METHOD VERIFICATION
PRELIMINARY STUDY: WITHIN RUN PRECISION
Method verification is a one- time process to determine performance characteristics before a
test system is utilized for patient testing. The experimental confirmation of performance
usually include preliminary study (precision check), linearity, comparison study and
reference range.
Preliminary Study
The preliminary study is applied only to newly installed instruments or new measurement
procedure in the laboratory. This is very important to verify the precision performance of
method in the insert kit is fulfilled before the new method is utilize for patient sample testing.
The basic statistics calculation as below will used to verified the precision of the new
method:
· Mean = ( Σ x)/ n , n= number of observation
· Variance, σ2 = Σ (Xi – Mean )2 / (n-1)
· Standard Deviation, SD= √ σ2
· Coefficient Variation ( CV ) = (SD/mean) x 100.
The main objective of preliminary study is to verify the precision performances in the insert
kit are fulfilled before we use the new method in our laboratory. These include CV within
run precision and CV intermediate or total precision.
We will use control material at 2 or 3 levels as recommended by the manufacturer to asses precision within run and total precision for verification of manufacturer’s claim. We will
accept the manufacturer precision within run and total precision if precision experiment
within run and total precision is less than manufacturer claim.
The experiment procedures for the preliminary study is to run triplicate per day within two
hours for five days using control material as recommended by the manufacturer. We
calculate the daily mean and variance (σ2) analytical. From daily mean and variance
analytical, we calculate overall mean, standard deviation (SD) and coefficient variation (CV) analytical for within runs, between days and total CV. Next, we compare CV analytical withinrun and total CV with manufacturer claims. If CV analytical within run and total CV analytical is less than CV within run and total precision manufacturer claim, we will accept the
precision manufacturer claims. If more, we will continue to calculate SD verification (SDV)
from CV manufacturer, chai- distribution and degree of freedom. We only accept if SD
analytical less than SDV.
Preliminary Study: Within Run
The procedure of the preliminary study is run all the quality control (QC ) material as
recommended by manufacturer triplicate within two hour for five day and record all the data.
Daily mean and variance (σ2) were calculate.
Example : Albumin, the total number of level QC are two.
The calculation for this preliminary study, only for level 2 QC. The mean level 2 is 38 g/l, CV within run is 0.5% and Total CV is 1.1%.In real situation, we must perform precision study
for level 1 and level 2. The data triplicate for five days with mean and variance as below:
Day
|
Run 1
|
Run 2
|
Run 3
|
Mean
|
Variance
|
1
|
37.7
|
37.5
|
37.8
|
37.73
|
0.0633
|
2
|
37.3
|
37.5
|
37.7
|
37.50
|
0.0400
|
3
|
38.4
|
37.6
|
38.1
|
38.03
|
0.1633
|
4
|
38
|
37.3
|
38.2
|
37.83
|
0.2233
|
5
|
38.1
|
37.8
|
38
|
37.97
|
0.0233
|
Note: formula for mean and variance
Mean = ( Σ x)/ n , n= number of observation or number of run
Variance, σ2 = Σ (Xi – Mean )2 / (n-1)
Overall mean and variance
Overall mean = (37.73 + 37.50+38.03+37.83+37.97)/5
= 37.81 g/l
Overall variance = (0.0633+0.0400+0.1633+0.2233+0.0233)/5
Standard deviation (SD) and coefficient variation CV)
SD = √variance
= √0.1027

CV =(SD/overall mean) x 100
=( 0.3204/37.81) x 100
= 0.8474% CV analytical
CV manufacturer is 0.5% and CV analytical is 0.8474%, That means analytical CV is more
than manufacturer CV . Next step is we calculate the SD verification (SDV)
Standard deviation verification (SDV)
SDV = {(CV manufacturer/100 x overall mean) x √C} / √V
C= Percentage point ( Ӽ2 (0.05/L, V), L= number of QC level (L=2)
V= Degree of freedom ( L*n), L= number of QC level, n = number of day (n=5)
SDV = {(1.1/100 x 37.8) x √ 20.48} /√ 10
= 0.12706
SD analytical is 0.3204, therefore the SD analytical > SDV means the precision for within
run manufacturer is not acceptable. The manufacturer must perform troubleshooting and
preliminary study for within run and re-evaluate.
Note: The value C is obtain from chi-square table or from window excel
( Ӽ2 (0.025,10).
No comments:
Post a Comment