Component Description
The NHANES program suspended field operations in March 2020 due to the coronavirus disease 2019 (COVID-19) pandemic. As a result, data collection for the NHANES 2019-2020 cycle was not completed and the collected data are not nationally representative. Therefore, data collected from 2019 to March 2020 were combined with data from the NHANES 2017-2018 cycle to form a nationally representative sample of NHANES 2017-March 2020 pre-pandemic data. These data are available to the public. Please refer to the Analytic Notes section for more details on the use of the data.
The audiometry examination component (AUX) consists of four parts:
1) Pre-exam
audiometric questions – a series of questions to identify conditions that would
affect how audiometric testing is conducted, or how results are interpreted.
Questions include whether the participant has pressure equalization (ear)
tubes, a current cold or earache, exposure to loud noise or music through earphones
in the past 24 hours, or perceived asymmetry between ears. NOTE: more extensive
data relating to a survey participant’s hearing status and history of noise
exposure are contained in the audiometry questionnaire (AUQ) section of the
NHANES Sample Person Questionnaire.
2) Otoscopy – a brief otoscopic screening (physical) exam of the ear canals and eardrums to identify abnormalities, which would require alternate audiometric procedures or influence interpretation of test results, and to identify conditions which might require medical referral. The exam screened for excessive or impacted ear cerumen (wax), physical abnormalities, or collapsing external ear canals.
3) Middle
ear testing – a measurement of eardrum compliance by placing a probe with a
rubber tip into the entrance of the ear canal. It is used to identify middle
ear pathologies that might contribute to a hearing loss. There are three middle
ear tests: tympanometry, wideband reflectance, and acoustic reflex. The P_AUX
data file consists of tympanometric quantitative measurement variables for
middle ear volume, pressure, compliance, and gradient. It also includes
tympanogram quality rating and score variables. The raw data for the participant's
tympanometric curves (P_AUXTYM), wideband reflectance curves (P_AUXWBR), and
acoustic reflex curves (P_AUXAR) are released separately.
4) Pure
tone air conduction audiometry – a measurement of hearing sensitivity conducted
by presenting pure tone signals to the ear through earphones and varying the
intensity of the signals until a participant’s hearing threshold at that
frequency is determined. Testing is performed at frequencies across the range
of human hearing.
Eligible Sample
All participants aged 6–19 years and 70 years and older in the NHANES 2017-March 2020 pre-pandemic sample were eligible. There were no precluding conditions for any part of the audiologic exam. However, participants using hearing aids who were not able to remove them for testing and participants who had sufficient ear pain at the time of the exam that they could not tolerate headphones were excluded.
Protocol and Procedure
All
audiometry exam sections were performed by trained examiners on participants in
a dedicated sound-isolating room in the Mobile Examination Center (MEC).
Hearing threshold testing was conducted on both ears of participants at seven
frequencies. Children aged 6–11 years were tested in the following frequency
order so that data could be obtained for more important frequencies first in
case the child’s attention waned before all frequencies could be completed: Ear 1: 1000, 2000, 1000 (repeat), 4000, 8000 Hz. Ear 2: 1000, 2000, 1000 (repeat), 4000, 8000, 6000, 3000, 500 Hz. Ear 1: 6000, 3000, 500 Hz. Adolescents aged 12–19 years and adults aged 70 years and older were tested in the frequency order 1000, 500, 1000 (repeat), 2000, 3000, 4000, 6000, 8000 Hz for ear 1 and then for ear 2. The ear tested first was varied to prevent biasing the data. However, if the participant indicated one ear hearing better than the other, the ear with better sensitivity was tested first. Testing was conducted according to a modified Hughson Westlake procedure using the automated testing mode of the audiometer. The approximate audiometer output constraints for standard and insert earphones was 100 decibels (dB) at 500 to 6000 Hz and 90 dB at 8000 Hz. Some earphones might produce slightly higher output than others. A threshold greater than 90 dB can be obtained using a manual testing mode. Observed values, therefore, varied between -10 and 110 dB. If a participant did not respond to the signal tone at any level for one or more frequencies because of deafness or severe hearing loss, a code value of 666 was entered. Manual testing was conducted on children aged 6-8 years. Manual testing was also conducted when the participant could not operate the response switch or responded too
slowly or inconsistently for the audiometer to accurately record the response.
In some instances, if a pure tone audiometric signal is sufficiently loud, it can "cross over" and be heard by the opposite ear via bone conduction. When this occurs, it is difficult to determine if the threshold obtained is truly the threshold of the test ear, or an artifact of the non-test ear (which may have better hearing). For the NHANES hearing threshold testing, a crossover retesting protocol was performed whenever the observed threshold at any given frequency was poorer in one ear than the other by 25 dB at 500 and 1000 Hz or 40 dB at any higher frequency. Retesting was accomplished using insert earphones, which are smaller and have less direct contact with the bones of the head. Thus, a much louder stimulus is required before crossover occurs. Due to the complexity of the procedure, masking was not employed in the NHANES audiometry exam.
The
NHANES 2017-2018 and the 2019-2020 Audiometry Procedures Manual are available
on the NHANES website NHANES
2017-2018 Audiometry Procedures Manual and NHANES 2019-2020
Audiometry Procedures Manual.
Quality Assurance & Quality Control
Instrumentation for the audiometry component included an ViAcoustics Audiometric Research Tool (ART) system with standard TDH-49P headphones and Etymotic EarTone 3A insert earphones. Interacoustics Titan (middle ear analyzer) was used to conduct middle ear testing. Ambient noise levels were measured daily using the ART system and monitored continuously during testing with a Quest Model BA-202-27 Bioacoustic Simulator. AUX data were entered directly into the computerized NHANES database system. Data from the ART system and Interacoustics Titan were captured electronically and uploaded into the survey information system automatically.
The ART system was used to conduct the pure-tone air conduction hearing test on survey participants. The ART system is a comprehensive, highly configurable audiometer designed for the research application. The ART system consists of dedicated National Instruments hardware and ART software as an audiometer. The ART software has been programmed to conduct the threshold test automatically. The system can also be used to conduct a guided manual test, in which the software suggests the presentation level and threshold determinations using the same algorithms as the automated test, but signals are presented manually by the tester, who can override the system suggestions when deemed appropriate.
The ART system contains an integrated calibration system, called Trident, to measure the sound levels put out by the system. A G.R.A.S. calibration kit with couplers for each type of earphones is used in conjunction with Trident to monitor calibration. ART calibration checks were conducted at the beginning of each survey location, daily before the exam, and at the end of each survey location to ensure that the accuracy has not shifted. Daily calibration procedures included checking the calibration of the G.R.A.S. microphone, the background noise measurement, an acoustic calibration check to measure the output and distortion, and a listening check to evaluate tone quality, attenuator accuracy, headphone cord integrity, and crossover. Calibration checks were accomplished for each headphone type on alternating days (i.e., TDH-49 headphones one day, insert earphones the next). At the beginning and end of data collection at each survey location, the daily procedures were conducted on both sets of headphones, along with a linearity check across all seven test frequencies. In addition to these calibration checks, each ART system received an exhaustive calibration once a year or whenever unresolvable problems were discovered during the calibration checks. The calibrations were accomplished by a certified laboratory to ensure that the calibrations are traceable to the National Institute of Standards and Technology (NIST).
An Interacoustics Titan was used to conduct middle ear testing. The physical volume was checked on the Titan to ensure that it was within the calibration limits. In addition, the MEC Health Technicians ran a sample test on one ear to ensure that the unit was producing clear, normal graphs for the tympanogram and wideband reflectance. The Titan calibration checks were conducted at the beginning of each survey location, daily before the exam, and at the end each survey location. The Titan also received an exhaustive NIST-traceable calibration check annually.
A Bioacoustic Simulator was used to continuously measure the background noise levels in the audiometric test room during the exam. Pure tone audiometric testing was not performed if ambient noise levels in the test booth exceeded maximum permissible levels. The microphones and sound measurement function of the bioacoustic simulator were checked by a certified laboratory annually or whenever unsolvable problems were discovered.
As an additional quality measure, all audiograms, whether conducted in automated or manual mode, tested the 1000 Hz frequency twice in each ear as a measure of the reliability of the participant’s responses. Pure tone audiograms were not accepted if there was more than a 10 dB difference between them.
Tympanometry and/or wideband reflectance measures were sometimes repeated once per ear when the initial result was unsatisfactory (acoustic reflex testing was no repeated due to the number of loud tones presented in a single test). The best of each of the two curves was retained in the data. Subjective quality ratings for tympanograms are the variables AUAREQC and AUALEQC. These represent a qualitative assessment of each tympanogram by the consulting audiologist. The consulting audiologist also scored these tympanograms using a common classification system developed by Liden (1969) and modified by Jerger (1970). Two variables AUATYMTR and AUATYMTL provide information about what the tympanometry results mean in terms of the participant’s middle ear status.
The MEC health technicians who performed the audiometry examination component of NHANES were professionally trained by a certified audiologist from the National Institute for Occupational Safety & Health (NIOSH). NIOSH also monitored the performance of each health technician on a regular basis. Field visits to each MEC were conducted by the NIOSH audiologist at least three times per year. Additionally, NCHS project officers visited the MECs approximately twice per year to observe the audiometry examinations and verify that standard testing procedures were being followed. NIOSH consultants provided the MEC health technicians with annual retraining and protocol updates.
Data Processing and Editing
The wording of some questions asked before the audiometry exam has been simplified for children aged 6–11 years. Therefore, there are two separated sets of variables for the pre-exam audiometric questionnaire: one for children aged 6–11 years and another for adolescents aged 12–19 years and adults aged 70 years and older.
All data were captured in the NHANES computerized database system, with audiometric and middle ear analyzer data automatically uploaded. On a continuous basis, a consulting audiologist performed a clinical review of all data for each participant as it was received, checking for quality and consistency. In addition, a computerized data editing program was developed to check for logical inconsistencies in the data and technician errors, and to cross-check other issues affecting data quality (consistency in identifying potential instances where crossover effects might have occurred, assurance of randomization of the initial test ear, etc.). Back-end edits of the data were performed as needed when errors were detected.
Analytic Notes
The COVID-19 pandemic required suspension of NHANES 2019-2020 field operations in March 2020 after data were collected in 18 of the 30 survey locations in the 2019-2020 sample. Because the collected data from 18 locations were not nationally representative, these data were combined with data from the previous cycle (2017-2018) to create a 2017-March 2020 pre-pandemic data file. The resulting exam sample weights in the demographic file should be used to calculate estimates from the combined cycles. These exam sample weights are not appropriate for independent analyses of the 2019-2020 data and will not yield nationally representative results for either the 2017-2018 data alone or the 2019-March 2020 data alone. Please refer to the NHANES website for additional information for the NHANES 2017-March 2020 pre-pandemic data, and for the previous 2017-2018 public use data file with specific weights for that 2-year cycle.
Audiologic data analysis is a complex procedure and requires a thorough knowledge of the specialty content area for valid results to be obtained. If an analyst does not have professional experience in this area, it is recommended that audiologic consultation be obtained to help formulate and review the results of the analysis. Data analysts should be especially aware of the fact that the number "666" in all primary audiometric frequency data fields (AUXU1K1R through AUXU8KL) as well as in all repeat test frequencies (AUXR1K1R through AUXR8KL) is a qualitative code for nonresponse at a particular frequency and does not represent actual measured decibel hearing threshold values. These nonresponse values should be appropriately edited prior to any numerical data analysis.
The algorithm for obtaining automated audiometric thresholds has changed slightly over NHANES cycles since 1999. From 1999-2016, automated thresholds were obtained using an Interacoustics AD226 audiometer, which defined threshold as the level at which the subject responded at least 50% of the time to a minimum of 2 out of 3 trials following ascending or descending presentations; if a subject responded consistently to the first two presentations at a given level, a third presentation at that level was not required. From 2017-2020, automated thresholds were obtained using the ViAcoustics ART software described above. The ART system defined threshold as the level at which the subject responded at least 50% of the time to ascending only signal presentations. In the 2017-2018 cycle, the software was programmed to require a third signal presentation even if a subject responded consistently to the first two presentations at a given level. The software was modified for the 2019-2020 cycle to eliminate the need for a third presentation when a subject responded consistently to the first two presentations at a given level.
The algorithm for obtaining manual audiometric thresholds has been unchanged since 1999. Threshold for manual audiometry was defined based on ascending presentations only following a minimum of two out of three responses. A third presentation was not required when a subject responded consistently to the first two presentations at a given level. This threshold definition is identical to the automated threshold determination algorithm used in the 2019-2020 cycle.
An internal analysis of response histories from a separate study indicated that inclusion of descending presentations toward threshold shifted the center point of the psychometric function by less than 2.5 dB, indicating that the influence of descending responses on center and width is very small for tests using the Hughson-Westlake procedure. Analysis also found that requiring the third ascending response resulted in a different threshold determination approximately 9% of the time, typically better by 5 dB. Thus, the differing threshold algorithms used across NHANES from 1999-2020 should have minimal impact on results over time. Nonetheless, analysts should be aware of these differences when comparing thresholds across various NHANES cycles.
NHANES thresholds in cycles prior to 1999 were all obtained manually using various threshold search procedures. Analysts comparing back to cycles prior to 1999 should refer to the data documentation from the earlier surveys to understand how differences in procedures could affect the thresholds obtained.
Please refer to the NHANES Analytic Guidelines and the online NHANES Tutorial for further details on the use of sample weights and other analytic issues.
The NHANES audiometry exam also collected the Words-In-Noise data in 2019-March 2020 as a convenience sample file, AUXW_R (Audiometry - Words-In-Noise), available through the NCHS RDC.