178
10 Interlaboratory Comparisons
10.9.1 The Assigned Value from Model Equation
In some situations, the value of the standard deviation used for the evaluation of
the performance of a laboratory can be determined on the basis of a general model,
for example, Horwitz function. So, it is possible to predict the standard deviation
depending on the concentration of the determined analyte. The main limitation of this
method is that the function does not always sufficiently reflect the real reproducibility
of the measuring procedure achieved practically by the laboratory.
10.9.2 The Assigned Value from PT Round
In this case, the measurement results from all participants of a given round are used.
The standard deviation of the mean value is calculated with the use of robust statistics,
in which the influence of the extreme values on the mean value is limited.
When z-score is used, following evaluation criteria are used:
|z| ≤ 2.0 satisfactory performance
2.0 < |z| < 3.0 questionable performance
|z| ≥ 3.0 unsatisfactory performance.
Another way to measure the performance of the laboratory is to use the z
-score,
which is similar to z-score, but with an extended denominator. In the equation describing the z-score, the uncertainty of the reference value is not included. Thus, z
-score
has been proposed—one that includes both the target range (as included in the equation for z-score) and the standard uncertainty of the reference value (u X ).
z
X lab − X ref /
√ σ
2 + u
2
X
(10.5)
where u X stands for the standard uncertainty for the reference value.
The criteria for the evaluation of results with the use of the z
-score is the same
as in the case of the z-score.
When z
-score is used, the following evaluation criteria are used:
|z
| ≤ 2.0 satisfactory performance
2.0 < |z
| < 3.0 questionable performance
|z
| ≥ 3.0 unsatisfactory performance.
NOTE! The value of the z
-score will significantly differ from the value of
z-score if the uncertainty value of the laboratory is significant.
10 Interlaboratory Comparisons
10.9.1 The Assigned Value from Model Equation
In some situations, the value of the standard deviation used for the evaluation of
the performance of a laboratory can be determined on the basis of a general model,
for example, Horwitz function. So, it is possible to predict the standard deviation
depending on the concentration of the determined analyte. The main limitation of this
method is that the function does not always sufficiently reflect the real reproducibility
of the measuring procedure achieved practically by the laboratory.
10.9.2 The Assigned Value from PT Round
In this case, the measurement results from all participants of a given round are used.
The standard deviation of the mean value is calculated with the use of robust statistics,
in which the influence of the extreme values on the mean value is limited.
When z-score is used, following evaluation criteria are used:
|z| ≤ 2.0 satisfactory performance
2.0 < |z| < 3.0 questionable performance
|z| ≥ 3.0 unsatisfactory performance.
Another way to measure the performance of the laboratory is to use the z
-score,
which is similar to z-score, but with an extended denominator. In the equation describing the z-score, the uncertainty of the reference value is not included. Thus, z
-score
has been proposed—one that includes both the target range (as included in the equation for z-score) and the standard uncertainty of the reference value (u X ).
z
X lab − X ref /
√ σ
2 + u
2
X
(10.5)
where u X stands for the standard uncertainty for the reference value.
The criteria for the evaluation of results with the use of the z
-score is the same
as in the case of the z-score.
When z
-score is used, the following evaluation criteria are used:
|z
| ≤ 2.0 satisfactory performance
2.0 < |z
| < 3.0 questionable performance
|z
| ≥ 3.0 unsatisfactory performance.
NOTE! The value of the z
-score will significantly differ from the value of
z-score if the uncertainty value of the laboratory is significant.