100 / 596
4
1
Test Portion Analysis
2
3
Collaborators were instructed to follow the appropriate preparation and analysis as outlined in
4
the study protocol for each matrix for both the 3M MDA 2 -
Listeria
method and reference
5
method. For both matrices, each collaborator received 72 test portions (12 high, 12 low and 12
6
un-inoculated controls for each method to be performed). For the analysis of the deli turkey test
7
portions by the 3M MDA 2 -
Listeria
method, a 125 g portion was enriched with 975 mL of
8
Demi-Fraser (DF) broth, homogenized for 2 minutes and incubated for 24-28 hours at 37 ±1
o
C.
9
For the raw chicken breast fillettest portions analyzed by the 3M MDA 2 -
Listeria
method, a 25
10
g portion was enriched with 475 mL of DF, homogenized for 2 minutes and incubated for 28-32
11
hours at 37 ±1
o
C.
12
Following enrichment, samples were assayed by the 3M MDA 2 -
Listeria
method and,
13
regardless of presumptive result, confirmed following the USDA/FSIS MLG 8.09 reference
14
method. Both matrices evaluated by the 3M MDA 2 -
Listeria
method were compared to
15
samples analyzed using the USDA/FSIS MLG 8.09 reference method in an unpaired study
16
design. All positive test portions were biochemically confirmed by the API
Listeria
biochemical
17
test or by the VITEK 2 GPbiochemical identification test, AOAC Official Method 2012.02 [8].
18
19
Statistical Analysis
20
21
Each collaborating laboratory recorded results for the reference method and the 3M MDA 2 -
22
Listeria
method on the data sheets provided. The data sheets were submitted to the study director
23
at the end of each week of testing for statistical analysis. Data for each matrix was analyzed
24
using the probability of detection (POD)statistical model [9]. The probability of detection (POD)
25
was calculated as the number of positive outcomes divided by the total number of trials. The
26
POD was calculated for the candidate presumptive results, POD
CP,
the candidate confirmatory
27
results (including false negative results), POD
CC
, the difference in the candidate presumptive and
28
confirmatory results, dLPOD
CP,
presumptive candidate results that confirmed positive (excluding
29
false negative results), POD
C,
the reference method, POD
R
, and the difference in the confirmed
30
candidate and reference methods, dLPOD
C
. A dLPOD
C
confidence interval not containing the
31
point zero would indicate a statistically significant difference between the 3M MDA 2 -
32
Listeria
and the reference methods at the 5 % probability level. In addition to POD, the
33
repeatability standard deviation (s
r
), the among laboratory repeatability standard deviation (s
L
),
34
the reproducibility standard deviation (s
R
)and the P
T
value were calculated. The s
r
provides the
35
variance of data within one laboratory, the s
L
provides the difference in standard deviation
36
between laboratories and the s
R
provides the variance in data between different laboratories. The
37
P
T
value provides information on the homogeneity test of laboratory PODs [10].
38
39
40
41
42
43
44
45
46
47
48
49
AOAC Research Institute
Expert Review Panel Use Only
OMAMAN-29 A/ Collaboartive Study Manuscript
OMA ERP June 2016
ERP Use Only