Jump to main content.

Procedures for Detection and Quantitation

Technical Workgroup

Federal Advisory Committee on Detection and Quantitation Approaches and Uses in Clean Water Act (CWA) Programs

Summary of Technical Work Group Conference Call #18
February 15, 2006
1:00 – 3:00 p.m. EDT

The next Technical Work Group call is scheduled for Wednesday, March 1, from 1:00 - 3:00 PM EST. The call-in number is 866-299-3188; access code is 2025661045#.

Future Meeting Dates

Technical Work Group calls are scheduled as follows:

Subgroup meetings are set as follows:

Action Items

The Technical Work Group agreed to the following action items during the conference call:

The Regression-based Subgroup (formerly known as the Calibration Design Study Group, Multi/Inter-Lab Subgroup) has a call scheduled on Tuesday, Feb. 21 at 1:00 PM EST to

  1. complete its design,
  2. identify which questions the design would answer related to procedures and uses, and
  3. identify questions the FACDQ needs to answer relative to uses, MQOs/DQOs, and what procedures should do to allow final design and implementation of the pilot.

The Single-Lab Subgroup has a call on Friday, Feb. 24 at 1:15 EST. It is to focus on a straw proposal for a single-lab pilot test using existing data. The group will identify additional datasets needed to fill holes and indicate what the datasets should look like. It will also identify questions the FACDQ needs to answer relative to uses, MQOs/DQOs, and what procedures should do to allow final design and implementation of the pilot.

Combined Data-Gathering Pilot Test Design Subgroup: Richard Burrows, Larry LaFleur, Jim Pletl and Dick Reding will participate in a call on Tuesday afternoon, Feb. 21, at 3:30 PM EST to identify the goals and objectives for a combined (single and regression-based) pilot study based on existing data, to develop a straw study design, and to identify timing.

Ken Osborn, as the sponsor of the LAB QC detection procedure, will send updates to the procedure to Triangle which will distribute them to the Technical Work Group.

Top of Page

Welcome and Introductions

At 1:00 PM Bob Wheeler, facilitator, welcomed participants to the call, conducted a roll call of Technical Work Group members and observers, and briefly reviewed the agenda for the meeting.

Discussion of Draft Meeting Summary

Mr. Wheeler asked for comments on the draft February 1, 2006 Technical Work Group meeting summary. There were no comments and the summary was approved as drafted.

Top of Page

Review of FACDQ Assignments and Technical Work Group Products - Status Report

Mr. Wheeler noted that the Technical Work Group had about a month to finalize its work on the following five assignments from the FACDQ December 2005 meeting:

  1. Further develop the pilot study for multi/interlaboratory process
  2. Further develop the design for a single lab pilot, using existing procedures and existing data
  3. Offer recommendations for removing procedures
  4. Present modifications to procedures that were recommended by the sponsors
  5. Revise the glossary.

He then listed the ten products the Technical Work Group was working on:

  1. Regression-based pilot study design and report on existing data
  2. Single lab pilot design, using existing data
  3. Single laboratory study design, using existing procedures
  4. Analytical methods recommended to be tested in a pilot study
  5. Normalized matrix
  6. Short document on the pros and cons of LD
  7. Definitions of LC for censored methods
  8. Glossary
  9. Narrowing of procedures
  10. Procedure modifications - such as for Ken Osborne's LAB QC procedure

Mr. Wheeler mentioned there was also the possibility of combining #1 and #3 into a data-gathering study using existing data.

While discussion showed there was general agreement that the list of ten products was about right, the group agreed that it should give priority to its assigned tasks.

Top of Page

Regression-based Subgroup Report

Larry LaFleur reported for the subgroup. He said the subgroup had developed a "proposed regression-based plan." The group had talked a lot about procedures and approaches to confirm that a given procedure achieved the detection and quantitation limit for its objectives and was close to finalizing the details. Three issues had come up that he said were not unique to a regression-based approach:

Top of Page

Single-lab Subgroup Report

Richard Burrows reported that the single-lab subgroup had looked at two approaches:

  1. testing procedures the FACDQ selected by asking labs to run the tests or
  2. ask labs to run spikes at different concentrations (below LC and above LQ) and use results to evaluate the policy question - what is a reasonable criterion for quantitation, precision and accuracy?

We could use the results to see if there is a big difference between LD and LQ. We could use data to optimize the procedures for a second test.

He said the subgroup had been leaning toward the second approach but then it realized that this was really asking labs to do a multi-laboratory study. Then the subgroup thought perhaps it made sense to combine the single and multi-lab study.

The subgroup had also talked about how to select labs for the test: it could be a mix of commercial, state and wastewater treatment plant labs, large and small. If the pilot tested a range of methods, no one lab would have to test all; they could be spread out among the labs. Concentrations given to labs would have to cover the range, LC - LQ. The MDL, despite its weaknesses, would be used to design spike levels. If the single and muli-lab studies were combined, one would have to check the qualifications of labs to do the work at those concentrations.

Mr. Reding noted the discussion of a two-phased pilot study. He asked if the Technical Work Group were far enough along to propose a phase 1 pilot - to test a specific number of procedures and labs - and ask for FACDQ approval to run it after the March meeting and receive a report at the summer meeting. The advantage would be that EPA could get started on scoping phase1 of the pilot. Then, the next step would be to optimize procedures for a phase 2 pilot.

Top of Page

Discussion of Whether to Gather Data or Set MQOs First

This proposal opened up a discussion about what was most important for the FACDQ to agree on first for the pilot study to go forward: 1) to identify the procedures to test or 2) to define the MQOs and what procedures should do - so the uses would drive the data-gathering. (Mr. Wheeler noted that all agreed the FACDQ would ultimately have to agree on MQOs; the question was at what point the decision needed to be made.)

Dick Reding indicated that EPA could take the data that labs generated to get a feel for where they were on LC, LD and LQ. Mr. LaFleur pointed out that data-gathering was different from evaluating procedures. Mr. Reding noted that it would be hard to specify what EPA wanted until it knew what the capabilities were. Mr. LaFleur pointed out, by contrast, that agreement on uses would determine the level of certainty needed.

Another point of discussion was whether the approach - "one size fits all" - would work or whether multiple approaches would work.

Ken Osborn noted that MQOs were lab objectives for data reporting while DQOs were client driven to support decision making; the latter varied with the client and the purpose. Since labs control MQOs, it was suggested that they should be the focus.

Mr. Wheeler summarized the choices under discussion as 1) having MQOs drive the procedures or 2) having data-gathering capabilities drive the setting of MQOs. He noted that a similar difference of opinion had emerged in discussions of the MQO Subgroup of the Policy Work Group. That group's discussions had led them toward a two-phase pilot. Phase 1 would consist of gathering data to inform the FACDQ and phase 2 would be more procedure-specific. This raised questions about whether or not LD, which the FACDQ had agreed it would not use, should again be considered in the testing.

In response to a comment that the Policy Work Group might not agree with the MQO Subgroup, Mr. Wheeler agreed that the Policy Work Group would have further discussions and the FACDQ would make the final decision.

Mr. Wheeler then asked what the Technical Work Group could take to the March meeting that would be helpful. Mr. LaFleur, who said he was not opposed to a two-phase pilot, suggested that it would be safest to take two proposals to the FACDQ: 1) an evaluation of existing procedures and 2) a data-gathering process where MQOs would be set later. Both needed to be fleshed out so the differences between them would be clear.

John Phillips said that with the data-gathering approach, more work would still be needed on MQOs/DQOs so the Technical Work Group would know how to analyze the data. Mr. Burrows suggested that decisions on MQOs could be made on a parallel track. Mr. LaFleur said that even if pressure to identify MQOs for the pilot had ebbed, MQOs were still needed for various uses.

In response to a question asking if the regression-based approach was specific enough for a procedure-based design, Mr. LaFleur indicated that he would need to know if the data gathered would answer all the questions the subgroup had. He said he thought they could present the concept of the approach without having to define the MQOs first.

Mr. Wheeler summarized the proposals for products for the March FACDQ as follows.

Mr. Reding suggested that outcomes in March could be either FACDQ authorization for the first of two phases of the regression-based approach or FACDQ go-ahead to gather data for single or multi-lab designs. If either were to happen, then between March and the next FACDQ meeting, they could gather data and test both single and multi-lab designs.

Mr. Wheeler said he thought the Technical Work Group could present a recommendation to the FACDQ.

Mr. Reding noted that at the first FACDQ meeting, the group agreed that when it could agree on uses and procedures, it would like to test procedures via data-gathering. If the FACDQ were to come out with a real agreement on uses, that would be a big step and critical for judging procedures.

Jim Pletl said that he understood the desire to gather data for decision making but he said that one had to have goals for a study and decide in advance on how to interpret the data. He urged the group to think long and hard about the specific objectives of a study before starting the study. David Kimbrough agreed that data gathering in and of itself would not get the FACDQ any closer to a decision; it would just delay that step. Mr. LaFleur agreed; he said the FACDQ needed to figure out MQOs rather than rushing to a pilot study. Mr. Reding also agreed that if the pilot wasn't feasible, the FACDQ should move on to the decision.

Mr. Pletl said he agreed that we needed a pilot test but the focus should be on why we were doing the pilot. He said he thought it would work to make progress on MQOs in a parallel effort but he thought the chances of success were higher if there was agreement on the objectives of the pilot and on how we would measure results.

In response to a question if the previous discussion meant that no data would be collected until the fall, Mr. Reding said he understood Mr. Pletl to favor identifying the questions to ask and agreeing on how we would evaluate the results.

There was discussion of the Episode 6000 series data -- its relevance and problems.

Mr. Wheeler again summarized the implications for product-development and future work:

After discussion the group agreed that presenting two approaches made sense because it would give the FACDQ two approaches to respond to: test procedures to see if they work or generate information and design a new procedure that meets our needs. From a product-development standpoint, this would involve

In response to a question about how this approach would work for EPA, Mr. Reding said it would be a heroic effort and he was uncertain about funding.

Mr. Pletl again raised the importance of deciding what we want from procedures before getting into a pilot because it would help the FACDQ reach consensus on the study design. In response to a question about MQOs, Mr. Pletl said that narrative statements would work; MQOs did not have to be numbers. For example, he said, do we want a procedure that addresses bias? If so, one could use that to screen procedures. It is an issue, he said, that the FACDQ would have to discuss now or in the future. In his opinion, collecting data would be the easy part.

Mr. Wheeler asked about the relevance of the MQOs and DQOs in the matrix. Mr. Pletl said that he viewed the matrix as an information tool to decide which procedures we want but it is a list of all the characteristics. The FACDQ hasn't said what we want from a procedure.

Mr. Wheeler asked how the group would get its products ready for the March FACDQ meeting.

Mr. Burrows suggested a study of Episode 6000 series data to see what data we want to pull out. Mr. Pletl suggested that if there were concerns about a dataset, one should identify them to help design the studies. He noted that Ken Miller was a good person to look at the data but that Mr. Miller would need direction on what information we wanted to extract from the data. Mr. LaFleur suggested compiling a list of questions we need to answer from the broadest perspective and figuring out which ones can be answered using the 6000 series data and which ones it won't answer and figure out how else we can answer them. That list would direct Ken Miller's work.

Because of time constraints, Mr. Wheeler noted that Ken Miller's report of his work on the matrix would be on the agenda for the next Technical Work Group meeting.

Mr. Wheeler also noted that he would work with Richard Burrows and Tom Georgian on the paper on LD. Mr. Phillips indicated that he intended to condense the document that summarized the discussions on LD into a shorter document. Its purpose he said was to be a resource, to identify issues. Before the next meeting, he said he would prepare a short summary of the pros and cons of LD for the single lab study.

Top of Page

Wrap-up and Closing

Mr. Wheeler summarized the action items from the discussion and thanked members for their hard work. He adjourned the meeting at 3:10 p.m. EST.

Top of Page

Attendance

Technical Work Group Members

Triangle Associates: Robert Wheeler and Vicki King

Observers

Top of Page

Local Navigation

Jump to main content.