Welch Allyn OAE Hearing Screener - User Manual - Manual (Page 37)

Welch Allyn OAE Hearing Screener Directions for Use 705353 Rev F 36
DPOAE Measurement-Based Stopping Rules per Test
The system default DPOAE protocol stops the test when the conditions for an overall test Pass or
Refer occur. For example, if the first 3 frequencies “Pass”, then the final frequency will not be
tested since the overall pass criterion is that 3 out of 4 frequencies must pass. Likewise, if the
first 2 frequencies “Refer” then the final two frequencies will not be tested since there is no chance
for an overall Pass result even if the final two frequencies Pass. This has the advantage of
shortening the test time.
If you would like the system to continue through all of the test frequencies rather than stopping
when conditions for an overall Pass or Refer are achieved, refer to the section entitled
Changing
the Default Protocol to Continue through All Frequencies.
DPOAE Pass or Refer Criteria
For an overall PASS result to display on the LCD at the completion of a DPOAE test when using
the DPOAE system default protocol, three of the four test F2 frequencies must meet the response
conditions defined for a PASS.
The PASS/REFER criteria incorporated into the screener are based on published data by Gorga,
Neely, Ohlrich, Hoover, Redner and Peters (1997)
1
. In Gorga et al’s study, hearing impairment
was defined as a behavioral, pure tone audiometric threshold of 25 dB HL or greater at a
particular frequency. The interested reader is referred to this study for more detailed information.
Screener Sensitivity and Specificity Estimations
If normal hearing at a frequency is defined as having an audiometric threshold of 20 dB HL or better, and
an equivalent hardware is use to record responses and generate stimuli with the same characteristics
described in the article Gorge, et al. 1997 (i.e., high frequency and low frequency stimulus tone intensities
are kept at 55 dB SPL and 65 dB SPL, respectively while the frequency ratio is kept as 1.22 as is the case
with Welch Allyn OAE Hearing Screener protocol), then the device’s clinical performance in terms of its
sensitivity and specificity can be estimated from the cumulative distribution functions published by Dr.
Gorga and his colleagues (Gorga et al., 1997; Dorn et al. 1999). With the use of cumulative distribution
graphs for normal and hearing impaired populations, the sensitivity and specificity for each tested frequency
using the specific protocol parameters can be determined for the DPOAE amplitude and SNR criteria
separately and then the statistical theory can be used to calculate the results for combined DPOAE and SNR
criterion. Furthermore the overall expected test sensitivity and specificity ranges for the representative
sample population (ages 1.3 years to 96 years) can be determined by using probability theory. The overall
sensitivity and specificity values for the screening protocols estimated using the methods described above is
provided in the Table below. Frequency specific sensitivity and specificity values are given in the next
Table.
Overall DPOAE Performance Estimates
Sensitivity Specificity
Highly Correlated 0.89 0.85
Independent 0.992 0.904
Table: Overall Sensitivity and Specificity range estimates of screening protocols that provide an overall
Pass/Refer recommendation. A minimum of 3 individual frequency ‘Pass’ results are required for an overall
‘Pass’. Note that ‘highly correlated’ designation represents the case where individual tested frequencies are
considered to be clinically correlated to each other, whereas ‘independent’ means they are considered
individually and not correlated to each other. Since some degree of correlation is expected in real life, but
the correlation amount is not known, the test sensitivity and specificity is expected to be between the
boundary values represented in the table.