on the Federal CCF. If you work across, Lab A is the lab that's doing the testing. The first column is the first result obtained for a specimen. If the laboratory is doing specimen validity testing, which is still optional as a policy, and the laboratory has an additional result for that specimen, then the third column indicates the boxes they would check on the Federal CCF.
The fourth column indicates if you need a remark, and the bottom half tells you what remark must appear on the electronic report or the Federal CCF. These are to the best of our knowledge, current policy.
The inspectors and laboratories have the reporting, and we are making a best effort to ensure that they comply with this, and it standardizes everything that is on the CCF or electronic report.
Hopefully, over time, the need for the lab to contact the MRO on invalids will probably decrease as MROs become more familiar with the reasons why a specimen is invalid and they understand what each comment means. We will work on that over time and see how it goes. But right now we believe the labs and MROs should discuss the invalid result.

MR. STEPHENSON: As often happens, when you think you have the final result -- and you alluded to this earlier. You said as soon as you come up with the final form of any document that immediately raises questions. The issue of your belief that over time there would be issues that would come up, how about right now? The fact that there is an interest in making a comment about one of the elements on the reporting table that DOT wishes to address at this time. I have no problem with doing this. This is the process that we have used successfully. This is a group of interested parties and this is where discuss or to try to update things that are going on.

MR. EDGELL: Inadvertently, with this table, we have a problem and I will point that out. In the top left-hand block you have "negative" and then you have instructions on checking an invalid box with a negative specimen. I am not sure how the confusion developed, but let me say this: that a specimen that is negative means that it has tested negative for the 5 drugs of abuse and it is acceptable with respect to specimen validity testing. By that I mean that it has not beendetermined to be adulterated or to be substituted. It could be dilute.

Negative and invalid is akin to saying that a specimen is negative and positive. At the end of a street at a T, you either go left or right. The light is either on or off. The computer bit is either zero or one. The patient is either alive or dead. The specimen is either negative or non-negative, to use a term that we coined in the new Part 40 that is used with your NLCP checklist. You would not under any circumstance have a negative specimen that was also invalid.
There is also a note that the laboratory should call the MRO to discuss an invalid result. I think that that as an idea to initiate is quite acceptable. However, our Rule, 49 CFR Part 40, instructs the MRO to contact laboratories to discuss with the laboratory the possible explanations around an invalid specimen.
If this chart is used by laboratories, then they need to take my comments into account, and my suggestion would be to correct this. If there's anyone here that is going to take this handout at this public meeting and print it in their newsletter, let's say, then by all means reflect the comments of DOT with respect to that negative and invalid result.

DR. VOGL: If you look at the top part, in the third column, if you go across, negative and invalid, you mark only invalid result on the Federal CCF or on the electronic report. You don't mark both the negative and invalid boxes on the CCF. I agree with you, you cannot have both. A negative implies it is negative with no other problems.

For invalid, if you look down for comments required, there are many cases where the immunoassays may be negative for all 5 drugs, but it is invalid for some other reason. Therefore, I think is still acceptable because you are not marking it negative and invalid.
As far as the requirement to contact the MRO, we have a "should call the MRO." If it does not happen, "should" does not mean that it is going to be rejected and action is taken against the laboratory for not calling the MRO. It is a "should." We would like it. We want them to do it, but it is not necessarily mandatory. If MROs want to call the laboratories when they get the invalid results, fine. We are not opposed to the MRO calling the laboratory when they get this result. We want to make sure there is contact. We want to make sure they talk about the results so that both parties understand exactly what the comment means and what additional information they have on that specimen. This way when the MRO discusses the result with the donor, he or she has a better understanding of perhaps what information the donor is going to find which may either, somehow beyond my imagination, support that invalid result with some legitimate reason or, you know, the MRO can feel comfortable actually telling or informing the employer that perhaps another specimen should be collected immediately. It is an issue that will take care of itself over time as both parties become familiar with the table and the comments that are on it.
A laboratory may make a mistake on a specimen. They may incorrectly mark the negative and invalid boxes on the CCF. That may happen. Hopefully, there still should be contact between the MRO and the laboratory to discuss the result, and they will find out why they marked it negative. If all 5 immunoassay test results were negative, they would provide that information to the MRO.

MR. STEPHENSON: Ken, did this satisfy any of your concerns in terms of Walt's explanation, or do you want to have this addressed in some other way for the Federal CCFs that are used in your process?

MR. EDGELL: This table is incorrect. That is as clear as I can put it.

MR. STEPHENSON: Okay. So the issue of Walt's explanation -- as I said, this is where we come sometimes, and it is not over time; it is now, and it is addressed. And if words have been said by DOT to reflect their issues and concerns, then we will work to try to see if we can resolve that.

MR. IRVING: Actually, as a laboratory I read this to reflect how you want it reported. Once I check invalid, I will never report a negative result. The way I read how Walt's done the table is I have invalids for two different reasons. One, the sample's invalid for pH or some other reason and I still have a good immunoassay; I still report that out as invalid, no comment about the negative immunoassay. Or I could have a sample that meets all the other criteria and my immunoassay is invalid; again, I report out invalid.

We read the table to cover all the possibilities. I don't think the laboratories would read this as the ability to check invalid and negative. I think that's pretty clear to the labs. Once you check invalid, the only other result you can put out is a positive result along with a valid, never a negative.
I think what you want is how, at least as a laboratory, we interpreted it. It may need some clarification, but as far as the lab's concerned we would report it out exactly the way you want it.

DR. CHILDS (DTAB member): I wanted to add a comment about invalids -- what our findings have been. For example, we actually have been able to identify, for example, a positive cocaine by GC/MS, but perhaps there's an interfering substance, not able to identify the cannabinoid that has screened positive. What we're finding from the MRO discussions we're having is a tremendous amount of confusion that the report says positive but invalid, and they go: Well, if it's valid for this, why isn't it valid for that?

My question is why not be able to report the GC/MS that was interfered with in the report that goes the first time? For example, we were able to confirm the cocaine, but the THC -- I'm sorry. Is that better? -- that the THC is unable to reconfirm. It gives the information in the report. Because I can tell you that where we are reporting out maybe several invalid samples a day in a laboratory, good-size laboratories, it is quite a burden to try to make a call to an MRO. They have an answering machine; they call back. We are on voice mail; we call back.
It's really a very challenging opportunity to do those discussions from the front end, and even the follow-ups, to try to give them the information that they need to have to make the decisions they need to do.
My question is, why not put the information in the electronic report or on the hard copy report when it's first going to them? That first communication should be where that information takes place.

DR. VOGL: You are still reporting cocaine positive if I understood correctly, and the invalid result. You are doing both?

DR. CHILDS: Right.

DR. VOGL: On the invalid result, it would have a comment, "GC/MS interference," because you said there was some sort of interference.

DR. CHILDS: But we can't say what drug.

DR. VOGL: Right.

DR. CHILDS: And sometimes that drug is very critical because it gives them an idea, oh, it's this kind of an adulterant. We have tremendous experience that certain adulterants are targeting certain types of internal standards.

DR. VOGL: That is why we have the asterisk. There should be contact between the MRO and the laboratory to explain what this interference is. Verbally, they can talk about each drug and explain what the problem is.

MR. STEPHENSON: Let me suggest that before we spiral down into the kinds of issues that are likely to happen here, that we do two things. Number one, recognize that no good deed shall ever go unpunished and that, despite our best intentions to clarify, we have not, and that, instead of a final, this is another version of a rolling reality that happens in our world and will continue to happen despite all of our best intentions to finalize anything.

Having said that, take into account the comments that Ken has made on behalf of DOT, and we will look at how to address some of these issues, and recognize that this is simply another opportunity the dialogue on issues that are incredibly complex and have a truly infinite number of opportunities for messing us up despite our intentions to get it right. That's just the nature of the very complex business that this is and our natural desire to simplify. Sometimes they come on collision courses and don't allow you to resolve these things the way you'd like to.
I am willing to accept that. I have to live with that every day. But again, the whole point of this is our life is made complicated because in this country we have a certain freedom that exists nationally that says that you're allowed to manufacture, distribute, and use with impunity adulterants and you're not held accountable for your own behavior, and it's not a lost opportunity for those who want to make money in the process of selling and distributing it.
Our life isn't going to get any simpler as long as these issues are out there. I'd like to move on.

Agenda Item: Alternate Matrices : II. Analyzing for Opiates

DR. BUSH: If Bob Stephenson was going to say we're going to move on to something simpler, we'll see, because we are going to take the next step along our path of exploring and evaluating alternative specimens and matrices for drug testing.

Over the course of time, in the past two Drug Testing Advisory Board meetings we have evaluated the pilot performance testing program information gathered by RTI and provided by laboratories interested in participating in the pilot program. We looked at marijuana, we looked at cocaine and their metabolites. And today we're going to take a look at opiates.
I'm going to turn the podium over to John Mitchell. He gathers all the information andhas the mind to put it all together and hopefully share it with us in a smooth style that we can all understand.

Note: The following presentation is found in a PowerPoint format under:

Workplace Drug Testing Documents and Publications

Regulations/Guidance Under Development

DR. MITCHELL (RTI): I would like to make one thing clear. It's not my mind that has brought all this together, because I've had to have help. I recognized when I was looking at this again last night that I had made the fatal error of not putting in a slide to recognize those individuals who worked with me to make this possible, and there's a lot of people at RTI. There's Andy McDaniel and Dale Hart who work with me a lot on this; and a new member of our staff, Dr. Esposito, who recently joined the staff at RTI. Of course, HHS has had to review and go through this and point out my errors or omissions. It's a team effort.

There are a couple of things I want you to look at as we go through the opiates. Those that have been here through the other ones, realize that this is a long and tedious task. One of the things that we need to look at very carefully as we go through this is variation. In other words, of participating labs, how is the variance? We have looked at the variance and we find it to be rather large. And people say, well, does it mean the drug's not there? No, it doesn't mean the drug's not there. Well, why do we have this component of concentration, being able to determine the concentration of the sample? A lesson John Irving and I learned a long time ago in the Navy is that one of the first things that will tell you whether or not a lab is operating correctly, if procedures and the methods are correct in that laboratory, is their ability to quantitate accurately. And if it's not, as we found in the Navy system, we had to go through and we had the standardize procedures in order to, as a system, have an acceptable variance among the laboratories for quantitation.
From a PT standpoint, quantitation will be the first thing that tells you whether or not a laboratory is having a problem with their system, whether be it a sample switch or whether be it an individual who is having problems carrying out the procedures within the laboratory, or whether there is some other problem in their systems. Identification. If you have two samples that contain THC, if you do not know the quantity of those two samples, they can be switched. They're still positive for THC. They can be switched, but you would never see the problem within the system. Remember that as we go through this morning.
Next slide.
The source of the analytical values we're going to be looking at this morning come from two occasions of maintenance PC samples for the urine, which is cycles -- I mean, Occasions 61 and 62. We have had, as you remember, four pilot PT cycles for hair, and we are only taking the six labs, data from the six labs that made it through all four cycles.
Then with oral fluids we are looking at 9 labs, data from 9 labs that completed all 3 cycles.
Next slide.
In the confirmation we need to know a little bit or we want to keep in mind what type of methodology is being used. With urine it's fairly uniform. We use GC/MS. Quantitations are in nanograms per mil. We're looking at, among the analytes, anywhere from parts per million to parts per billion, for those who are used to seeing analytical values in things other than nanogramsor per mill.
Hair, mostly the labs are using GC/MS. We have two labs that are using the quadrupoles, the MS/MS technology, which is a step above GC/MS in its ability to give you sensitivity. The quantitations in hair have been picograms per milligram. That's what we've required them to report. And that is in the parts per billion range.
For oral fluids, most of the labs again are using GC/MS, and again we have two that are using the MS/MS technology. Again, the quantitations, just as they are on hair, are in the parts per billion. We require them to be reported as nanograms per milliliter of the neat oral fluid that was provided.
Next slide.
To give you some idea of the sensitivities that are required, for codeine and morphine if we set the sensitivity required in urine to be 1, then in hair you would need 1,000 times that sensitivity, in oral fluids you need about 1600 times the sensitivity that's required for analysis in urine.
So you can see automatically we're going way beyond the sensitivity that's required for urine when we go to these other two matrices.
Next slide.
For 6-acetylmorphine, we see that the sensitivities are a little bit closer. If we consider 1 for urine, it would be the sensitivity required for hair would be 5 times that.
In the oral fluids, it is 4 nanograms per mL, so they need a sensitivity of about 80 times that that's required for detection of 6-acetylmorphine and urine. So that's just to give you some idea of the relative sensitivities that are required for these methodologies.
Next slide.
In evaluating we have used the same parameters that we used over the past two presentations. We have graphs of the reported values given by each of the labs or reported by each of the labs versus the mean, and we show the values relative to 20 and 50 percent quantitation ranges. I'll point those out when we get there.
We also show the relative position of the cutoff, and we'll look at the distribution below and around the cutoff. The variance, again we're working with industry-agreed upon cutoffs. We're looking at means of reported values calculated for samples with three or more values.
Sometimes when you're working with different requirements within the PT program we required labs to -- or we asked labs to analyze samples two ways and some of the labs didn't have sufficient amount of material to analyze both ways. So we would only get one value, and I'll talk about that as we go around, as we go through this.
We also talk about the percent coefficient of variation. That is nothing more than a mathematical term that we use for measuring. It's an index of variation.
Next slide.
So we'll look at the urine first since it, quote unquote, as Donna has often said, is the -- is it gold or the platinum standard.
One of the things I'll point out is that, do we have variance within the program from time to time within the urine? Yes, we do have labs which will have a value out, as you can see up here. For morphine we have a lab which had a concentration that was way out. Now, what happens is when we have concentrations that are much above the cutoff, labs will often do a dilution. So in this case we had a lab which had an error in their dilution problem or in theirdilution system, and so they ended up with an analytical value which was way out of kilter.
So within the NLCP we went through that lab and went through a process of evaluation. Now, was this sample negative? No, it wasn't negative. The problem happens to be within the system when they go outside of the normal routine in the dilutions. Dilutions are not normally made for any sample that has values in this range. It's when we get above the linear area of the immunoassay, because that's what they use to determine whether or not they need to make a dilution.
But if we do not have strict standards which allow us to determine whether or not that value is outside of the normal range or outside of what is expected, then we would have no idea that they're really having a problem. As you can see with urine, the labs have gone through proficiency testing and performance testing. They have their systems in line such that they normally obtain values which are very close.
The blue line or the blue dotted line represents plus or minus 20 percent of the mean. The red is plus or minus 50 percent. So this is the data from morphine.
Next slide.
When we look at the coefficient of variations for the two occasions, we can see for morphine that, on looking at all the data, the coefficient of variations is 10 percent or less, right around 10 percent or less. And that's because we're a mature system. We have not always been there, as those who have been in this program for a while. But this is a mature system.
This is the type of thing that we want to be able to mimic in our alternate matrices as we go into those programs and as we begin a certification program for the other matrices.
Next slide.
6-acetylmorphine. As you can see, the variation as we get around the cutoff is very tight, at 10 nanograms per mg. We do have some variations as we get further out in concentration. 6-acetylmorphine, as everyone knows, is a difficult assay, a difficult assay to spread over a large concentration range. It's easier to get accurate quantitation close to the cutoff.
But in this case, our laboratories do show an acceptable range. But again, this is the most sensitivity that's required of this system, is with this particular analyte.
Next slide.
When we look at this, on each side you can see what the sensitivity looks like. It looks very good as far as their variation about the 20 percent, the blue lines; they're well within that.
Next slide.
Again, the variation mimics that that we saw for morphine, where we're at somewhere around 10 percent or less. I guess it's about 12 percent out there on the end. That's at the higher concentrations.
Next slide.
The values that we see for codeine, the results are very similar to that that we see with morphine. We have the tight variance around the cutoff. As we get out in concentration, it becomes a higher variation in the values that we're seeing.
Next slide.
The variance, as we see here, the variance becomes about 12 percent when we go out in a high concentration of 23,000, which 10 times the cutoff concentration.
Next slide.
Now let's go to the first of our alternate matrices, hair. The initial test for the opiatemetabolites is 200 nanograms per mg. That's the cutoff for the initial test, which is an immunoassay. The confirmatory test has the same cutoff, 200 nanograms per mg for morphine and codeine and 200 for 6-acetyl morphine.
Next slide.
One of the things I'd like to point out on this slide is the hair preparation, and this is a possible source of some of the variation that we see, that we see in the matrices, is that out of the six labs we have four different methods of preparing the sample for the GC/MS test. In 3 of these we have initially the hair is washed with some type of solution or solutions to remove outside contamination.
Now, you say what type of contamination? Well, most people, if they have long hair, are using different types of conditioners on their hair. One of the interesting things is that when we're making our PT samples we kept seeing this white, fluffy layer that we were getting, emulsion that we were getting in the solution.
We found out -- we took some hair that someone had used conditioner in, someone what didn't, and looked at the hair, and we got the same thing. So the conditioners that are on the hair could potentially affect the recovery of the drugs in the hair. I'm not sure of the effect of this. I have yet to investigate this, but it appears to be a potential problem as to how much conditioners can affect hair if it's left on there, or even if it's not left on there, if it's taken off. How much -- is there drug that can be extracted, because most of these conditioners are an oil-like material. They help in the hydration of the hair and in making the hair soft. So that can cause problems.
But you see that we have three that are doing the prewash and one that was not. That's one variation.
Then we get down, you've got hair, it's a solid matrix, as we well know. So there's got to be some way to get the drug out of the hair. We have one -- two laboratories that are taking the hair and digesting it somehow, either enzymatically or chemically, to tear down the hair structure, the structure of the hair, so the drug can be released.
We have one lab that is powdering the hair after they've washed it. When we started this program we were trying to use powdered hair for the PT and there are several things that you know about powdering hair. It has to be ground in some manner in order to make the powder, which it's very difficult to recover everything that you have ground. So I think that could be another source of variation, causing the variation.
We have one lab that cuts the hair up into small pieces and then chemically extracts the drug out. It's been shown in other PT programs, I think Dr. Goldberger early on showed, that with morphine you saw a difference between labs that did a chemical extraction versus those that did a digestion. So now we have a third means of differences between the labs could be a source.
Then of course, the fourth, we had the lab that did the powder without a wash.
So I think that what we're seeing here is that somehow we need within the industry to come to standardization, which is in essence what has happened within the urine program, is that the procedures have been standardized such that they are not exactly the same, but the general methods are very similar.
For the samples, we had challenges for all analytes in all cycles, in all four cycles for hair.
Next slide.
This is what I'm talking about with the variation. We can see the 20 percent lines, the 50 percent variation lines from the mean, and you can see that the values from lab to lab vary. Wecan see that down in this area we have around the cutoff, though, that we're getting some pretty good -- I mean, it is possible in certain areas to get pretty good variation. It's much better than it is as we get further out, even 2 times. Whenever we get out here to 2 to 4 times, we start seeing an increase in variation.
Next slide.
It looks better when it's on the smaller scale, but you can see that they're all somewhere around in the 50 percent or less variation from the mean. But the values are spread out. Now, agreed we only have, we're looking at six values. So trying to compare it back to the urine, where we have 50 values squished into the same amount of area, is somewhat an optical illusion. It gives the wrong impression.
Next slide.
When we look at the variance, though, as you can see, the variance for the morphine in the hair in the first cycle was 25 to almost 50 percent. In the second cycle it went for a sample down near the cutoff, just below the cutoff, from about 15 percent, as we go on up 20, 60 plus percent.
In cycle four, where we're looking at washed and unwashed, we asked the laboratories, okay, how about washing the hair, but we want you to analyze part of the unwashed hair, and we were trying to see if that would help in reducing variation. There may have been some effect with the not-washed. It seemed to be a little bit more -- a little less variation. But still, it ranged from about 5 to 40 percent. Overall, the washed hair had a little bit higher variation if you look at the average. There were some samples which were equivalent to that of the not-washed.
Next slide.
Now, the reason for this, I think, the variation between the two, is that what we're seeing with the washing is the removal of some of the morphine from the hair PT sample. So again, here we have in the washed samples the variation in the amount that's removed by the different washes is going to affect the variation. That's just an explanation of why we don't see exactly the same level of variation in nonwashed and washed.
Next slide.
Go to 6-acetylmorphine. Not all the labs -- well, I think all the labs attempted to do 6-acetylmorphine. It was when we got down near the cutoff that we had some problems. We were not, in these particular samples, we were not below the cutoff of 200 picograms per mg.
But you can see we do have considerable variation, just as we saw with the morphine.
Next slide.
We can see as we blow it up that we only had like three values as we got closer to the cutoff. So there's some problems with sensitivity in some of the laboratories trying to reach these concentrations, and still considerable variation.
Next slide.
But we see the opposite thing with 6-acetylmorphine than we saw with morphine as far as washed or not washed. The not-washed, unwashed samples, had considerable more variation than we had with the washed. There are several explanations possible, but one of them might be that, since not all of the laboratories were doing the washed, then the ones that had the greater sensitivity were able to do the washed samples and so their values are a little more consistent, the people with the MS/MS for example. But that's yet to be seen. It's just a feeling that I have is the MS/MS may provide -- because of its increased sensitivity, may give us a little bit better reproducibility on the quantitation of these samples.
Next slide.
Again, on the washed samples, at least at lower concentrations, it was somewhat inconclusive, but it would appear that, like the others, we do have removal of 6-acetylmorphine from the sample with washing. I'd like to have more values so I could say things a little more definitely.
Next slide.
Codeine, which is of interest analytically, shows the same type of variation that we saw with other samples.
Next slide.
And we can see that the variation is somewhere in the 20 to -- we had some samples in the not-washed samples that were fairly high and we'll be talking about that. But anyway, it appears that with washing with codeine we can get a decrease in the variance that we see. We had a high variance in the not-washed samples.
Next slide.
Going to oral fluids, in the oral fluids, looking at the opiate metabolites, initial test cutoff at 40 for morphine and codeine, 40 nanograms per mL, 6-acetylmorphine, 4 nanograms per mL.
Next slide.
One of the things -- I just gave this piece of information. Of the labs that were working or participated, we were looking at the collection devices that we had and the information that we were able to get and we had three that were using OraSure collection devices, one AVITAR, one ANSYS, and two were not using -- had no collection device.
Now, in this program we provided each laboratory with two milliliters of oral fluid that had been spiked with drugs and we gave them that and allowed them to test that. We asked them to test it both as the neat fluid, that is just as oral fluid, and then we also in certain samples asked them to take and pipette onto their device the theoretical amount of fluid that that device would absorb. In some cases it's one mill, in some cases I think it was 0.4 mLs was the amount that was required. Then they were asked to analyze these as they normally would.
Next slide.
The information -- you'll see that we have oral fluid neat. That means without using the device. And the concentrations, as you can see, we did straddle the cutoff of 40 nanograms per mL. We had one sample out here at about ten times the concentration. We saw some variation, had one lab had a value outside of 50 percent.
Go to the next slide, which shows the values a little bit better as we look at them.
It blows up that area around the cutoff, and you can see that the variation here is not quite what we saw with the hair, but still there's a good bit of variation. When we look at the variance, which is in the next slide.
Go back to the previous one.
This is again for the neat samples. Then the next slide shows the few labs that were able to do both neat and with collection device, and you can see the variation. It doesn't seem to be that much difference with morphine.
Go to the next slide, which gives the variance that we saw. The variances are somewhere in the range of 10 to 40 percent. You can see with the neat versus device we got very similar results on those samples in cycle 3. When we go to the next slide --
I thought I had another slide in here which showed the recovery of the morphine versus,device versus neat. I must have deleted that one somehow.
Let's go to looking at 6-acetylmorphine. Again, we were able to straddle the cutoff. We have some at half the cutoff, we have others right around the cutoff. We have again considerable variation in samples all around, up to and above 50 percent of the 6-acetylmorphine cutoff.
Next slide.
The previous one was for the neat. With the device you can see it looks a little bit less, but we have very few values here, so it gets more difficult to look at this with few values.
Now, this is another problem that we're going to have in this program that we don't have with the urine at this time, is that with the smaller number of laboratories the statistics are going to be more difficult, because you're going to have higher variations whenever you have fewer labs, because n does play a part when we're looking at statistics.
But still, plus or minus 20 percent, plus or minus 50 percent, remains constant no matter where we are.
(Screen.)
You can see with 6-acetylmorphine, we did not have 6-acetylmorphine in the first cycle, but we did have it in cycle two and cycle three. You can see the variations for 6-acetylmorphine were in the 10 to 15 percent range as neat, but when we went to device we saw an increase in the variation. This may also have something to do with the number of samples that we had.
Next slide.
With codeine we had -- the variation appeared to be a little bit better in the codeine than it had been in the morphine, but I'm not sure that that is significant with the number of values that we have. I know initially, working with the industry, they really didn't -- one of the issues was, well, why do we care about codeine? And most of the devices were set up to detect morphine.
But it seems that the laboratories have done a fairly acceptable job with the codeine.
Next slide.
You can see that on this side of the cutoff the values look pretty good. You don't see any values that are outside the 50 percent.

We see the variance. Again, with codeine the variance is low, as we would expect. Just looking from the graphs, it's somewhere around, oh, 16, 17 percent or less. So we get, at least with codeine, in cycle three we were able to get into, start getting into the range of what you would consider maybe acceptable, because we use plus or minus 20 percent as a judge within urine as to whether a sample is acceptable. So it looks like a variation is getting into that range in the codeine and the 6-acetylmorphine in cycle three.
We would hope that as we go through something we would see an increase in the - or a decrease in the variance, that is a more uniform result, from the laboratories if we're starting to see standardization and labs coming together in their procedures.
Next slide.
I changed that slide. What we have here is the device versus the mean at concentrations, we have -- 6-acetylmorphine, as you can see, with oral fluids the device had essentially no effect. You can see with -- this is morphine and this is codeine. You can see that the two values, whether it was from the device or whether it was neat, the analysis gave essentially the same results.
So if there is an effect there's not much on these samples. Again, we had some limitedvalues, trying to get paired values from the laboratories.
Next slide.
Urine summary. The confirmatory procedures do provide sufficient sensitivity and precision for the present analytes. Now, as the labs move into new analytes they're going to go through the same type of issues in developing procedures that will be standardized to provide acceptable results. Before new analytes are put into play within the program, it's been our procedures in the past to PT these samples for quite a while and until we receive acceptable results. And by "PT" we mean we go through, we give them samples, they analyze them; we go back to the lab, we tell them what the problems are that we see, and then we provide -- they provide us input as to what they're doing to try to prevent that problem from occurring again.
So it's an education process. It's a process of bringing things together. Unfortunately, in a pilot program, in a non-regulated, you don't have that option. You are depending upon the laboratories and their goodwill to provide, to do the analysis and to provide you with the results. You don't go back and say: Well, you had a 50 percent error, you've got to correct this and tell us what you're going to do, and then PT them again.
Next slide.
Hair summary. I think I've covered most of this previously. These are just some summary slides. The wash procedure appeared to lower opiate concentrations in hair PT samples and appeared to decrease the variability for 6-acetylmorphine determinations as well as codeine. Efforts to demonstrate analytical sensitivity, precision, and accuracy and to evaluate the appropriateness of cutoffs have been hampered by the limited number and variability of the analytical values. This has been a problem all the way through.
Future PT cycles. We intend to provide -- we have them made now. We're ready to begin a new set and we want to look at inter, that means between the labs, and within-the-lab variation using these samples and looking at different points both below and above the cutoff. So that's our plans with the future PT cycles, for which we now have the samples made.
Next slide.
Oral fluids. Values obtained for neat and device treatments were equivalent for the opiates. Most laboratories seemed to demonstrate sensitivity below the suggested opiate cutoff. In cycle three the variability was usually less than 40 percent of the group mean for morphine, 30 percent for 6-AM, and 20 percent or less for codeine.
In future sets we plan to do the same thing as we do with hair. We plan to look at the inter-lab and intra-lab variation, as well as assess and investigate issues associated with collection device. Since at this point in time there's a requirement for labs to report their results based upon the neat oral fluid, that is how much -- the volume that was taken from the individual, the issue of the collection, of the current collection devices which collect a certain amount, how accurate are they in obtaining that, because otherwise, if we're looking at collection devices and trying to do quantitation in a PT program there has to be accuracy or else it's going to increase the variability that we'll see on the PT program.

Agenda Item: Analysis of Initial Test Results for Cannabinoids and Opiates

Note: The following presentation is found in a PowerPoint format under:

Workplace Drug Testing Documents and Publications

Regulations/Guidance Under Development

DR. MITCHELL: If you go to the SAMHSA website, you'll find that we have the results from each of the 4 cycles that we sent out. Now, one of the things that we provide on that for initial test is we say, well, how many were positive and how many were negative for a particular sample or a particular analyte.

Well, what we are attempting to do here is go beyond that and look at the initial tests that were used in the oral fluids and in the hair. We start off, as an example, with some of the results that we've gotten from urine.
Next slide.
This is something that's totally new. We have never done this before at DTAB. So what we're going to do is go into some very simple statistical type analysis, Bayesian analysis. It's an analysis that allows you to look at the predictive value or the ability of a test or a procedure to predict something. If you go out on the website and put "Bayesian" in, you'll see that there is actually an International Society for Bayesian Analysis, and it's used in medicine, it's used in science, it's used in different places.
In the way that we are looking at immunoassays, immunoassay is the first test that's conducted and it's going to give you a positive result or a negative result. So Bayesian analysis uses that result and what is known about the sample in order to make some predictions or look at some properties of the way that a immunoassay behaves.
It utilizes the number of true positives, which these samples -- we'll go through the definition of these -- true negatives, false positives, and false negatives to determine four parameters. The four parameters we're going to be looking at are sensitivity, specificity, positive predictive value, and negative predictive value.
Next slide.
A true positive is nothing more than a sample which contains an analyte, a drug for example, cocaine for example, at or above a predetermined cutoff, a predetermined value. In this case we're using the cutoffs that we had or we published for the different analytes or different drug classes that we're going to look at today, being cannabinoids and opiates.
A true negative is a sample that contains drug analytes at concentrations less than that defined cutoff. It can actually have the analyte or it might not have the analyte, but it's going to be less than the defined cutoff.
A false positive, again, is a sample that has analyte below the cutoff, the defined cutoff, but which is found to be positive by the particular test that's being used.
A false negative is a sample that has analyte above the cutoff, but is found to be negative.
Next slide.
First we're going to look at sensitivity. It's a very simple equation, where sensitivity is equal to the true positives that were identified as positive, divided by the true positives that were identified, plus the false negatives that occur. Actually, that bottom, the denominator, is nothing more than the total number of true positives that were included in the test set.
Of course, all it does is reflects the portion of the true positive samples that were identified by the test. So if you have a value of 1 it means that it identified all of the true positives, that's all it means.
Now, 1 minus sensitivity, just as an aside, will give you the portion of them that were falsenegatives.
Next slide.
Specificity. The same way, it's the number of true negatives that were identified divided by the total true negatives, and that's composed of the true negatives that were identified as well as the false positives. False positives are really true negatives. And the specificity reflects the portion of the true negative samples that were identified as negative, and 1 minus the specificity gives you the portion that were false positives.
Next slide.
Now, positive predictive value. Now we're getting into colors, slight differences in grey. It's the true positives that were identified divided by, in essence, all of the positives that were identified, which is composed of true positives and false positives. Again, one minus PPV would reflect the number of false positives.
Next slide.
NPV is similar. It's the true negatives divided by the total negatives that were identified by the assay, which is composed of true negatives and false negatives.
Next slide.
Let's look at some data from urine for an example. This happens to be an EMIT result for THCA in urine, PT Occasion 61 and 62. You can see the data gives us true positives being 153, true negatives 570, and we had 3 false positives by the immunoassay in that set. This was -- we found that the cause for this was that that sample contained glutaraldehyde and it evidently --we've talked about this before, how different adulterants can affect different assays different ways, and glutaraldehyde had a tendency in some systems to cause false positives for THC. We had no false negatives.
Now, with the sensitivity -- remember, sensitivity reflects how many or the portion of the true positives that were identified. So we identified all of the true positives. The specificity, since we have a false positive, is going to be less than 1.
The positive predictive value again is going to be less than 1 because we had the false positive. And the negative predictive value is 1 because we identified all the negative -- the negative predictive value is the true negatives divided by the total negatives.
Next slide.
This is another result in which we had a false positive, and you can see similar results. This is just -- we only have I believe one or two labs that are using FPIA at this time. True positives are a small number, 6; true negatives, 19; false positives, 1; false negatives, zero. The sensitivity again is 1. Both specificity and PPV are affected by that false positive and the negative predictive value stays at 1.
Next slide.
This is the results for cannabinoids in urine, PT Occasions 61 and 62. As you can see, the different assays that we have -- we have CEDIA, DRI, EMIT. Let me say, with EMIT we've taken all the different EMIT and Ciba and Behringer and put them all into one, which is probably not fair, but we did it that way anyway. We have FPIA and the KIMS reagent from Roche.
We looked at these as far as true positives, true negatives, false positives, false negatives, and so forth. You can see that we only had the one lab that we previously -- I mean, the one instance where we had the false positives.
Next slide.
You can see again the predictive values, the specificity and predictive values, are less than 1 because of those false positives in the FPIA and the EMIT.
Next slide.
The same thing for opiates. Again we have the FPIA problem with the one false positive. All the others, we had no false positives or false negatives. Therefore in them we would expect the values to all be 1.
Next slide.
Again sensitivity and PPV are affected only in the FPIA. All others are 1.
Now, this is a simple example. Let's take on something a little bit more complex.

MR. STEPHENSON: Before you get any further, I just want to say that, for those that are looking at this as interested parties, but not necessarily coming at it from the training or experience as a forensic toxicologist, a false positive as defined in this environment coming out of actual operational data is a screening assay. It's a presumptive test that in the normal course of events would continue on through GC/MS. It would not sustain itself as a false positive in the report. I need to say that because I don't want to have a flag go up saying we're identifying false positives in the system that involves the current testing technology.

DR. MITCHELL: That's correct and I should have made that clear. We are only considering results based upon the immunoassay that we're looking at. Of course the specimens were confirmed either true positives or true negatives in the total sense by a confirmation, and that's how we found our true positives and true negatives.

Next slide.
Okay. We're going to go to cannabinoids in hair and opiates in hair, and we're actually looking at an initial test which has one picogram per mg as the cutoff. But we also in this case, we need to consider the confirmatory cutoff, which is 0.05. When we're looking from an operational standpoint, we want to -- one of the things that you would like to measure is how well does an initial test predict the success that you're going to have with the confirmatory test.
The success of the confirmatory test is going to be based upon the confirmation cutoff, even though the initial test cutoff is at 1 picogram per milligram. Now, with THC we have this in urine, we have the same problem, is that the cutoff for the initial test is at 50, whereas the cutoff for the confirmatory test is at 15. The reason for this is that with the cannabinoids there are various products which are produced when it's taken into the body. So these immunoassays, they have reactivity to a basic shape and composition of a molecule, and so it will have cross-reactivity with many of these different compounds that are produced in the body as a result of taking in THC.
So what we're looking at with immunoassay is, what we used to say, is a combination of the compounds or analytes that are produced as a result of the ingestion of THC into the body. The same thing is going to happen in hair, because we should see compounds which result from the ingestion of THC into the body.
The primary one we're looking at, which is the one that the confirmatory test is established upon, is the THCA, which is the 9-carboxylic acid metabolite of THC.
In opiates, the initial confirmatory tests are both 200, so we don't have to worry about some of these things. But there are other issues with opiates that we'll talk about as we gothrough them.
Now that everybody is confused, let's go forward and see if we can't increase the confusion.
Next slide.
With cannabinoids, when we go through the calculations like we did, like we showed you, with THC, we find that the sensitivity based upon THC is all below 0.67, 0.67 or less. Now, what that means is that the true positives, that is the sample that had THCA at a concentration greater than or equal to the cutoff of 0.05 picograms per mg, only 67 percent of those were identified by this immunoassay, or 62 percent were identified, and so forth. That's based upon THCA, the metabolite that we looked at.
We had some other data, just to confuse things further. Trying to look and determine what the specificity, if there was some indication of what the specificity was, we saw that if we took one picogram per milligram of THC content in that hair and we looked at that and did this analysis, then it would appear that this and the ELISA and the RIO would identify 80 percent, 80-plus percent, or 0.8 of the samples that contained THC. So it indicates -- it doesn't indicate. It is possible that it might be that these particular assays are a little more attuned to THC than they are to the THCA concentration, or a little better at predicting THC concentration.
Now, when we look at the cannabinoids we consider both at a concentration equal to 0.05 and we see that the predictive value is not very good. But in that one case I did it just a little bit different so that we might be able to get some information.
So the sensitivity says that we're doing less than or we're able to identify less than 67 percent of the positives with the assays that we're looking at here. Now, the specificity tells us what? Anybody remember?

VOICE: Is it better at identifying negatives?

DR. MITCHELL: Yes, are we able to identify the negatives. You can see that we're much better at identifying true negatives, because, using THC, we're able to identify 77 plus percent of the negatives, of the true negatives, by this assay. So it's better at identifying the negatives.

Next slide.
Again, we can see the positive predictive value. We would expect it to be low because of the true positives, but it also indicates that there are, what, a lot of false negatives, is that correct? Because in the equation we have true positives divided by true positives plus false negatives. So we can say that for the predictive value, the ability to predict the true positives, it ranges from 48 percent up to 83 percent for these assays.
In the negative predictive value, again how well at predicting negative values, again it does very well. That's reflected from the sensitivity.
Next slide.
Going to opiates, you can see that with the opiates it appears that we do a little bit better job with the sensitivity. We had one assay which had fairly poor sensitivity no matter which analyte you looked at at a cutoff of 200. The others appeared to be about 80 percent or greater, depending upon the analyte.
We liked to use -- whenever, a long time ago when we were looking at the efficiency of assays, we thought that if we got 80 percent or better that was a fairly good, from a productionstandpoint that was fairly good. You'd like to have it 100 percent, but sometimes you'll lose things if you go to 100 percent. So that 100 percent is not always the best.
The specificity, again it appears to be very good at identifying true negatives. You can see that it looks like most of it is in the 90 percent or better range.

Next slide.

Now, the question -- you'll see, just looking at this in general, it appears the positive predictive value is pretty good. However, the ELISA-1, if I remember right, they were -- go back one slide.
The ELISA-1, you see that they had a pretty good record as far as the true positives divided by the true positives plus the false negatives, which means that they had very few false negatives.
Next slide.
So why do we end up with a positive predictive value that's less? It's got to be false positives. So even though they're identifying true positives and they have few false negatives, they are having false positives with that particular assay. That means from an operations standpoint that's more samples going into the GC/MS, which for a laboratory is time-consuming and money-consuming.
But it doesn't mean that they're going to produce false negatives in their final results. It just means the number of samples that's going to be going into the confirmation process that will not produce a positive when you go to confirmation.
But overall, it would appear that we have, based upon morphine, all the samples are fairly good in predicting that morphine is there at or above the concentration, and the NPV again are pretty good for these assays.
Next slide.

DR. KWONG (DTAB member): I think when we calculate predictive value it's very important to take into account the prevalence. The exercise you've gone through, which is very, very interesting and good, illustrates that there's more to sensitivity and specificity, okay. But the analysis you did was based on the PT samples, which means the prevalence is very, very high.

DR. MITCHELL: That's correct.

DR. KWONG: If you extend that to your real life situation where your prevalence, the positive rate, is 1 in 100, 1 in 1,000, 1 in 10,000, 1 in 100,000, then your positive predictive value, even with the specificity of 99 percent, will be down to about 10 percent.

DR. MITCHELL: That's correct.

DR. KWONG: I just want to emphasize that since you are dealing with very high prevalence here, the predictive error actually looked pretty good. But when you use the same analysis on a real life situation, even a 99 percent specificity will give you a very poor positive predictive value.

DR. MITCHELL: That's correct, and that was one of the points I was going to make at the end, is that, okay, now that I've said all this, this has no relationship to the real world. We're using thisas a means of doing an exercise in trying to understand the assays that are being used, and that's under artificial conditions, which is the performance testing.

But it does not change some of the issues that are associated with false positives and false negatives. One of the things that we have here, we don't have it so much with hair, but we would have it with urine and we would have it with oral fluid, is that because of the way we collect those particular specimens the people are not using other drugs. So what we've done is we've removed from those matrices the interferences that were caused by over-the-counter preparations and their interference with immunoassays causing -- sometimes they react, sometimes they cause other types of problems with the immunoassay.
So this is a very artificial environment that we're looking at, but I think from a point of looking at immunoassays, whether or not a cutoff is close to where it should be, whether or not it's going to, if presented with samples of this type, if it's going to give us an acceptable result, I think that it does have some merit. It's not the whole story, please. I don't mean to give that impression, that this is the whole story on immunoassays.
With oral fluids, we'll look at THCA and opiates and the initial test for THC has a cutoff of 4 nanograms per mL, confirmation 2 nanograms per mL; for opiates, the initial and confirmatory tests are the same, 40 nanograms per mL.
Next slide.
When we look at THC at the cutoff, we find that at least one of the assays was pretty good at identifying true positives -- well, two of them actually pretty good, 0.7 and 1.0, and two were fairly poor. Now, the reason for that may well be that the cutoff on those particular assays is high, is much higher than the intended cutoff in reality.
The specificity, as we see, they were able to identify the true negatives. The positive predictive values were pretty good, meaning that they were not producing false positives, and the negative predictive value for THC was very good.
So with the immunoassays for oral fluids, again we're looking for the parent compound, the THC in the oral fluid.
Next slide.
Analysis of the initial test for opiates for oral fluids. I tried to look at codeine, morphine, and the total opiates because the 6-AM was a tenth of the concentration, but morphine's probably going to -- well, it does overpower anything that you would see from the 6-AM. You can see that all but one of the assays able to identify all the true positives at the concentration of the cutoff that we're looking at. Codeine became less important and when you consider the combined it didn't do anything to increase the true positives that were being identified.
Specificity, we were dealing in the 80-plus percent as the ability to identify the true negatives.
Next slide.
With the positive predictive value for ELISA-1, again it went down, and we said the reason for that is the false positives that we would see in that particular assay. The positive predictive value of all of these was fairly low using these particular samples that we were looking at. The negative predictive value appeared to be very good with these PT samples.
I believe that's the last slide.
(Screen.)
Oh. That's okay. That was just in case questions came up about the morphine in oralfluids, if we actually had some of the data available.
In summary, this is not the total answer. It appears that to really test the sample the way we would like to it would probably be best to have only one analyte in a sample at a time where we're dealing with mixtures, and so that further complicates this type of analysis.
With the limited number of samples that we've had or have been able to formulate at this time, we've tried to make the most of the confirmation tests. That's what these were really designed for, was looking at the confirmation. So the analysis based upon the initial test is again not a classical analysis that you would do trying to look at the sensitivity, specificity, and some of the other values that are brought forth by this.
Again, as Dr. Kwong said, when you go to the real world where you're dealing with positive rates of 1 percent or so, these values are going to be reflected differently.

DR. BUSH: While we qualify this in the context of the application of the statistics to these values, you all need to understand that we're sitting in a position here where we have to determine and evaluate cutoffs. We have to figure out what the best and most appropriate cutoff is that's going to hopefully preclude the identification of false positives all the way through the system, preclude false positives from being reported out even through the confirmation procedure.

But we also want to make sure that with immunoassay and any type of screening technique that is used that we appropriately identify those that need to go on for confirmation, because what could possibly happen in a worse case scenario, we could have the most perfect confirmation test linked with a screening test that doesn't identify an appropriate number of specimens that contain drug.
You can have this perfect confirmation test at the back end of something that's not identifying anything to gate into the confirmation procedure. But on the other hand, we don't want a sieve on the front end where essentially everything moves to confirmation. That's not the point either.
We are wrestling with the different types of technologies, with their individual sensitivities and specificities, which of course varies analyte by analyte across drug classes. And of course, we're looking at five of those at least and maybe more for the future. So we're trying to evaluate the technologies, how to marry screening and confirmation best, and any and all suggestions on how to get there from here would be greatly appreciated.

MR. IRVING: I want to go back in a little history. When we established the cutoffs for drug testing for THC, we didn't have the capability of creating PT samples that reflected real life. Basically, we spiked with THCA, and in order to evaluate the laboratories at the lower cutoffs we sent directed samples. It was not until we developed PT samples that mimicked real life, where you could spike it in and have a low level of THC, that the PT program really reflected real life.

I think in the case of the opiates I don't see a problem with this. But in the case of THC we don't know what a real life sample looks like. I know for hair we don't as far as what is in there. I don't think the case for opiates for saliva is as bad, but it's almost like we have to back in with real samples and create the PT program based on what the real life samples are. The relationship between the screening cutoff and the confirmation cutoff for THC was really experienced from real life rather than scientific direct research on it. We sort of backed into it.
It's almost like we're going to have to do the same thing here with the alternate matricesfor complex samples such as THC.

DR. BUSH: I think you're right on with that when in fact the industry working group for hair testing suggested that the immunoassay or screening cutoff was 1 picogram per milligram, yet in fact when you drill down and look at THCA in the hair specimen looking at a confirm process, confirmation process, it's 0.05. So you see a factor of 20 difference between the initial screen and then taking it on to confirmation, allowing for just a large number of possible cross-reacting analytes and what might be in there.

The point is well taken.

MR. STEPHENSON: I'm sure a part of that is the lit fuse of being overwhelmed a little bit perhaps, or pondering questions. I think it was a great presentation, both of them. Thank you very much. They are helpful in looking at the exploration of the utility of a screening assay under development, and the purpose is not to score them. The purpose is to help the industry themselves identify those assays and techniques which can reduce ultimately total cost by having a higher predictive positive value for sending something to confirmation, which, as you said, in a lab is both a time and a money consuming process. That ultimately determines the success of any deployment of technology out there.

It can vary from drug class to drug class, as we've seen. It can include elements that we truly understand now and some of the analytes that may be present that we have not directly identified. One of the examples in the oral fluid area is the issue of total opiates. You get great cross-reactivity and in certain areas that may be a very good thing to have, but not if you're looking for a very specific single.
But in terms of all of these, I will tell you that under the domain of a future contract which is now in the process of being developed we will have adequate -- we have defined tasks and cost categories that will allow adequate funds and staffing concentrations to be applied to some of these areas.
How quickly added money and staff can in fact resolve these things is still dependent upon the goodwill and the participation of the scientific community and the industries themselves. We're not going to be able to drive this from a federal resource perspective alone, but I think we'll be able to concentrate a lot more capacity to doing some of these things.
That's going to cut across the board. As we look at these issues, we're going to have to expand to look at the opportunities for those same adulterants that we have identified in urine, that have consumed a great deal of our time and effort in the past, and how those may in fact also impact alternative specimens. And we still need to monitor that whole process out there because it's still a part of what we are in this country.

MR. STEPHENSON: All right. If we don't have any other issues -- are there any other issues of importance to the Board members that they would like to address at this time, that we haven't discussed in these issues? I want to make that as an opportunity. (No response.)

Agenda Item: Public Comments

MS. VARLOTTA (Flight Attendant): Mr. Stephenson, members of the Drug Testing AdvisoryBoard. Allow me to introduce myself again. My name is N.B. Varlotta. I am a career flight attendant with an excellent employment history. In March of 1999, my employer terminated me one month shy of my providing 20 years of faithful service. I was falsely accused of substituting a drug screen sample under the standards established by this Board.

After much research, I have found that there is no avenue in place for an employee without representation or for a non-certificate holder, a flight attendant such as myself, to clear their name in a similar circumstance.
I have found thus far corporations and laboratories take the positions that rights, meaning the right to challenge laboratory results and a termination in court, are preempted by federal law. Unfortunately, against the intent of Congress the courts have sustained this position to date. This means that no employee has the right to contest results or a termination in court. The employee has to go to the Secretary of Transportation for relief. However, the Secretary of Transportation has no power to reinstate a wrongfully fired employee.
I have also discovered laboratories are claiming findings can never be challenged in court. Corporations argue that laboratories comply with the FAA procedures and, if challenged, the FAA will conduct an investigation which the supports the corporations.
Accordingly, any claim based on the test results is preempted. Again, the sole remedy is to go to an agency having no power or desire to provide a remedy. Even in the instance when indisputable evidence is provided, is presented that a laboratory did not comply with FAA procedures, the corporation argues that the wrongfully fired employee has no right to sue the laboratory or the corporations because this is preempted by federal law as well.
Corporations also argue that an employee accused of substituting a drug screen sample, low in creatinine as in validity testing, has no right to present to the employer contrary evidence. This means that a false or even fraudulent accusation a corporation to fire an employee without any right of appeal, with no right to challenge the validity test in court, and of course no right to challenge the decision of the corporation to fire the employee.
Additional research unveils that corporations also contend a fired employee has no right to present evidence of their innocence, to have the corporation conduct an investigation, to have evidence considered, or to have a fair hearing before being fired.
Presently corporations argue that the medical review officer has no obligation to investigate an alternative explanation for substitution, for example low in creatinine in validity testing, nor any obligation to gather information from the accused employee.
Although there have been some discussions and alerts sent out to the medical review officers, there is no protection or avenue of procedure in place to clear an employee of a false accusation short of the medical review officer simply choosing on their own personal accord or whim to cancel a test.
I am again asking this Board to please discontinue the so-called validity testing immediately and until the science is foolproof. Failing such action, since careers continue to be threatened and since the science continues to be questionable and since human error can still never be removed from the equation, I would continue to suggest that it is incumbent on this Board to adopt specific step by step procedures acceptable to all employers which would allow any employee to clear their name with a remedy in the event of a false accusation.
Mr. Stephenson, all members of the Drug Testing Advisory Board, thank you for your time.

MS. SHANNON (AFA): Good morning, ladies and gentlemen of the Drug Testing Advisory Board, my courageous flying partners, and guests.

My name is Lucy Shannon. I am a member of the Association of Flight Attendants at United Airlines. I am responsible for the delivery of peer-based EAP services at AFA United councils worldwide. I am here today to provide you with a summary of the Association of Flight Attendants' history with the concerns about substitution testing. I believe that I am well qualified to speak about the members' issues around this developing technology as I was to represent the first flight attendant terminated from United Airlines for substitution testing a little over three years ago in my capacity then as local president.
In my current position as Master Executive Council EAP Chair, I educate and support flight attendants, their family members, their doctors, specialists, union presidents, and grievance representatives as they attempt to understand and defend against the charge of cheating on a drug test.
In the last 90 days, I have had three such cases. All three flight attendants' lives have been turned upside down for no other fault than they can provide, naturally provide urine underneath the DOT levels of substitution. All of the cases have been exonerated as cheaters. They are now back to work, but not without prejudice and not without significant and quantifiable economic loss.
AFA's direct experience with substitution testing began about three years ago when a United flight attendant, Tana Laura, tested just above the 5.0 creatinine and 1.001 specific gravity. Mistakenly, the lab rounded her creatinine results down to 5, which resulted in her termination from United Airlines.
Following that experience, Tana participated in the DOT water load study. In that study Tana is the donor who is listed as testing closest to the substitution threshold at 5.2 and 1.001. For AFA, when we look at Tana's substituted results at United and her results in the DOT water load study, we knew that it was time to fasten our seatbelts.
Twice, Tana naturally produced urine that hovered around the "you are a cheater and you will be fired" threshold. She produced the second result by drinking less than two cups of water an hour. For flight attendants who must get acclimated to an atmosphere with little or no moisture, this level of water consumption should not be characterized as water loading.
As a special note, at the direction of the medical community, you'd be hard-pressed to find a flight attendant who was not either having water with them or consuming quantities of water while working.
Tana's results triggered concerns by AFA around lab variation. What would happen to someone who naturally hovered around the threshold of substitution when lab variation is factored into the formula? We searched for Tana's split results of the DOT water load study, but have yet failed to see these results.
A second flight attendant, Michelle Nelson, validated for AFA that one can naturally test below the substitution thresholds, as Tana's test results strongly suggested. Michelle produced a second substituted sample under direct observation during a follow-up DOT test. Michelle's first substituted sample was cancelled due to problems with lab quality assurance. For AFA it was a sentinel event and thought that it should be for DOT and HHS. Unfortunately, the only response that AFA received after sharing this information is that the flight attendant could have had a catheter up her vagina.
Nest, the Association of Flight Attendants presented to DOT and HHS the paired creatinine and specific gravity test results of Alaska flight attendants whose urine was collected and processed by Quest Lab following complaints of air quality issues. Two of the samples were substituted. Again, this is another sentinel event for AFA under substitution testing. Flight attendants who had no reason or motive for doctoring their urine still had substituted results.
Most of you are aware of the experiences of a flight attendant Julia Jones. Her results are not inconsistent with what we saw with Tana and Michelle and the Alaska flight attendants. Her results also suggest that a variation between labs does exist and it exists at a level that does impact accuracy and reliability.
A newly substituted flight attendant, Joan Goodwin, shared her successful testing replication results during the recent meeting of the stakeholders in Tampa. Not included in the formal presentation were the results of a variation testing conducted by the Association of Flight Attendants. The same lab at the same location tested splits of the same sample of Joan's urine one month apart, with significant outcomes. A 5.9 and 1.001 was reported as 4.1 and 1.002, and a 6.4 and 1.001 was reported as 6.0, 1.001.
Today you are being presented with information of yet another recently substituted United flight attendant, Anne-Britt Svellingen, who has traveled across the country at her personal expense and time to share with you her experiences of substitution testing. Anne, like Tana, like Michelle, like the Alaska flight attendants, like Julia and like Joan, have all tested substituted without cheating. That's seven flight attendants.
Today I offer the Drug Testing Advisory Board a challenge, a challenge which requires each and every member to set aside its self-interest as beneficiaries of the drug testing industry. I challenge you to participate in the creation of a rigorously honest report to the Appropriations Committee on the accuracy and reliability of substitution testing.
I thank you for your time and for your consideration.

MS. SVELLINGEN (Flight Attendant): Good morning, members of the Drug Testing Board, and thank you for letting me address my experience that occurred in drug testing procedure. My name is Anne-Britt Svellingen and I'm a flight attendant that on December the 18th, 2002, after a flight gave a random drug test.

This is my second public speech. My first one I had when I was seven years old on Father's Day. Too bad this speech is about such a humiliating subject as my urine, which has made me feel like an alien for two months.
The nightmare started in December the 24th, 2002, when the MRO told me my urine test had come back inconclusive; it was not consistent with normal human urine. At first I could not comprehend what I just had been told. The MRO explained that the specific gravity and the creatinine, two words I had never known the meaning of before, was below the number the Department of Transportation has set forth as criteria for normal human urine. Therefore my urine was considered substituted, and that again meant refusal to be tested for drug and my drug test was now considered positive. Although I had never been tested for drugs, I failed to validity testing.
I was in shock. I knew the urine I gave to the nurse at the testing site was my urine, so how could this be possible? I was still not clear about what he meant. I do not use drugs, so how could my urine be considered positive when they hadn't even tested it for drugs?
The MRO didn't give me any guidance in regard to what the next step to take other than I had to contact a physician and he would have to provide them with a medical reason for why my urine came off like this before I could come back to work. Now, I felt I was scared. I was confused about the whole thing.
During the first week after this happened, I got misleading information. If I had gotten the right information in the beginning, I could have saved a lot of money. I started running around to doctors, giving different urine tests under observation, which was very humiliating. I gave blood tests, several blood tests. I had my hair tested for drugs; calling different people.
I was denied unemployment because the information they had gotten was that I had tested positive for drugs, so therefore I could not -- I didn't qualify for unemployment benefits. So I had to write a letter to appeal their decision. I had bills to pay, but I was not paid for two months. I had my vacation ruined.
I had to lie for my family, my mom that is 70 years old and wouldn't understand anything that I was talking about. I'm European and she don't understand the system here. I couldn't talk about this thing to her, so I just said I was furloughed and I was looking for a job, so that's why I couldn't come home for vacation.
I live a healthy life. I eat the right food, I drink water, green tea. I have no reason to go out and use drugs. I don't do drugs. I've never done drugs. I have done everything in my power to show my innocence in this case. I am still very confused about the whole thing. I am worried about my record. I want my record cleared because I am innocent. I haven't done anything wrong.
This has given me a lot of stress for two months, many nights not sleeping, stomach ache. I felt humiliated. I felt I couldn't talk with everybody about this, but it's just a few friends that I dare to talk to about this, because I felt it was not too many that would understand this. I didn't understand it in the beginning. My physician was confused. So it took a long time before I found somebody who actually did understand.
So I hope that this will never happen to another like me. Thank you.

MS. KLOTH (AFA): Good morning. My name is Elaine Kloth. I'm with the Association of Flight Attendants and I am a United Airline flight attendant.

I recently attended the colloquium on specimen validity testing in Tampa, Florida, and deemed it important to share my thoughts and reactions on that process. You may or may not have been aware of it, but AFA coincidentally submitted the language for a 250,000 dollar study on the accuracy and reliability of validity testing for inclusion with the FAA funding authorization.
Attending with me at the colloquium were flight attendants Julia Jones and Joan Goodwin. Joan's ability to replicate substituted results in a clinical setting parallels that of Julia. Ironically, even though flight attendants have been a significantly impacted work group, we have never been invited to this closed meeting. Julia and Joan's attendance was afforded after initiating many last-minute conversations with DOT and we do appreciate Mr. Edgell's support in allowing us to be able to be there and share our field expertise.
At the conference I heard that Julia and Joan were principally the only donors with claims of falling below the levels. What happened to AFA's submitted documentation about other flight attendants? What happened to the MRO reports of two additional United flight attendants who tested substituted and later had the tests cancelled with the failure of bottle B to reconfirm? Surely those cancelled results add to the evidence of the risk of individuals to fall below the standards and challenges the variation of labs.
I was disturbed to hear that the provision of MRO review was considered a safeguard to donor due process. It can only be a safeguard if the MRO knows what they are looking for. MRO's have rejected all submitted nephrologist letters of interpretation that an evaluated flight attendant could produce the substituted results just three weeks ago. It can be a safeguard if the substitution test that is under review is reliable and accurate to begin with. It can only be a safeguard if the process for obtaining a valid medical explanation is available, accessible, and timely for all.
The two most recently exonerated flight attendants needed an average of 54 days to prove they did nothing wrong. This is called fast-tracking when we look at the 18 months needed for Julia Jones.
Never did I hear a discussion about donor costs, though government and employer costs have always been considered.
While pulling the mandatory substitution language is an important step, giving employers authorization to conduct substitution testing without the rapid exchange of substitution findings creates more problems.
In summary, we need the three C's: competence of the MRO's, costs considered for donors, and communication about evolving findings back to the field.
Thank you.

DR. LAPPE: My name is Murray Lappe. I'd like to make a comment about the urine point of collection testing progress of the DTAB. Is it appropriate to bring that up at this time?

MR. STEPHENSON: It's a public comment period.

MR. LAPPE: Mr. Stephenson, members of the DTAB: In James Greenwood's recent letter to Secretary Tommy Thompson of the Department of Health and Human Services, Mr. Greenwood, Chairman of the Subcommittee on Oversight and Investigations, calls on the DHHS to assume control of alternative specimen test policy. The DTAB charter clearly states that SAMHSA seeks to improve the quality of services for forensic workplace urine drug testing, to assess the science and technology used in urine drug analysis, and to guide national policy in these areas.

Today I've heard no new progress on urine point of collection testing. For six years since April 1997, DTAB has been evaluating alternative specimens and technologies and it appears that we are still several years away from implementation. The DTAB may not be adequately staffed or funded to accomplish the depth of work required to investigate, evaluate, and advise HHS on the myriad of alternative technologies for drug testing.
The time has come to face the current realities of forensic science in the best interests of the vast community of employers and employees subject to federally mandated drug testing programs. Alternative specimens, such as hair, sweat, and oral fluids, must be disassociated and unbundled from alternative urine technologies, such as urine point of collection testing. Unlike oral fluids, sweat and hair testing, urine point of collection testing has an established precedent in the existing federally mandated drug testing procedures.
First, current drugs of abuse tests are limited to urine-based immunoassay screens and lab-based GC/MS confirmations. Urine point of collection tests would not deviate from these basic methodologies.
Second, point of collection tests have been successfully administered for ten years under federal mandates for alcohol testing.
There doesn't seem to be any significant scientific, analytical or legal barriers to urine point of collection testing. Yet, for six years this proven method of urine drug testing has been delayed by the policy of all or none of alternative matrices of the Board. More than 8 million urine point of collection tests were performed last year with no significant legal challenges that I know of.
The DTAB could develop a high-quality urine point of collection procedure in months, not years, and I strongly urge the DTAB to unbundle urine point of collection testing from the alternative matrices and to assess the science and technology used in urine point of collection testing and to establish a working group to draft guidelines for urine point of collection testing, to improve the quality of services for forensic workplace urine drug testing, and to guide national policy in these areas.
Thank you very much.

MR. STEPHENSON: Are there any other comments? (No response.)

At this time, I'd like to bring to a close the open session of the Drug Testing Advisory Board.

The open session was adjourned at 11:45 a.m..