P&B, IDE, AND IQE Defined for Ideal Water-Testing
ASTM Committee D19 on Water has produced three standard practices
that fully characterize the performance of analytical methods
at the interlaboratory level. Developed for water-testing methods,
these standards have broad applicability where an increased level
of the analyte results in an increased response. The standards
can be applied using data from a single interlaboratory round-robin
study, provided the selected concentrations span the trace-level
and working ranges of the method. They explicitly incorporate
a common phenomenon: an increased level of the analyte results
in an increased standard deviation of the measurements.
Created by large task forces of over 200 chemists, laboratory
technicians, and water-industry specialists in Subcommittee D19.02
on General Specifications, Technical Resources, and Statistical
Methods, the standards are:
D 2777, Standard Practice for Determination of Precision and Bias
of Applicable Test Methods of Committee D19 on Water, aids development
of round-robin-study designs, evaluation of data and rejection
of outliers, and calculation of precision and bias over the working
range of a method.
With a few exceptions, the requirements of D 2777 must be satisfied
for approval of new D19 test methods; therefore D 2777 is referred
to as the Method Validation or the Precision and Bias standard
practice. Released in 1998, it provides extensive guidance to
planners of interlaboratory studies. Procedures for selection
of test concentrations, management of participating laboratories,
role and responsibilities of a study supervisor, and Youden-pair
study designs are included. Also provided are application of Grubbs
method and lab-ranking to screen data for outliers, and formulas
for predicting the precision and bias of a method.
D 6091, Standard Practice for 99%/95% Interlaboratory Detection
Estimation (IDE) for Analytical Methods with Negligible Calibration
Error Practice, calculates the lowest level of reliable detection.
Released in 1997, D 6091 depends on the interlaboratory-study
approach of D 2777. Within a population of laboratories, the IDE
is computed to be the lowest concentration at which one can be
90 percent confident to achieve detection 95 per cent of the time,
and no detection for 99 percent of measured blanks.
D 6091 provides guidance on inclusion of concentration levels
that are consistent with determination of detection, for example,
the addition of a few low-concentration samples to the historically
working-range samples of D 2777. The IDE document describes additional
analysis of the study data, after outlier removal, to compute
the IDE. An innovation is the choice and use of models that relate
the interlaboratory standard deviation (ILSD) to concentration.
To generate an IDE value that reflects data-quality objectives,
the user is able to select alternatives for the confidence level,
the true-detection rate, and the non-detection rate for blanks.
Nearing publication, D 6512, Standard Practice for an Interlaboratory
Quantitation Estimate (IQE) covers calculation of the lowest level
of reliable quantitation. The IQE is computed to be the lowest
concentration with a marginally acceptable relative standard deviation
(RSD = 30%, 20%, or 10%; all, some, or none of these values may
be achievable for a given test method) among a population of laboratories.
D 6512 uses the data and the interlaboratory-study approach of
D 2777, and involves modeling the interlaboratory standard deviation,
as in D 6091. The user can evaluate the suitability of methods
by comparing each methods IQE values to data-quality objectives.
Collectively, these standards provide a new, technically sound
means for method-performance characterization. D 2777 forces the
refinement of a method to the point that multiple laboratories
can perform the test and get comparable results. This precision-and-bias
document also characterizes the behavior of a method within its
working range, which might span two or more orders of magnitude
in spike concentrations. Consistent with recent ISO quantitation
and detection concepts in Nomenclature in Evaluation of Analytical
Methods, Including Detection and Quantification Capabilities,
Pure and Applied Chemistry, Vol. 67, No. 10, 1995, by L. Currie,
the IQE and the IDE standard practices allow determination of
a methods lower levels of usefulness. The IQE can be treated
as the lowest level at which a measurement value might be recorded
for routine use, and the IDE can be seen as the lowest level at
which one can confidently achieve detection.
ASTM standards are available by calling Customer Service (610/832-9585)
or through the Web site (www. astm.org).
For further technical information on D 6091 (IDE) or D 6512 (IQE),
contact D19 member David Coleman, Alcoa Technical Center, 100 Technical Dr., Alcoa Center, PA
15069 (724/337-5913). For further D 2777 (P&B) information, contact
member Lynn Vanatta, Air Liquide, Box 650311, M/S 301, Dallas, TX 75265 (). Committee
D19 meets June 24-28 in Tucson, Ariz. For meeting or membership
details, contact manager Bruce Noe, ASTM (610/832-9719). //