Procedures for Detection and Quantitation

Federal Advisory Committee on Detection and Quantitation for Uses in Clean Water Act Programs

Meeting #10

Development of Quantitation Limits for National Use

The Technical Workgroup asked the Multi/Inter Lab TWG subgroup (Dick Reding, Tim Fitzpatrick, Larry LaFleur, Ken Miller, Brad Venner, Rick Rediske, and John Phillips) to discuss:

How to develop a recommended process for determining a QL_nat.
Develop recommendations on multi/interlab approaches
What does pooling mean?

On August 29, 2007, the TWGrp reviewed and discussed the findings of this subgroup. The Technical Workgroup is recommending the following for consideration by the FACDQ.

Multilab Approach to Develop QL_nat

We discussed the relative merits of the two approaches used in the pilot: interlab/regression (e.g., several labs develop data from which a single QL_nat is calculated) and multilab (several labs develop their QL-lab which are "pooled" to form a QL_nat). We agreed that a multilab approach was acceptable. It allows use of the wealth of QL-labs that labs will be routinely developing and reporting to clients or databases, such as ICIS. For future method development, it allows use of the FACDQ's single lab QL procedure to develop the QL_nat. The multilab approach does not close the door on use of regression techniques where feasible or appropriate.

To develop QL_nats we considered two types of methods:

QL_nat for future Part 136 candidates, i.e. new methods, and
QL_nat for future updates to existing Part 136 methods.

1. Future Part 136 Methods

For future methods we will have less time to collect data, and fewer labs with experience in the new chemistries and technologies. We discussed specific recommendations for: minimum number of labs, how long to collect data, and how to pool these single lab limits into a QL_nat.

We generally agreed on a minimum 6-7 labs, collecting data over 3- 6 months with a minimum of 20 to 30 QL spikes used in the calculation of each single lab limit. Because of the size of this data set we recommended that QL_nat be chosen to be the highest single lab limit.

Some members noted that not all new methods needed data from 6-7 labs, and suggested instead that in some cases, that method analytes be promulgated without a QL_nat. We would then wait for a more robust set of data to emerge from the user community that would allow development of a QL_nat more along the lines described next for future updates to existing Part 136 methods.

2. Future Updates to Existing Part 136 Methods

As the FACDQ single lab procedure is adopted, labs will be developing data with certain widely used Part 136 methods. Therefore our universe of labs and time to collect data is not as limited as for future methods. There are many techniques available to cull this data and use it to set a QL_nat. We note that a substantial amount and type of metadata would be needed to interpret and the QL-labs and preferably be available in a central system, such as ICIS, to facilitate QL_nat development.

Top of Page