153 / 350
February 20, 2013
19
Statistical Analysis
1
For the method comparison, results obtained from samples analyzed by the 3M Petrifilm SALX System
2
were comparable to those analyzed by the FDA/BAM Chapter 5, USDA/FSIS-MLG 4.05 and ISO 6579
3
reference methods. A Mantel-Haenszel chi-square analysis (for unmatched test portions) was used to
4
compare the 3MPetrifilm SALX System samples and the reference method samples.
5
6
For the low inoculum level and the un-inoculated control level, the probability of detection (POD) was
7
calculated as the number of positive outcomes divided by the total number of trials. The POD was
8
calculated for the candidate presumptive results, POD
CP,
the candidate confirmatory results, POD
CC
, the
9
difference in the candidate presumptive and confirmatory results, dPOD
CP,
the reference method,
10
POD
R
, and the difference in the confirmed candidate and reference methods, dPOD
C
. POD analyses
11
were conducted for each time point and of the 3M Petrifilm SALX System and the reference methods.
12
13
Based on the results obtained from the MPN analyses, the target levels of 0.2-2 CFU/ test portion for
14
the low inoculum level were achieved for each matrix tested. In addition, the requirement of
15
obtaining fractionally positive results (5-15 positives out of 20 replicates) was achieved by either the
16
candidate or reference methods or both for each of the matrices analyzed in this study.
17
18
19
RESULTS AND DISCUSSION
20
High Microbial Foods
21
Raw Ground Chicken
22
For raw ground chicken test portions, there were 15 confirmed positives by the 3M Petrifilm SALX System
23
at all four time points (18 hours and 24 hours of primary enrichment transferred into RV[R10] secondary
24
enrichment and streaked at 8 and 24 hours) and 14 confirmed positives by the USDA/FSIS-MLG method.
25
All un-inoculated control samples were negative by both methods. A χ
2
value of 0.12 was obtained
26
between the reference methods and the candidate method when using the alternative confirmation
27
indicating no significant difference between the candidate and reference method at all four time points. A
28
χ
2
value of 0.0 was obtained between the reference method and the candidate method when using the
29
traditional confirmation, indicating no significant difference between the candidate and reference method
30
at any of the time points.
31
32
For the low level inoculum, a dPOD
C
value of 0.05 with 95% confidence intervals of (-0.22, 0.31) was
33
obtained between the USDA/FSIS-MLG reference method and the candidate method samples confirmed
34
using alternative confirmation at all four time points. A dPOD
C
value of 0.00 with 95% confidence intervals
35
of (-0.26, 0.26) was obtained between the USDA/FSIS-MLG reference method and the candidate method
36
samples confirmed using traditional confirmation at all four time points. The confidence intervals obtained
37
for the dPOD
C
values indicated no significant difference between the two methods using either the
38
traditional or alternative confirmatory procedure. A dPOD
CP
value of 0.00 with 95% confidence intervals of
39
(-0.26, 0.26) were obtained between presumptive and confirmed results using either the traditional or
40
alternative confirmatory procedure at all four time points. The confidence intervals obtained for dPOD
CP
41
indicated no significant difference between the presumptive and confirmed results.
42
43
Frozen Uncooked Shrimp
44
For frozen uncooked shrimp test portions, there were 12 confirmed positives by the 3M Petrifilm SALX
45
System at all four time points and 10 confirmed positives by the FDA/BAM method. All un-inoculated
46
control samples were negative for both methods. A χ
2
value of 0.39 was obtained between the reference
47