 |
|
Are they clinically accurate?
Introduction
Although current and future miniature audiometric test devices are promising, these devices represent a departure from standard audiometric procedures. Therefore, their accuracy needs to be studied.
This accuracy, compared to that of standard audiometers, was assessed in a clinical trial of 76 subjects. Results showed good sensitivity and specificity even though screenings were performed outside a sound booth environment and by a less experienced tester. This promises well for the further development of miniaturized audiometric devices.
One of the hoped-for benefits of smaller sized audiometers is that their portability will enable health professionals to make greater use of them, resulting in increased referrals for hearing aids (1,9). Manufacturers are currently offering audiometers and other test devices in increasingly smaller formats: small (1) desk-top audiometers (8,10), handheld screeners (9), and hand-held audiometers (6). Manufacturers say they have plans to offer an even wider range of tests in small or hand-held formats (7).
In order for these benefits of size to be realized, however, the accuracy of the device must not be compromised. In addition, since the use of miniaturized devices often represents a departure from standard testing methods, it is important to study these departures: use by non-traditional testers who are less well trained, use outside typical sound booths and closer proximity to the client, resulting in possible tester cues.
This article addresses these issues by means of a clinical trial of the Oto-Screen 1, a hand-held audiometer manufactured by Handtronix, Inc., in which an earphone is handheld to the ear, as opposed to the standard headset.
Testing was performed under conditions similar to those expected to occur in a typical physician's office, i.e., in a nonsound treated room and with a tester having minimal experience. It is felt these data are pertinent both to the present device and the development of future miniaturized test equipment.
The Handtronix OtoScreener device is a semi-automatic pure-tone screening device with specified characteristics (2, 3, 4). The device was first developed with the Utah Innovation Center, Salt Lake City, UT, in conjunction with Campbell Scientific, Logan, UT, in 1988. Many upgrades and improvement have been incorporated over the years. It was hoped that the small size and relatively low cost would encourage increased usage by physicians and other health care providers. Two separate buttons utilizing pressure-sensitive "silent" switches operate the device.
The "Scan" button selects one of four frequencies (500, 1000, 2000 or 4000 Hz), then presents the tone after a 1.0 second delay to avoid tester cues. Repressing "Scan" moves the device to the next frequency, whose tone is then automatically presented after the delay. The device is offered in two versions, 20 - 40 dB HL (Pediatric) or 25 - 40 dB HL (Standard). Other intensity combinations are also available.
The "Boost" button, which presents the stimulus at 40 dB HL on either model, is used primarily as a conditioning tone. The device is calibrated on an NBS 9A coupler using reference equivalent sound pressure levels developed for the device according to ANSI 1969.3 The device utilizes a 9 volt battery with a "low- battery" light indicating when current is insufficient for proper function.
Clinical Trial
Seventy-six subjects (38 male and 38 female) were tested both with standard calibrated audiometric equipment (Grason-Stadler Model 16 or Madsen Model OB 822) and with the OtoScreener. Subjects ages ranged from 4 to 87 years (mean = 43 years). Testing order was counterbalanced between the audiometer and the screener to avoid order effects.
Testing was performed at a large otolaryngology-audiology center in Salt Lake City by a certified audiologist. Screening with the Oto-Screen was performed with the 25 dB version by a female audiology student with minimal testing experience in an adjacent nonsound treated room simulating the environment in which such testing may be performed. Background noise levels were tested in one-third octave bands and compared to allowable levels according to ANSI S 3.1 1991.5.
Measurements were taken under conditions of maximum typical noise and found to be adequate at 1000, 2000 and 4000 Hz, but just above allowable levels at 500 Hz. Screenings were nevertheless also performed and data collected at 500 Hz. Screenings and audiograms were performed by the testers without knowledge of the other's results.
Results
Table 1 shows sensitivity (percentage of true positives) and specificity (percentage of true negatives) for the Oto-Screen1. For data analysis, an audiometric threshold of 25 dB was considered a "pass." Though the screener did not obtain a true 50% threshold, testers felt that including 25 dB audiometric thresholds as "pass" results would yield more conservative sensitivity and specificity values than would be the case if audiometric thresholds of 25 dB were excluded. Table 1 also shows the percentage of false positives (those who failed the screener but passed the audiogram) and false negatives (those who passed the screener but failed the audiogram).
Table 1. Sensitivity/specificity (%) of OtoScreenI compared to standard audiometry (n=76) |
||||
Frequency (Hz) |
500 |
1000 |
2000 |
4000 |
Sensitivity |
97.9 |
94.9 |
98.4 |
98.7 |
Specificity |
77.9 |
98.9 |
94.4 |
97.3 |
False Positives |
22.1 |
1.1 |
5.6 |
2.7 |
False Negatives |
2.1 |
5.1 |
1.6 |
1.3 |
Conclusions
According to our results, sensitivity and specificity values are high for the handheld audiometer when compared to standard audiometry. The somewhat lower values at 500 Hz are possibly due to the slightly higher room noise at that frequency as could be typically expected in such an environment.
The use of a less experienced tester, probably less aware of the need to avoid extraneous noises, may also have affected the data. Use of the inexperienced tester, however, in a non-sound treated environment was intended to simulate realistic conditions in which this or similar devices may be used.
Apparently the closer proximity of the tester to the patient was not a significant factor for this device, since agreement with standard audiometrics was high. It should be noted the OtoScreen had been developed with features aimed at avoiding tester cues such as silent buttons whose axis of motion is at a right angle to the patient's head. The built-in delay prior to test tone presentation was also designed to avoid the sound of button noise being mistaken for the tone.
Though these results cannot be generalized to all hand-held devices, it appears the hand-held format has potential for accuracy in current and future applications.
The OtoScreenI provides two intensities (typically 25 and 40 dB HL) at standard screening frequencies (500, 1000, 2000 and 4000 Hz.
References
1. Alvord LS: Toward miniature audiometers. Hear J 42(6) 22-23, 1989.
2. Alvord LS: Manual for Oto-Screen I and Guide to
Hearing Loss. Handtronix, Inc. Salt Lake City, UT 1989.
3. American National Standards Institute: ANSI S 3.6 - 1969. Specification
for Audiometers. New York 1969.
4. American National Standards Institute: ANSI S 3.6 - 1989. Specification for
Audiometers (R S 3.6 - 1969). New York, 1989.
5. American National Standards Institute: ANSI S 3.1 - 1991. Maximum
Permissible Ambient Noise Levels for Audiometric Test Rooms.New York, 1991.
6. Bienvenue GR, Michael PL, Chaflinch JC and Zeigler J: The Audioscope:
A clinical tool for otoscopic and audiometric examination. Ear and Hear
6(5) 251 - 254,1985.
7. Lyon P: Vice-President, Handtronix Inc. Personal communication
December 30, 1992. Handtronix, Inc. Salt Lake City, UT 1992.
8. Monitor Instruments: Advertisement of Model 20 P Portable,
Hear Instrum42(9):26, 1991.
9. Sullivan RF: A pocket-portable, single frequency screening device. Hear
Instrum43(4):39 - 40,1992.
10. Tremetrics: Manufacturer's News, Model RA 500, Hear Instrum43(12):28,
1992.
Reprinted from HEARING INSTRUMENTS, Volume 44, Number 6, 1993
AN ADVANSTAR PUBLICATION Printed in U.S.A.