characteristic, however, is that their cause can be identified and corrected. There are
four common means of identifying this type of error:- Use of a ‘blank’ sample: This is a sample that you know contains none of the analyte
under test so that if the method gives a non-zero answer then it must be responding in
some unintended way. The use of blank samples is difficult in cases where the matrix
of the test sample is complex, for example, serum. - Use of a standard reference sample: This is a sample of the test analyte of known
composition so the method under evaluation must reproduce the known answer. - Use of an alternative method: If the test and alternative methods give different results
for a given test sample then at least one of the methods must have an inbuilt flaw. - Use of an external quality assessment sample: This is a standard reference sample that
is analysed by other investigators based in different laboratories employing the same
or different methods. Their results are compared and any differences in excess of
random errors (see below) identify the systematic error for each analyst. The use of
external quality assessment schemes is standard practice in clinical biochemistry
laboratories (see Section 16.2.3).
Random error (also called indeterminate error)
Random errors are caused by unpredictable and often uncontrollable inaccuracies in
the various manipulations involved in the method. Such errors may be variably
positive or negative and are caused by such factors as difficulty in the process of
sampling, random electrical ‘noise’ in an instrument or by the analyst being inconsist-
ent in the operation of the instrument or in recording readings from it.Standard operating procedures
The minimisation of both systematic and random errors is essential in cases where the
analytical data are used as the basis for a crucial diagnostic or prognostic decision as
is common, for example, in routine clinical biochemical investigations and in the
development of new drugs. In such cases it is normal for the analyses to be conducted
in accordance withstandard operating procedures(SOPs) that define in full detail the
quality of the reagents, the preparation of standard solutions, the calibration of
instruments and the methodology of the actual analytical procedure which must be
followed.1.4.2 Assessment of the performance of an analytical method
All analytical methods can be characterised by a number ofperformance indicators
that define how the selected method performs under specified conditions. Knowl-
edge of these performance indicators allows the analyst to decide whether or not
the method is acceptable for the particular application. The major performance
indicators are:- Precision(also calledimprecisionandvariability): This is a measure of the
reproducibility of a particular set of analytical measurements on the same sample
18 Basic principles