Abstract
Objective
The purpose of this study was to develop a cost-effective and easy to use audiometer to monitor progressive hearing change in school-aged children.
Design
The hardware of the audiometer developed included a computer, an external sound blaster and a headphone The hearing screening software was developed to control the pure tone sound level with modulability in the range of 0-45 dB HL at 1, 2, 4 and 0.5 kHz frequencies. Thirty sixth-grade children, aged 12-13 years old, were randomly divided into two groups for a hearing test in a conference room. Testing for one group was performed by a person with experience and the other group was examined by a graduate student who was not familiar with the device. After the hearing test, all children were immediately screened using a clinical diagnostic audiometer in a soundproof room by an audiologist.
Results
Most of the threshold dB values obtained by the audiometer in a conference room (55 ears, 93.2%) were significantly greater than those obtained by the clinical audiometer in a soundproof room. Most of the differences between these two measurements were within 5 dB (94.9%). Only 5.1% had a discrepancy within the maximum range of 10 dB. The correlation and intraclass correlation coefficients between the two measurements were 0.861 and 0.929, respectively. The results also indicated that the experience level of the screening personnel did not affect the testing.
Conclusion
The newly developed audiometer is quite cost-effective and can be easily operated. The threshold dB values obtained by the audiometer developed in a conference room were comparable with the results of a clinical audiometer in a soundproof room. The audiometer developed could measure the hearing threshold values and, therefore, be useful in monitoring progressive hearing change in school-aged children.
Introduction
Hearing impairment is difficult to monitor due to the limited availability of test equipment and trained specialists in many developing countries. 1 Even in the USA, paediatricians were reported not to recheck or refer more than half of the children who failed hearing screening. 6 It is therefore necessary to develop an audiometer that parents or school nurses or staff can use routinely to monitor the hearing level of school-aged children. The aim of this research was to develop an audiometer that is cost-effective and need not be operated by trained staff, through which progressive hearing change in school-aged children can be monitored.
Materials and Methods
Hardware
The three major hardware components of the audiometer included a computer (S6F, ASUS notebook, Taipei, Taiwan), an external sound blaster (iVoiceII, Macally Inc, Ontario, CA, USA) and a headphone (ATH-T3, Audio-technica Co., Shinjuku-ku, Tokyo, Japan). The audiometer was configured as shown in Figure 1.
Hardware components of the developed audiometer
A desktop computer might also be used. A notebook computer was utilized here for portability. The computer was equipped with the Microsoft Windows operating system, and a USB interface port was required to connect the external sound blaster.
The iVoiceII, an egg-shaped USB interface sound blaster, was utilized as an external sound blaster to generate the pure tones. The sound blaster had a line-out jack to accept headphones and a mute button to be used as a feedback control. When the mute button was pushed, the software would detect the on-off signal and recognize whether or not the test subject heard the auditing sound.
The headphones used were of the closed-air dynamic type, with 20-22,000 Hz frequency response and a 108 dB/mW output sound level. The 3.5 mm stereo connector of the headphones could be fitted into the output port of the iVoiceII.
Software
The graphical programming tool LabVIEW 7.1 (National Instruments, Austin, TX, USA) was utilized to develop the hearing auditing program. Referring to the ASHA guidelines, 7 the program was developed to control the sound level of pure tone at 1, 2, 4 and 0.5 kHz frequencies. Furthermore, to simplify the hearing auditing procedure, the program had been modular in the range from 0 to 45 dB HL for school-aged children. In order to run the hearing auditing program, the free download run-time engine software (National Instruments, Austin, TX, USA) was installed.
As indicated in Figure 2, the auditing program was designed to make the audiometer work semi-automatically. First, the test ear (right or left) was selected. When the start button for each screening sound level was pressed, the program generated a pure tone at the dB value in a sequence of 1, 2, 4 and 0.5 kHz frequencies. If the subject being audited could hear the tone, she/he would push the mute button on the top of the sound blaster. Then, the feedback signals could trigger the green lights shown on the screen. To pass the auditing dB value, the subject should hear all testing tones at the four frequencies.
The interface of the hearing screening program
For school-aged children, the module was started at 50 dB HL for a trial. Then, a 20 dB sound level was tested as a baseline. If the subject heard tones at all frequencies, the 15, 10, 5 and 0 dB sound levels were tested in sequence to determine the hearing level of subject. If the subject failed at the 20 dB level, she/he would be re-examined again at 20 dB. The subject could possibly have a 20 dB hearing level when she/he succeeded on the retest. If the subject could not hear the 20 dB again, larger sound levels of 25, 30, 35, 40 and 45 dB were tested in sequence. Finally, the test result was recorded and stratified 8 as 0-10 dB ‘excellent’, 11-20 dB ‘good’, 21-25 dB ‘fair’, 26-35 dB ‘alert’ and more than 35 dB ‘referral’.
The sustained time of each pure tone was preset to 1.5 seconds. The time interval between each frequency was 1 second. In case it was needed, the sustained time and time interval could be adjusted on the right top of the program. It should be stressed that the pure tones at low frequencies, such as 500 Hz, are easily affected by the ambient noise. 9 Therefore, another 5 dB was added at the 500 Hz frequency for each screening, as suggested by Weber. 9
Calibration
The digital-stimulus waveform generated by the software developed was used to calibrate the audiometer at the beginning of set-up. According to the ANSI S3.6-2004 standard, the headphones were coupled to a flat-plate and connected to a sound level meter (Model 1900, QUEST Technologies, Oconomowoc, WI, USA). Subsequently, the calibration parameters for the set of external sound blaster and headphone were bundled into the software. The variation after calibration was within 1 dB SPL, and different notebooks or personal computers would not significantly affect the variation. The set-up of the system is the main cost of the calibration.
Confirmation
Three volunteers were tested on the pure tone threshold by using the pure tone audiometry system available in Taipei Veterans General Hospital (clinical audiometer: GSI-61, Viasys Healthcare, Madison, WI, USA; earphone: TDH-50P, Telephonics Co., Farmingdale, NY, USA). In the same soundproofed room, the first author, an otologist, also utilized the audiometer developed to test hearing threshold for the volunteers. The test result indicated that these two audiometers detected the same pure tone hearing threshold for all three volunteers.
Participants
Thirty sixth-grade children, 17 girls and 13 boys aged 12-13 years from the same class in Taipei Tainmu Elementary School participated in the study. The trial was approved by the Institutional Review Board of Taipei Veterans General Hospital. Informed consent was obtained and forms were signed by the pupils’ parents.
Before the hearing screening, the students were taught how to wear the headphones and to push the button when hearing the tone. The children were randomly divided into two groups. One group was tested with the audiometer operated by an otologist who was familiar with system. The other group was examined with the same audiometer but by a graduate student who had learned how to operate the system only two hours before the test.
The test was performed in the conference room of the Tien-Mu Elementary School. The environmental noise was assessed with a sound level meter (2260 Investigator, Brüel & Kjær, Nærum, Denmark) every 20 minutes. The noise level detected in the room was acceptable based on the ANSI S3.1-1999 standard. The standard adjusts the maximum permissible ambient noise levels (MPANLs) for audiometric test room if the minimum hearing level to be tested is higher or lower than 0 dB HL. From a statistical viewpoint, if the attenuation supplied by an earphone is less than the mean value for an individual normal-hearing listener, the listener might experience a threshold shift greater than 2 dB as testing was done in a room having ambient noise levels equal to the values specified in MPANLs. 10
After the hearing test, the children were immediately transported to Taipei Veterans General Hospital. An audiologist utilized a clinical diagnostic audiometer (GSI-61, Viasys Healthcare, Madison, WI, USA) to test the pupils again in sound proof room.
Statistical methods
The intraclass correlation coefficient was calculated to estimate the average correlation coefficient across both methods. A two-sample t-test was utilized to estimate the difference between the means of two independent samples. P values less than 0.05 were considered to be statistically significant.
Results
Thirty students underwent hearing threshold screening. One child with right ear hearing impairment did not respond to the 50 dB trial test of the developed audiometer. The clinical audiometer also confirmed that the threshold of the ear was 70 dB. Therefore, one dataset was missing for the screening utilized by the audiometer.
Most of the threshold dB values obtained by the audiometer in a conference room (55 ears, 93.2%) were significantly greater than those obtained by the clinical audiometer in a soundproofed room. The difference was 2.20 ± 3.50 dB (mean ± standard deviation). A 10 dB difference was only shown in three ears (5.1%). The other 56 ears (94.9%) had a difference within 5 dB. Figure 3 showed the frequency percentage of the threshold difference between these two audiometers.
The distribution of threshold difference observed between the two audiometers
The non-parametric correlation between the two measurements by utilizing developed audiometer in a quiet conference room and the clinical audiometer in a soundproofed room was 0.861, which was significant at the 0.01 level. The intraclass correlation coefficient of these two measurements was 0.929 (95% confidence interval 0.881-0.958).
The mean and standard deviation for the group of pupils tested with the audiometer operated by an otologist were 7.04 ± 8.58 dB. The values for other group, performed with the same audiometer but by the newly trained graduate student were 8.00 ± 7.50 dB. The difference between the means of the measurements by different persons, familiar with or new to the developed audiometer, was not statistically significant.
For the clinical audiometer in a soundproofed room, the mean and standard deviations for measurements by a skilled person versus that of the new trainee were 1.67 ± 3.50 and 2.83 ± 3.39 dB, respectively. However, the difference was not statistically significant.
Discussion
Hearing impairment is one of the most common disorders for school-aged children,1, 2 and it has been suggested 11 that regular screening of children of school-entry age will ensure that these children begin their school life without hearing impairment. Mass hearing screening in a school system is thought to be useful for detecting hearing losses; 12 however, hearing impairment is a low priority for health systems in the developing world. 13 Even in Taiwan, hearing screening in school-aged children is not routinely performed every year. Hearing screening technology is very costly, and the materials and human resources are limited, 13 so a cost-effective audiometer that does not need to be operated by trained staff would be valuable.
The audiometer developed in this study was calibrated and confirmed to be comparable with a clinical audiometer in a soundproofed room. The audiometer can be calibrated in less than 10 minutes by a skilled technician, and once the calibration parameters were bundled into the software, the audiometer required no further calibration. The calibrations were performed by an electrical laboratory for 5 US dollars per system. These costs are expected to be lower if the audiometer is produced on a large scale.
The concept of the design was quite cost-effective. For example, the two major parts, the external sound blaster and headphones, cost less than 100 US dollars. Therefore, the audiometer would be affordable to most elementary schools, or even to parents for routine hearing check-ups.
The test results in this study indicated that most (94.9%) of the differences between threshold dB values obtained by the developed audiometer and by the clinical audiometer were within 5 dB. Only 5.1% had a 10 dB maximum discrepancy, demonstrating that the results using the newly developed audiometer in a quiet room were comparable with those using a clinical audiometer in a soundproofed booth.
As expected, the hearing threshold screened in a soundproofed room was lower than that obtained in a quiet room. 9 The study also demonstrated that the threshold dB values obtained by the audiometer in a conference room were significantly greater than those obtained by a clinical audiometer in a soundproofed room. Only four ears had a lower dB threshold in the conference room, but the difference was within 5 dB, which was in the acceptable range of measurement error.
The hearing auditing program with graphical interface was conducted without difficulty and without encountering any compliance difficulties. The program was acceptable to the students, and the experience of the screening personnel did not affect the testing. Using such a program, trained audiologists would not be required for mass hearing screening. 1
Environmental noise level is one of the most common concerns for hearing screening.9, 14 However, most schools or families do not have a sound level meter to measure the ambient noise level in the area used for screening. To substitute the sound level meter, a biological noise level has been proposed. 9 If a subject with known normal hearing can establish a threshold at 10 dB below the screening level at all frequencies, the area can be utilized for screening. This study also demonstrated that the result of using the newly developed audiometer in a conference room of a school was comparable with the clinical audiometer in a soundproofed booth.
Other than cost, one of the most important features of the newly developed audiometer was that it would test the threshold value of the school-aged children after screening. Currently almost all screening audiometers only have pass or fail results. Using the new audiometer the threshold value for each child could be recorded and monitored for progressive hearing change. Testing results were recorded and stratified 8 as 0-10 dB ‘excellent’, 11-20 dB ‘good’, 21-25 dB ‘fair’, 26-35 dB ‘alert’ and more than 35 dB ‘referral’. This stratification offers information for patients without the need for a professional audiologist. The developed audiometer could be used in schools or even for routine check-ups.
Conclusion
The newly developed audiometer is cost-effective and easily operated. The trial demonstrated that the threshold dB values obtained in a conference room were comparable with those of a clinical audiometer in a soundproofed room. The preliminary results also indicated that the experience of the screening personnel did not affect the testing. Moreover, the screening result of the audiometer is a threshold value, which can be utilized in records for monitoring progressive hearing change in school-aged children.
Footnotes
Acknowledgements
This study was supported by the National Science Council, Taiwan ROC (grant numbers: NSC 95-2221-E-075-001-MY2 and NSC 96-2414-H-075-001). The authors would like to thank Ms HC Lee, Division of Experimental Surgery, Department of Surgery, Taipei Veterans General Hospital, for the assistance in statistical analysis.