Method Verification

The HowM.L. Jane Weitzel ALACC Chair

The What

Released early 2008

http://www.aoac.org/alacc_guide_2008.pdf

Meet ISO 17025 Requirement

Categories of Methods

The six categories of chemical analytical methods are:

1. Confirmation of Identity, a method that ensures a material is what it purports to be or confirms the detection of the target analyte.

2. Quantifying an analyte at a low concentration. 3. Determining if an analyte is present above or

below a specified, low concentration (often called a Limit Test). The specified concentration is close to the LOQ.

Categories of Methods

4. Quantifying an analyte at a high concentration.

5. Determining if an analyte is present above or below a specified, high concentration (often called a Limit Test). The specified concentration is substantially above the LOQ.

6. Qualitative test.

Tables 2 -6

Requirements of Method Verification for the Six Categories of Chemical Test Methods (Tables 2–6)

Category 1: Confirmation of Identity

ISO Technical Specification ISO/TS 21748, Guidance forthe use of repeatability, reproducibility and trueness estimates in measurement uncertainty estimation

ISO/TS 21748Template for Verification ISO 21748 is a thorough guide to verifying

a method. The examples will be based on ISO 21748

“Specification for the Method”

The results of collaborative study yield performance indicators (sR, sr) and, in some circumstances, a method bias estimate, which form a “specification” for the method performance.

In adopting the method for its specified purpose, a laboratory is normally expected to demonstrate that it is meeting this “specification.”

Repeatability

In most cases, this is achieved by studies intended to verify control of repeatability and of the laboratory component of bias, and by continued performance checks (quality control and assurance)

Many added benefits

Method verification Estimate of Uncertainty Better understanding of the method

Equation

y= μ +δ +B+ Σcix′i +e

Equation Components

y = µ + δ + Β + e + Σ cix′i

Result = ideal result + method

bias + lab bias + repeata

bility +sum of effects NOT included in collabora-

tive study

What if your method was not collaboratively studied?

The approach described by ISO 21748 is still relevant. Intermediate precision could be substituted for

reproducibility Fewer variables would be included in intermediate

precision study than during a reproducibility study There would be more effects included in Σ cixi’ The approach provides a detailed, organized procedure for

identifying and evaluating these effects

Method Bias

δmethod bias

Collaborative StudyNote: the study may not include this

component

Lab Bias

Β

lab bias

Collaborative StudysL

2

sR2 = sL

2 +sr2

Thus

sL2 = sR

2 - sr2

Repeatability

e

repeatability

Collaborative Studysr

2

Effects NOT includedΣ cix′i

sum of effects NOT included in collaborative study

Examine the equation and procedure.

Table Summary

y µ δ Β e Σ cix′i

Result Ideal result Method bias Lab bias Repeatability

Sum of effects not included

in collaborative

study

Source of Specification

for the method

Collaborative Study. Note:

the study may not include

this component

CollaborativeStudysL2

sR2 = sL

2 +sr2

ThussL

2 = sR2 - sr

2

CollaborativeStudy

sr2

Examine the equation and

procedure

Method Bias

δmethod bias

Collaborative StudyNote: the study may not include this component.

The study may include method bias, if for example the results are corrected for a known method bias. The procedure for assessing method bias is included in

the ISO standard in detail in “Incorporating Trueness Data”.(Uncertainty associated with CRV becomes important)

Repeatability - Equation

erepeatability

Collaborative Study sr2

Do repeatability study in lab. Calculate sw.Compare to sr

2 using the F test.

Bias

Often “accuracy” must be verifiedTerminology – bias

Can use same data from repeatability If Certified Reference Material is available

Lab Bias

Β

lab bias

Collaborative StudysL

2

sR2 = sL

2 +sr2

Thus

sL2 = sR

2 - sr2

Bias Specification

Follows ISO Guide 33 Δ = m - µ = mean - CRV

|Δ| < 2σD

|Δ| < 2√ (sR2 - sr

2 + sw2/n)

sL2 = sR

2 - sr2

n

Choose n so that the uncertainty of the bias is not significant

choose n such that the uncertainty √sw

2/n < 0.2sR

Good rule of thumb uncertainties less than 0,2 sR lead to changes of

under 0,02 sR in the overall uncertainty estimate.

Σ cix′i Look at equation to identify any source of

uncertainty that was not included in the collaborative study.

It is often convenient to consider each of the three factors the sample, the laboratory and the method when identifying gross uncertainties

Specificity

If your sample is not identical to those included in the collaborative study or method validation, you must assess the impact of the differences.

AOAC Food triangle can be useful

Useful Table 8 in ALACC Method Verification Guide Lists Parameter The difference from the validated method Required Equivalence Study

Useful Table 8

Table 8 Example

Uncertainty

For a collaboratively studied method and if you follow the process from ISO 21748, you will also end up with an estimate of the uncertainty for the method.

Table Summary

y µ δ Β e Σ cix′i

Result Ideal result Method bias Lab bias Repeatability

Sum of effects not included

in collaborative

study

Source of Specification

for the method

Collaborative Study. Note:

the study may not include

this component

CollaborativeStudysL2

sR2 = sL

2 +sr2

ThussL

2 = sR2 - sr

2

CollaborativeStudy

sr2

Examine the equation and

procedure

Examples