0% found this document useful (0 votes)
593 views9 pages

Iso TS 21748

ISO TS 21748

Uploaded by

Diana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
593 views9 pages

Iso TS 21748

ISO TS 21748

Uploaded by

Diana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 9

ISO/TS 21748: A new guide to uncertainty

estimation using interlaboratory study data


S Ellison
LGC

/Eurachem-UK/1

Overview

• Reproducibility and uncertainty estimation

• Introduction to ISO TS 21748

/Eurachem-UK/2

What is Measurement
Uncertainty?

“A parameter, associated with the result of a measurement,


that characterises the dispersion of the values that could
reasonably be attributed to the measurand”
(ISO Guide)

The number after the ±

/Eurachem-UK/3

Method Validation Seminar June 06


© 2006 LGC Limited 1
Models and Uncertainty
xi xj xk

y = f(xi, xj, xk, .) y

/Eurachem-UK/4

Models and Uncertainty


xi xj xk

y = f(xi, xj, xk, .) y

u(y)

/Eurachem-UK/5

Models and Uncertainty


xi xj xk
+ +

y = f(xi, xj, xk, .) y

u(y)

/Eurachem-UK/6

Method Validation Seminar June 06


© 2006 LGC Limited 2
ISO Guide approach

• Specify the measurand


– including complete equation
• Quantify significant uncertainties in all parameters
– A: from statistics of repeated experiment
– B: by any other means (theory, certificates, judgement...)
• Express as standard deviation
• Combine according to stated principles

/Eurachem-UK/7

Combining uncertainties
(ISO)

u 2 +u 2
1 2

u2

u1

• Standard deviations
• Established error propagation theory

/Eurachem-UK/8

Example: Forensic alcohol standard


titration
Oxidant conc.

Extent Oxid'n

Titration vol.

Density

TOTAL

0 0.02 0.04 0.06 0.08 0.1 0.12


Uncertainty contribution (mg/100ml)

/Eurachem-UK/9

Method Validation Seminar June 06


© 2006 LGC Limited 3
Implementing ISO in Chemistry
Building models

• Every determinand is unique


– Every element, every molecule, every formulation
• Every ‘matrix’ is unique
– Different interactions with substrate
• Interactions with environment and substrate rarely
understood
• Models are difficult to build!

/Eurachem-UK/10

Two views of influence factors


xi xj xk
+ +

y = f(xi, xj, xk, .) y

u(y)

/Eurachem-UK/11

Two views of influence factors


xi xj xk
+ +

Analyst 1 y1
Analyst 2 y2
y3
Analyst 3
y4
Analyst 4
Combined
distribution sR u(y)

Nominal range Values

/Eurachem-UK/12

Method Validation Seminar June 06


© 2006 LGC Limited 4
A simpler model

• The best available estimate of precision


– An effect varied representatively during a precision
experiment requires no further study
• The best available estimate of bias and its uncertainty
• Other significant effects evaluated
– By experiment, or from standing data

/Eurachem-UK/13

Collaboratively trialled methods

• Collaboratively trialled method


– bias and precision verified in-house
• Assessed against collaborative study estimates
– ISO/TS 21748
– demonstrate method bias and repeatability are under control

/Eurachem-UK/14

Collaborative trial data

Method Bias
Reproducibility (collaborative
sd (sR) study)

• sR - random
MU • Use of data
selection of most requires correct
main effects collaborative
study and sound
QC/QA
Other
Effects
/Eurachem-UK/15

Method Validation Seminar June 06


© 2006 LGC Limited 5
Process for evaluating
uncertainty according to ISO/TS
21748

• Obtain repeatability, reproducibility and bias estimates


from collaborative study
• Establish whether laboratory bias is within that expected
on the basis of the collaborative study
• Establish whether laboratory precision is within that
expected on the basis of the collaborative study
• Where laboratory bias and precision are under control,
combine effects appropriately to form a combined
uncertainty estimate

/Eurachem-UK/16

TS21748 uncertainty budget

()
uc2 (y ) = sˆR2 + u 2 δˆ + ∑ c i2 u 2 (x i )
=
i k,n

• Combined uncertainty
• Adjusted reproducibility estimate

• Uncertainty associated with method bias

• Effects not covered, e.g. sampling uncertainty

/Eurachem-UK/17

Demonstrating acceptable
laboratory bias

• Study of CRMs
– Simple test for significant bias
– Ideally with small bias check uncertainty

• Comparison with known test method


– Typically paired tests on several materials

• Comparison with other laboratories


– Proficiency testing or other collaborative study

/Eurachem-UK/18

Method Validation Seminar June 06


© 2006 LGC Limited 6
Including trueness data in an
uncertainty estimate

()
uc2 (y ) = sˆR2 + u 2 δˆ + ∑ c i2 u 2 (x i )
=
i k,n

• Combined uncertainty
• Adjusted reproducibility estimate

• Uncertainty associated with method bias

• Effects not covered, e.g. sampling uncertainty

/Eurachem-UK/19

Demonstrating acceptable
laboratory precision

• Requires a check on repeatability standard deviation


– F-test at appropriate level.

• If within-laboratory precision differs from Study sr


– Replace study sr with in-house repeatability SD

• Otherwise, continue using published sR

/Eurachem-UK/20

Adjusting for changes in test


item

• Allow for inhomogeneity and sample prep via additional


uncertainty contributions
• Allow for changes in precision with analyte level
– Linear model most common
– Often sufficient to use %RSD

/Eurachem-UK/21

Method Validation Seminar June 06


© 2006 LGC Limited 7
Including reproducibility data

uc2 (y ) = sˆR2 + u 2 δˆ + () ∑ c i2 u 2 (x i )
=
i k,n

• Combined uncertainty
• Adjusted reproducibility estimate

• Uncertainty associated with method bias

• Effects not covered, e.g. sampling uncertainty

/Eurachem-UK/22

Other effects

• Simple “worst case” calculations can often show an


effect is negligible
• Formal uncertainty calculations often show negligible
components
• Typical criteria:
– u<umax/3 ( <6% effect on combined uncertainty)
– u<umax/5 ( <2% effect on combined uncertainty)
– ISO/TS 21748 effects <0.2SR

/Eurachem-UK/23

Numerical calculation

Calculated
result y
u(y)
y(x+u(x))
y(x)
u(y) ≈
y(x+u(x)) - y(x)

x x+u(x)

/Eurachem-UK/24

Method Validation Seminar June 06


© 2006 LGC Limited 8
Other effects: A simple
uncertainty experiment
( y2 − y1 )
Measurement Gradient =
result y ( x2 − x1 )
y2

u(y)
y1

x1 x2
u(x) Input parameter xi

/Eurachem-UK/25

“Other effects”

()
uc2 (y ) = sˆR2 + u 2 δˆ + ∑ c i2 u 2 (x i )
=
i k,n

• Combined uncertainty
• Adjusted reproducibility estimate

• Uncertainty associated with method bias

• Effects not covered, e.g. sampling uncertainty

/Eurachem-UK/26

Conclusions

• ISO TS 21748 provides a recipe for using collaborative


study data
• Requires checks for consistency with study performance
• Allows for changes in the test item type

• Often reduces to a simple reproducibility SD

/Eurachem-UK/27

Method Validation Seminar June 06


© 2006 LGC Limited 9

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy