Nothing exists until it is measured (Niels Bohr)
To measure is to know, or “nothing exists until it is measured“ as Niels Bohr
said. Producers, consumers and regulatory bodies alike impose ever stricter
quality and consistency requirements on products and processes (e.g. purity,
failure frequency and yield). The analytical measurement methods used to
determine these characteristics have to be valid and of known accuracy.
The Mathematics & Statistics group can assist you in assessing the validity of
an analytical method and its conformity with accuracy requirements. Our
resources are many years of industrial experience, state-of-the art software
and statistical expertise of the highest academic level. We are familiar with
international standards and laws.
Objective of a validation study
The objective of a validation
study is to assess the accuracy of an analytical method, to demonstrate that
the method is suitable for its intended purpose of use and that it satisfies
the relevant requirements.
Elements of a validation study
A validation study might contain
one or more of the following elements:
-
Accuracy: determine the trueness and precision of the test method for a range
of analyte concentrations.
-
Repeatability: for multiple copies of a single sample, determine the spread of
the test results within a single laboratory.
-
Reproducibility: for multiple copies of a single sample, determine the spread
of the test results between laboratories at different production sites.
-
Comparison of two competing methods: which of the two is more appropriate?
-
Linearity: determine the relationship between the analyte concentration and
the value of the response variable.
-
Selectivity: what happens in the presence of other components nearly identical
to the analyse?
-
Robustness: determine interfering factors and sources of error which affect
the accuracy of the method.
-
Quantification limits: within what limits does your method give reliable
measurement results?
-
Detection limit: is your method capable of proving the absence of a certain
substance?
-
Calibration and use of standard reference samples: determine the accuracy of a
calibration curve and future test results.
-
SPC: statistical process control to ensure the proper performance of the
method over time.
