Selection, Verification and Validation of Methods

1
SELECTION,
VERIFICATION AND
VALIDATION OF METHOD
08 JUL 2016
masih
Selection
2
ISO/IEC 17025:2005
Clause 5.4.2
The laboratory shall use test and/or calibration
methods, including methods for sampling, which meet
the needs of the customer and which are appropriate
for the tests and/or calibrations it undertakes
EURACHEM Guide. The Fitness for Purpose of Analytical Methods: A Laboratory Guide
to Method Validation and Related Topics, 2nd ed. 2014
Verification
3
The provision of objective evidence that a given item

fulfills specified requirements. [VIM 2.44]
Validation
4
Verification, where the specified requirements are

adequate for intended use. [VIM 7]
Confirmation by examination and provision of
objective evidence that the particular requirements
for a specific intended use are fulfilled. (ISO/IEC
17025)
Selection
Type of Analytical Application
Performance
Characteristic
Selectivity
Identification
Test
Quantitative
test for
impurity
Limit test
for
impurity
Quantification
of main
component
Limit of Detection
Limit of Quantitation
Working range
Trueness (bias)
Precision (repeatability
and intermediate
precision)
x signifies normally validated performance characteristics for four types of

analytical applications
Definition
6
Reagent Blank - reagents used during the analytical

process (including solvents used for extraction or
digestion) are analyzed in order to determine if they
contribute to measurement signal. The signal arising
from the analyte can then be corrected accordingly.
Method Blank reagent blanks that are taken
through the whole measurement procedure or
method process
Definition
7
Sample Blank these are essentially sample matrices

with no analyte present. These are difficult to obtain
but such materials are necessary to give a realistic
estimate of interferences that would be encountered
in the analysis of test samples.
EURACHEM Guide.The Fitness for Purpose of Analytical Methods: A Laboratory Guide

Performance Characteristics
Method Performance Characteristics

9
Selectivity
Limit of Detection (LOD) and Limit of Quantitation
(LOQ)
Working Range
Trueness (bias, recovery)
Precision (repeatability, intermediate precision,
reproducibility)
Ruggedness (robustness)
Method Performance Characteristics

10
***Measurement Uncertainty is not a performance

characteristic but a property of the results obtained
using that method.
Selectivity
11
The ability of a method to determine accurately and

specifically the analyte of interest in the presence of
other components in a sample matrix under the stated
conditions of the test. [EURACHEM Guide]
Selectivity
12
What to do
How many times
Analyze test sample

containing various
suspected
interferences in the
presence of the
analytes of interest
What to calculate
from the data
Examine
interference effect.
EURACHEM Guide - Selectivity
Comment
If detection or
quantification is
inhibited by the
interferences, further
method development
will be required.
Concepts on Low Concentration

Measurements
13
1. Establish a value that an analyte level is significantly

different from zero (critical value or decision limit)
2. Know the lowest concentration of the analyte that can
be detected by the method at a specified level of
confidence (limit of detection, minimum detectable
value or detection limit)
3. Establish the lowest level at which the performance is
acceptable for a typical application (limit of
quantitation)
Limit of Detection
14
The lowest amount of analyte in a sample which can

be detected but not necessarily quantitated as an
exact value. [ICH Q2A and ICH Q2B, ASEAN Guide]
The lowest content that can be measured with

reasonable statistical certainty. [AOAC-PVMC]
S0
(standard deviation from
results of measurement)
Limit of Detection
15
Results blank
corrected
during
routine?
YES
NO
s0 '
s0
n
Use the calculated s0 for

computing the value of LOD
and LOQ
s 0 ' s0
1 1
n nb
Limit of Detection
16
Instrument Detection Limit

is based on sample analysis, often a reagent blank
presented directly to the instrument (omitting sample
preparation steps). When the reagent blanks do not go
through the whole measurement procedure, the
calculated data will give the Instrument LOD.
EURACHEM Guide LOD (Instrument Detection Limit)
What to do
Replicate measurements
of reagent blanks
How many What to calculate from the

Comment
times
data
10
Calculate standard
for non-blank corrected results
deviation of the results, s0.
s0 '
or
Calculate s0.
of reagent blanks spiked
with low concentrations
of analyte
Calculate LOD
LOD 3 s0 '
This approach is
acceptable when blank
samples or test samples
at low concentrations
cannot be obtained
s0
For blank-corrected results
s0 ' s0
1
1
n nb
S0 standard deviation of result

at or near zero concentration
S0 standard deviation for
calculating LOD and LOQ
n number of replicate
observations averaged when
reporting results
nb number of blank observations

averaged when calculating blank
correction
EURACHEM Guide.The Fitness for Purpose of Analytical Methods: A Laboratory Guide
17
Limit of Detection
18
Method Detection Limit

LOD is based on the analysis of samples that have been
taken through the whole measurement procedure using
results calculated with the same equation as the test
sample.
Samples that are suitable for estimation of LOD or LOQ
are: blank samples (matrices with no detectable
analyte), test samples with concentrations of analyte
close to or below the expected LOD.
EURACHEM Guide LOD (Method Detection Limit)

What to do
19Replicate measurements
of blank samples
or
How many
times
10
What to calculate from the

Comment
data
Calculate standard
s
s0 '
Calculate s0.
of test samples with low
analyte concentration
Calculate LOD
LOD 3 s0 '
or
Replicate measurement
of reagent blanks
Or
with low analyte
concentration
s0 ' s0
1
1
n nb

reporting results
nb number of blank
calculating blank correction
Limit of Detection
20
LOD based on standard deviation of QC results at low

concentration levels under intermediate precision
conditions.
1.
2.
The standard deviation s0 is equal to s0

Since blank correction is no longer required, LOD
can be computed directly by using s0
multiplied by 3.
21
The lowest level of analyte that can be determined

with acceptable performance.
EURACHEM Guide LOQ

What to do
22Replicate measurements
of blank samples
or
How many
times
10

Comment
data
Calculate standard
s
s0 '
Calculate s0.
of test samples with low
analyte concentration
or
of reagent blanks
Calculate LOQ
IUPAC default kQ value is

usually 10
s0 ' s0
1
1
n nb
LOQ kQ s0 '

LOQ 10 s0 '

or
with low concentrations
of analyte
reporting results
nb number of blank
calculating blank correction
23
LOQ based on the standard deviation of quality control

results at low concentration levels under intermediate
precision conditions.
1. The standard deviation s0 is equal to s0

2. Since blank correction is no longer required, LOQ
can be computed directly by using s0 multiplied by
10.
Working Range
24
The interval over which the method provides results

with an acceptable uncertainty.
The lower end is bounded by the limit of
quantification and the upper end of the working
range is defined by concentrations at which significant
anomalies in analytical sensitivity are observed.
Working Range
25
Instrument Working Range

In method validation, the instrument working range is
assessed to: confirm the relationship (e.g. linear,
curvilinear, etc) of the instrument response between the
LOQ and upper end of the instrument working range;
Working Range
26
Method Working Range

Assessing method working range requires use of sample
blanks and samples with known concentrations. These
samples shall undergo the entire measurement
procedure (method) and preferably cover the whole
working range. Results are calculated based on the
method.
EURACHEM Guide Working and linear range

How many
What to do
What to calculate from the data
times
1.Measure blank and
1
Plot
27calibration standards at 6Response y-axis
10 concentrations (evenly
Concentration x-axis
spaced across the range of
Visually identify approximate
interest)
linear range and upper and lower
working range boundaries. Proceed
to step 2.
2.Measure blank plus 6-10
1
Plot
calibration standards, 2 to
Response y-axis
3 times. (calibration
Concentration x-axis
standard concentrations
Calculate appropriate regression
evenly spaced across
statistics. Calculate and plot
linear range.
residuals (difference between
observed y-value and calculated yvalue predicted by the straight line,
for each x value). Linearity is
confirmed when there is random
distribution of residuals. Systematic
trends indicate non-linearity or
change in variance with level
Comment
Visual confirmation if
working range is linear
or not
If the standard deviation

is proportional to
concentration, consider
using weighted
regression rather than
simple non-weighted
linear regression.
In certain circumstances, it
may be better to try to
fit a non-linear curve to
the data. Functions higher
than quadratic are
generally not advised.
EURACHEM Guide Working and linear range

How
What to do
many
times
283.Calibrate according to
1
Plot measured concentration y-axis against
the proposed calibration
test sample concentration x-axis.
procedure. Measure
Visually identify approximate linear range
according to the
and upper and lower working range
method, blank plus RMs
boundaries.
or spiked sample blanks
Calculate appropriate regression statistics.
2 to 3 times at 6-10
Calculate and plot residuals (difference
different concentrations
between observed y-value and calculated
evenly spaced across
y-value predicted by the straight line, for
the range of interest.
each x value). Linearity is confirmed when
there is random distribution of residuals.
Systematic trends indicate non-linearity
Comment
This is to assess
whether instrument
range and calibration
procedure are fit for
purpose.
Accuracy
29
The closeness of agreement between a measured

quantity value and a true quantity value of a
measurand [VIM 2.13]
Therefore, it is normally studied as two components,

trueness and precision. In addition, an increasingly
common expression of accuracy is measurement
uncertainty.
Trueness (bias)
30
The closeness of agreement between the average of

an infinite number of replicate measured quantity
values and a reference quantity value. [VIM 2.14]
Trueness is inversely related to systematic

measurement error which may be estimated as bias.
EURACHEM Guide Trueness

How many
What to do
times
10
31Measure RM using
candidate method

Calculate
bias,
Comment
bias taking into
account effect of
method and
laboratory bias
b x xref
%relative bias,
b (%)
x xref
xref
100
% relative recovery (apparent

recovery)
R (%)
Measure matrix
blanks/test samples
unspiked and spiked
with the analyte of
interest over a range
of concentrations
10
x
xref
100
Calculate
relative spike recovery,
R ' (%)
x 'spiked x
xspike
assess net recovery of

added spike
100
EURACHEM Guide Trueness

How many
What to do
times
10
32Measure RM/test
sample using
candidate method
and alternative
method

Calculate
bias,
alternative method
may be a reference
method
b x xref
%relative bias,
b (%)
x xref
xref
100
% relative recovery (apparent

recovery)
R (%)
x
xref
Comment
100
Estimated Recovery
33
Analyte
Concentration, %
Analyte
Ratio
Unit
Mean
% Recovery
100
100%
98 102
> 10
10-1
10%
98 102
>1
10-2
1%
97 103
> 0.1
10-3
0.10%
95 105
0.01
10-4
100 ppm
90 107
0.001
10-5
10 ppm
80 110
0.0001
10-6
1 ppm
80 110
0.00001
10-7
100 ppb
80 110
0.000001
10-8
10 ppb
60 115
0.0000001
10-9
1 ppb
40 120
AOAC Manual Peer Verified Methods
Precision
34
The closeness of agreement between indications or

measured quantity values obtained by replicate
measurements on the same or similar objects under
specified conditions. [VIM 2.15]
Measurement precision is related to random
measurement error [VIM 2.19] and is a measure of
how close results are to one another.
Precision
35
Repeatability
is a measure of the variability in results when a
measurement is performed by a single analyst using
the same equipment over a short timescale.
Intermediate Precision
gives an estimate of the variation in results when
measurements are made in a single laboratory but
under conditions that are more variable than
repeatability conditions (different analysts, extended
timescale, different pieces of equipment, etc).
Precision
36
Reproducibility
is a measure of the variability in results between
laboratories using the same method or using different
methods but intending to measure the same quantity.
EURACHEM Guide Precision

How many
What to do
times
a. same analyst
6 to 15
37and equipment,
replicates for
same laboratory, each material
short timescale
b. different
6 to 15
analysts and
replicates for
equipment, same
each material
laboratory,
extended
timescale
c. different
analysts and
equipment, same
laboratory,
extended
timescale
6 to 15 groups
of duplicate
measurements
obtained under
repeatability
conditions on
different days/
equipment for
each material

Comment
data
Calculate standard
Estimates Repeatability standard
deviation of results for each deviation sr for each material
material
Calculate standard
Estimates Intermediate Precision
deviation of results for each standard deviation sI for each
material
material
Calculate repeatability
standard deviation from
ANOVA results for each
material.
Calculate between-group
ANOVA and combine with
repeatability standard
deviation for each material

deviation sr for each material
Estimates Intermediate Precision
standard deviation sI for each
material
EURACHEM Guide Precision

How many
What to do
times
d. different
6 to 15 groups
38analysts and
of duplicate
equipment,
measurements
different
obtained under
laboratory,
repeatability
extended
conditions in
timescale
different
laboratories for
each material

Comment
data
Calculate repeatability
deviation sr for each material
ANOVA results for each
material.
Estimates Reproducibility standard
deviation sR for each material. This
Calculate betweenrequires a special inter-laboratory
laboratory standard
comparison (collaborative trial)
deviation from ANOVA
results and combine with
repeatability standard
deviation for each material
Estimated Precision
39
Analyte
Concentration, %
Unit
%RSD
100
100%
1.3
10
10%
2.3
1%
2.7
0.1
0.10%
3.7
0.01
100 ppm
5.3
0.001
10 ppm
7.3
0.0001
1 ppm
11
0.00001
100 ppb
15
0.000001
10 ppb
21
0.0000001
1 ppb
30
AOAC Manual Peer Verified Methods
Ruggedness
40
The ruggedness of an analytical procedure is a

measure of its capacity to remain unaffected by
small, but deliberate variations in method parameters
and provides an indication of its reliability during
normal usage. [ICH Q2A and ICH Q2B, ASEAN
Guide]
EURACHEM Guide Ruggedness

What to do
Identify variables which
41could have a significant
effect on method
performance.
Set up experiments
(analysing RMs or test
samples) to monitor the
effect on measurement
results of systematically
changing the variables.
How many times
Most effectively
evaluated using
experimental
designs.
e.g. 7 parameters
can be studied in
8 experiments
using a
Plackett-Burman
experimental
design.
Determine the effect of each

change of condition on the
measurement results.
Rank the variables in order of
the greatest effect on method
performance.
Carry out significance tests to
determine whether observed
effects are statistically
significant.
Comment
Design quality
control or modify
the method in order
to control the
critical variables,
e.g. by stating
suitable tolerance
limits in the
standard operating
procedure
Analytical Sensitivity
42
Analytical sensitivity is the change in instrument

response which corresponds to a change in the
measured quantity
Analytical Sensitivity
43
Useful Applications:
1.
Theoretical analytical sensitivity is sometimes known.
(e.g. ion-selective electrode show a Nernstian
behavior.)
2.
In spectrophotometric analysis, absorbance can be
predicted from Beer-Lamberts law.
*Measurement Uncertainty
44
Interval associated with the measurement result which

expresses the range of values that can reasonably be
attributed to the quantity being measured.
It is not a performance characteristic but a property
of the results obtained using the method.
45
Exercise
Exercise
46
Method Validation for Determination of Fe in food

sample was done using a AAS. The expected Fe
content of the sample is around 100g/g.
a.
What are the performance characteristics that
should be included in the study?
b.
The method for routine requires duplicate analysis
for the blank and test sample. Results were
g/mL 50mL
calculated as : Fe, g/g AAS Reading,
Weight Sample, g
Calculate for LOD and LOQ using the data
obtained from the validation study.
47
Trial
Wt (g)
Method Blank
Sample Blank
Reading (g/mL) Reading (g/mL)
1.0003
0.0691
0.2070
1.0003
0.0604
0.2017
1.0003
0.0702
0.2051
1.0001
0.0756
0.1998
1.0003
0.0740
0.2097
1.0006
0.0631
0.2615
1.0000
0.0633
0.1908
1.0000
0.0482
0.2042
1.0004
0.0561
0.1907
10
1.0004
0.0509
0.2179

Selection, Verification and Validation of Methods

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Selection, Verification and Validation of Methods

Uploaded by

Copyright:

Available Formats

1

The provision of objective evidence that a given item

Verification, where the specified requirements are

x signifies normally validated performance characteristics for four types of

Reagent Blank - reagents used during the analytical

Sample Blank these are essentially sample matrices

EURACHEM Guide.The Fitness for Purpose of Analytical Methods: A Laboratory Guide

Method Performance Characteristics

Method Performance Characteristics

***Measurement Uncertainty is not a performance

The ability of a method to determine accurately and

How many times

Analyze test sample

EURACHEM Guide - Selectivity

Concepts on Low Concentration

1. Establish a value that an analyte level is significantly

The lowest amount of analyte in a sample which can

The lowest content that can be measured with

Use the calculated s0 for

Instrument Detection Limit

EURACHEM Guide LOD (Instrument Detection Limit)

How many What to calculate from the

For blank-corrected results

S0 standard deviation of result

nb number of blank observations

Method Detection Limit

EURACHEM Guide LOD (Method Detection Limit)

What to calculate from the

For blank-corrected results

S0 standard deviation of result

LOD based on standard deviation of QC results at low

The standard deviation s0 is equal to s0

The lowest level of analyte that can be determined

EURACHEM Guide LOQ

What to calculate from the

IUPAC default kQ value is

For blank-corrected results

S0 standard deviation of result

S0 standard deviation for

LOQ based on the standard deviation of quality control

1. The standard deviation s0 is equal to s0

The interval over which the method provides results

Instrument Working Range

Method Working Range

EURACHEM Guide Working and linear range

If the standard deviation

EURACHEM Guide Working and linear range

The closeness of agreement between a measured

Therefore, it is normally studied as two components,

The closeness of agreement between the average of

Trueness is inversely related to systematic

EURACHEM Guide Trueness

What to calculate from the data

% relative recovery (apparent

assess net recovery of

EURACHEM Guide Trueness

What to calculate from the data

% relative recovery (apparent

AOAC Manual Peer Verified Methods

The closeness of agreement between indications or

EURACHEM Guide Precision

What to calculate from the

Estimates Repeatability standard

EURACHEM Guide Precision