BD Veritor System Bedienungsanleitung Seite 20

CLIA WAIVER STUDY
As part of a larger prospective study, as described in the Performance Characteristics section above, the accuracy of the
BD Veritor System for Rapid Detection of Flu A+B test was evaluated at five CLIA waived testing sites. A total of 31 operators
representative of CLIA waived sites (intended users) participated in the study. No training on the use of the test was provided.
The study included 515 nasal/nasopharyngeal swabs prospectively collected and 80 retrospective archived specimens. The
BD Veritor System results were compared with results obtained by an FDA cleared molecular influenza A and B assay, the
comparator method. Three specimens were excluded due to BD Veritor invalid results. The invalid rate was 0.5% (3/598) with
95% C.I: 0.2% to 1.5%.
The positive percent agreement (PPA) and the negative percent agreement (NPA) between the BD Veritor results and the
comparator method are presented in the tables below (refer to Performance Characteristics section for definition of terms).
Positive Percent Agreement and Negative Percent Agreement of BD Veritor Flu A+B Test with the Comparator Method
Total Number of
Samples
595
Positive Percent Agreement and Negative Percent Agreement of BD Veritor Flu A+B Test with the Comparator Method
Total Number of
Samples
595
Another study was designed to assess the capability of untrained users to test weakly reactive samples and deliver results
with accuracy. The BD Veritor System for Rapid Detection of Flu A+B assay was evaluated at three non-laboratory CLIA
waived sites using panels of simulated swab samples including two weak positives near the assay cutoff and one negative
sample. The positive swab samples were formulated at two levels: a "low positive" sample targeted at the assay limit of
detection; and a "high negative" sample targeted just below the assay limit of detection. The panels included two strains of
Flu A viruses (A/California/7/2009 and A/Victoria 3/75) and one Flu B virus (B/Lee/40). The swab samples were randomized
and masked with respect to their identity. There were two intended users at each of the CLIA waived sites (six operators in
total) and each site tested the panel on each of 10 days. The same panels of simulated swab samples were also tested at
three clinical laboratory sites as controls. The performance of the BD Veritor System with samples near the assay cutoff was
acceptable when used by intended users.
The tables below show performance of the test with samples near the cutoff of the assay for influenza A and influenza B in
the hands of untrained intended users (across all sites).
Sample Type
High Negative A/California/7/2009
H1N1
Low Positive A/California/7/2009
H1N1
High Negative A/Victoria 3/75
H3N2
Low Positive A/Victoria 3/75
H3N2
Negative
*Two (2) samples were excluded from the analysis due to errors in data recording.
PPA
Confidence Interval
82.1% (151/184)
PPA
Confidence Interval
79.7% (102/128)
Influenza A Viral Strains
Percent Detection
INFLUENZA A
95%
(75.9%, 86.9%)
INFLUENZA B
95%
(71.9%, 85.7%)
Untrained Intended Users
6.7% (4/60)
81.7% (49/60)
6.7% (4/60)
80.0% (48/60)
0.0% (0/118)*
20
NPA
Confidence Interval
98.1% (403/411)
NPA
Confidence Interval
99.4% (464/467)
95% Confidence Interval
(2.6%, 15.9%)
(70.1%, 89.4%)
(2.6%, 15.9%)
(68.2%, 88.2%)
(0.0%, 3.2%)
95%
(96.2%, 99.0%)
95%
(98.1%, 99.8%)