691 / 1195
VIDAS LPT - Summary 2012
EUROFINS IPL Nord SAS
–
1 rue du Professeur Calmette
–
59046 LILLE cedex
p 11/14
The relative accuracy values for the alternative method with respect to the reference method were calculated for
each of the levels and are given in the table below.
Level
AC%
LCL* %
L0
100,00
100,0
L1
75,00
66,8
L2
94,64
90,4
L1+L2
84,82
80,0
Total
89,88
86,6
* LCL : low critical value, defined by the EN ISO 16140 standard
3.4.3 Study of discrepant results
According to annex F of the EN ISO 16140 standard, the number of discrepant results from which a statistical test
needs to be performed to compare the two methods is 6.
Thirty four discrepancies being observed, a statistical test needs to be performed.
If the number of discrepant results is greater than 22, it is necessary to use the McNemar test with
²
distribution
for one degree of freedom. It consists of determining d = PD
–
ND and comparing d to a minimum value of d
defined for each number of discrepant results.
Number of discrepant results
d minus
d
Conclusion
34
12
12
–
22 = 10
Equivalency
The VIDAS LPT method can be considered to be equivalent to the ISO 11290-1 reference method (including
amendment A1) for the detection of
Listeria
in human food products and environmental samples.
3.5 Interpretation
3.5.1 Comparison of relative accuracy (AC), specificity (SP) and sensitivity (SE) values
The values obtained in both parts of the validation study are given below:
Study Value
Inter-laboratory study
Preliminary study
Exactitude relative (AC)
89,9 %
92,0 %
Sensibilité (SE)
93,0 %
89,7 %
Spécificité (SP)
86,6 %
93,8 %
AFNOR Certification requests that the sensitivity of both methods should be recalculated accounting for all
confirmed positives (this includes additional positives from the alternative method):
Alternative method
Reference method
(PA + PD) / (PA + PD + ND) =
93,8 %
(PA + ND) / (PA + PD + ND) =
88,7 %
3.5.2 Accordance (DA)
Accordance is the percentage of likelihood of obtaining the same result for two identical test samples analyzed in
the same laboratory under repeatability conditions, i.e. a single operator using the same apparatus and the same
reagents within the shortest possible time interval.
To calculate the accordance, it is necessary to calculate the probability of two identical samples giving the same
result, for each of the participating laboratories, and subsequently determine the mean of the probabilities of all the
laboratories.