Key Words
abnormal test results - normal audiogram - prevalence - pure tone threshold - retrospective
multi-site - Veteran
INTRODUCTION
Audiology clinics within the Department of Veterans Affairs (VA) are among the busiest
outpatient clinics in the VA health care system. Although audiologists primarily care
for Veterans with auditory deficits, many clinicians report testing patients with
normal pure-tone hearing thresholds; however, it is unknown exactly how frequently
these patients present to audiology clinics. A subset of these patients is likely
being referred to the audiology clinic with subjective complaints of difficulty hearing,
such as understanding speech in background noise or other difficult listening situations.
Little is known about the exact nature of their problems or what the most effective
treatment options may be; best practice guidelines for such situations have yet to
be formally established. An important first step to improving the diagnosis and treatment
of these individuals with normal pure-tone thresholds is to determine the prevalence
of these visits and to better characterize the results of other tests in the audiological
test battery.
It is not surprising that abnormalities in the audiological test battery may be present
even when pure-tone thresholds fall within the normal range. In fact, it is widely
accepted that normal thresholds do not necessarily indicate a normal auditory system.
Even among supposedly “normal-hearing” populations, there are individuals who experience
an unexpectedly large decrease in speech understanding when listening in challenging
listening environments ([Gates et al, 1990]; [Jerger et al, 1990]; [Stach et al, 1990]; [Cooper and Gates, 1991]). For example, normal-hearing individuals often struggle to understand speech in
background noise or in situations with rapid, reverberant, or otherwise degraded speech
([Middelweerd et al, 1990]; [AAA, 2010]). Differences between auditory function near threshold and at suprathreshold levels
have been an important distinction for many decades ([Plomp, 1978]). [Plomp (1978)] suggested that both attenuation (or reduced audibility due to hearing loss) and
distortion (deterioration of the representation of a signal due to hearing loss) may
play important roles in auditory perception. Attenuation can be resolved primarily
by improving audibility, whereas distortion in the auditory system is more difficult
to quantify and resolve. It may be that distortion within the auditory system, especially
at suprathreshold levels is at the root of hearing difficulties experienced by those
with normal pure-tone thresholds. Indeed, [Martin et al. (1998)] argued that pure-tone testing does not address distortion-related deficits. Patients
with normal hearing thresholds who report having hearing difficulties have been studied
for decades. Historically, these patients have been classified as having various conditions,
including: auditory inferiority complex ([Byrne and Kerr, 1987]), auditory disability with normal hearing ([Rendell and Stephens, 1988]), selective dysacusis ([Narula and Mason, 1988]), obscure auditory dysfunction ([Saunders and Haggard, 1989]), King–Kopetzky syndrome ([Kopetzky, 1948]; [King, 1954]; [Hinchcliffe, 1992]) auditory dysacusis ([Jayaram et al, 1992]), and idiopathic discriminatory dysfunction ([Rappaport et al, 1993]). Although definitions and symptoms of these conditions vary, a common characteristic
of these patients is their difficulty hearing even with normal pure-tone thresholds.
Currently, clinicians may classify this group of patients as having an auditory processing
disorder or deficit ([Jerger et al, 1990]; [AAA, 2010]). Reports from these patients show that a diagnosis of normal hearing combined with
a lack of treatment recommendations may result in feelings of dismissal and confusion
([Pryce and Wainwright, 2008]).
Within the VA, clinicians report frustration with determining what to do for Veterans
with subjective complaints of hearing difficulty in the context of normal pure-tone
thresholds. An informal survey administered to VA audiologists by [Saunders and Abrams (2009)] estimated that 92% of responding audiologists reported encountering at least one
Veteran per month with normal hearing who complained of hearing difficulties. The
proportion of audiologists that reported encountering such a patient one to three
times per month was 53% and 39% reported encountering four or more per month. Although
33% of these audiologists reported issuing a personal FM system and 26% reported recommending
auditory training, 30% reported being unsure of how to treat these patients. That
is, a third of respondents did not know how to treat individuals with hearing difficulties
and normal thresholds. This demonstrates a need to improve our understanding of the
extent and nature of this clinical challenge. The Veteran population is unique in
that individuals with normal hearing often present with comorbidities related to military
service, such as blast exposure and traumatic brain injury ([Gallun et al, 2012]), tinnitus ([Helfer et al, 2005]), posttraumatic stress disorder ([Gordon et al, 2017]) or other mental health conditions (such as depression and anxiety), or auditory
processing deficits ([Saunders et al, 2015]). Each of these factors may exacerbate hearing and processing difficulties.
The goals of this study were (a) to determine the prevalence of Veterans visiting
VA audiology clinics who have normal pure-tone hearing thresholds and (b) to determine
the prevalence of abnormal audiological measurements in these patients. This is the
first large-scale report that characterizes Veterans with normal hearing thresholds.
METHODS
Data Source
This retrospective descriptive study was conducted on data collected within the VA
health care system. We used data obtained from the U.S. Department of Veteran Affairs
Office of Acquisition and Logistics. The Denver Acquisition and Logistics Center (DALC)
supports the VA nationwide with acquisition and logistics services, including data
storage and distribution of hearing aids, hearing aid accessories, and assistive listening
devices. The DALC audiological repository data included in this study consisted of
hearing assessments from 259 VA sites in the United States between 1991 and 2015.
The repository is a useful dataset from which to establish the range of pure-tone
thresholds among Veterans seeking care within outpatient audiology clinics. The repository
contains more than three million audiograms from audiology visits by Veterans. To
date, however, only a handful of peer-reviewed works have been published from this
dataset and are primarily focused on hearing impairments including audiological notches
([Wilson, 2011]; [Wilson and McArdle, 2013]; [2014]).
It is worth noting that although the repository dataset is quite expansive, it is
not exhaustive or complete. Not all sites or audiologists within a site contributed
records to the repository from every patient or every test during the date range queried
for this study. An informal survey of VA audiologists we conducted demonstrated that
there has been a wide range of compliance with requests for audiologists to submit
records to the repository. Identification of patients’ sex was not available in the
repository, nor were there records of other health conditions, noise exposure history,
or specifics regarding hearing difficulty.
Data Analysis
Data were summarized and examined using descriptive statistics including frequencies,
percentages, means, and standard deviations (SD). All data were assessed using Excel
and SPSS software. All patients aged 90 yr and older were aggregated into a single
category in compliance with the Health Insurance Portability and Accountability Act
Privacy Rule, which classifies ages over 89 yr as identifiable data. Abnormal results
were determined by comparing with commonly accepted standards (shown in [Table 1]). The normative values listed in [Table 1] come from common and well-supported definitions of normal status that are applicable
to this study population. Normal hearing was defined as pure-tone thresholds ≤25 dB
HL at each octave frequency from 250 to 8000 Hz.
RESULTS
At the time the data were obtained, the total number of records in the DALC audiological
data repository was 3,641,326. [Figure 1] depicts the study population and how the analytic sample was determined. For patients
with more than one audiological evaluation on file, only the entry that included the
best audiogram was used for analysis, resulting in 2,322,771 unique patient records.
For these purposes, the “best” audiogram was defined as the one with the lowest binaural
average of air-conduction thresholds at 250, 500, and 1000 Hz.
Figure 1 Epidemiological decision tree for DALC audiological data repository audiograms. Final
analysis was performed on 235,091 audiograms with normal pure-tone thresholds within
the target age and date ranges.
Data included records for patients that were seen between April 1, 1991 and June 21,
2015, who were ≥19 yr of age. Those younger than 19 were excluded in compliance with
the approved institutional review board protocol. Subsequently, only records with
present and normal pure-tone thresholds (≤25 dB HL) at all octave frequencies (250,
500, 1000, 2000, 4000, 8000 Hz) in both ears were retained. The final analysis group
was composed of 235,091 patients with normal hearing thresholds. Most of these 235,091
patients also had pure-tone thresholds reported at 3000 and/or 6000 Hz (n = 229,270),
which were used to perform analyses related to audiological notches at these frequencies
but were not used to define exclusion criteria. Patients with pure-tone thresholds
above 25 dB HL at 3000 and 6000 Hz were still considered to have normal hearing thresholds
for the purposes of this study.
[Figure 2] shows a scatterplot of high-frequency pure tone averages (PTAs) for each record
plotted as a function of age. The mean age of participants was 37.14 yr (SD = 11.7
yr; range = 19–90+ yr). The mean bilateral high-frequency PTA (average of 1.0, 2.0,
and 4.0 kHz values) was 10.59 dB HL (SD = 5.0; range = −10 to 25 dB HL). [Figure 2] excludes participants 90 yr of age or older (n = 12) and participants with no age
listed (n = 77), and includes a linear trendline (R
2 = 0.0758) that illustrates a limited relationship. It should be noted that this value
will not represent the general relationship between age and hearing loss because thresholds
were limited to 25 dB HL.
Figure 2 Scatterplot of all audiograms. The PTA (1.0, 2.0, and 4.0 kHz) of each record is
shown as a function of age and demonstrates the large number of normal audiological
records that made up this dataset (n = 235,091). Y-axis values were jittered to display individual data points. Older Veterans exhibit
higher PTAs on average even within bounds of normal hearing, as seen by the dashed
linear trendline.
Prevalence of Normal Pure-Tone Thresholds
Of the 2,322,088 unique patients, 235,091 (10.12%) were determined to have normal
hearing, with pure-tone thresholds ≤25 dB HL at all octave frequencies. The number
of normal-hearing patients by age decade is displayed in [Table 2], and mean thresholds for different decade groups are shown in [Figure 3]. Thresholds generally increase with age, especially at high frequencies (4000 and
8000 Hz), meaning that on average patients are less likely to have normal hearing
as they get older. The majority (141,564 or 60.2%) of individuals fell between 19
and 39 yr of age.
Table 2
Breakdown of Included Records by Decade of Life
|
Age
|
Number
|
Percent
|
|
19–29
|
81,662
|
34.75
|
|
30–39
|
59,902
|
25.49
|
|
40–49
|
54,604
|
23.23
|
|
50–59
|
27,602
|
11.74
|
|
60–69
|
10,416
|
4.43
|
|
70–79
|
743
|
0.32
|
|
80 and older
|
85
|
0.04
|
|
Sub-total
|
235,014
|
100
|
|
No age available
|
77
|
—
|
|
Total
|
235,091
|
—
|
Figure 3 Mean pure-tone air-conduction thresholds for each ten-year age group. Veterans 70
yr and older were grouped together because of the comparatively small number of normal
audiograms available. Age groups show differences in high-frequency hearing, and there
is a visible elevation of hearing thresholds with increasing age, even within this
population of ostensibly normal-hearing individuals. Error bars display the standard
error of the mean, although they may be difficult to see because the large size of
the dataset produces very small standard error estimates.
Other Abnormal Audiological Test Results
[Figure 4] shows how many of the 235,091 normal-hearing patient records contained results for
each of several other audiological test measures. The most commonly recorded test
measure was word recognition (n = 217,396), followed by tympanometry (n = 156,078),
bone-conduction audiometry (n = 123,658), ipsilateral acoustic reflexes (n = 97,966),
and contralateral acoustic reflexes (n = 71,936). Abnormal results were determined
by comparing with commonly accepted standards (shown in [Table 1]). Contralateral acoustic reflex threshold (ART) tests had the highest prevalence
of abnormal results compared with the number of individuals tested (31.7%). Air-bone
gaps (ABG) had the lowest prevalence at 1.74% abnormal results of tests performed.
The overall prevalence of any abnormality in either ear for this normal-hearing population
was 41.10%. An abnormality was defined as an audiological notch at 3000, 4000, or
6000 Hz, or an abnormal test result in any of the tests listed in [Table 1].
Figure 4 Number of normal-hearing VA patient records containing each of five different audiological
tests. Not all tests were completed for the full sample; gray bars represent the total
number of records where results for the given test were available. Black bars represent
the number of abnormal results for each test. The proportion of abnormal tests to
total tests performed is expressed next to each bar as a percentage.
[Table 3] shows the mean depth and prevalence of unilateral and bilateral notches for 3000,
4000, and 6000 Hz notches. Notch depth was defined as the difference between the threshold
at the notch frequency and the average of the thresholds at its two adjacent octave
frequencies. Left-ear notches were found to be more prevalent than right-ear notches
at all frequencies. Notches were also deeper in the left ear than in the right ear
for all notch frequencies, although no statistical analysis was performed to determine
significance because of the descriptive nature of the study. Bilateral notches were
the least common, consistently appearing less than half as often as right-ear notches.
Average thresholds for 3000 and 6000 Hz notches are shown in [Figure 5]. Most of the 235,091 records contained pure-tone thresholds at 3000 and/or 6000
Hz resulting in a total of 229,270 records used for notch analysis. Notches at 6000
Hz were deeper on average than 3000 Hz, but 6000 Hz was tested less frequently (211,536
audiograms compared with 228,161). Notches at 4000 Hz may be shallower in this dataset
because the threshold at the notch center frequency was always ≤25 dB HL because of
the inclusion criteria.
Table 3
Depth and Prevalence of Unilateral and Bilateral Notches
|
Notch Frequency (Edge Frequency)
|
Right‐Ear Notch Depth (Standard Deviation)
|
Left‐Ear Notch Depth (Standard Deviation)
|
Right-Ear Notch Prevalence (%)
|
Left-Ear Notch Prevalence (%)
|
Prevalence of Bilateral Notch (%)
|
|
4 (2, 8)
|
12.728 (2.94)
|
13.29 (3.20)
|
4.79
|
6.66
|
2.18
|
|
3 (2, 4)
|
12.69 (4.01)
|
13.34 (4.62)
|
2.22
|
2.54
|
0.46
|
|
6 (4, 8)
|
14.13 (5.04)
|
14.40 (5.36)
|
5.85
|
6.41
|
2.21
|
Figure 5 (A) Right-ear and left-ear notches at 3000 Hz. On average, left-ear notches are deeper
than right-ear notches in this population. (B) The 6000-Hz notches show a similar
pattern, but have a greater average depth (see [Table 3]). Error bars display standard errors of the mean, but may be difficult to see because
of the large sample size and the resulting small standard errors.
DISCUSSION
An important premise of this paper is the definition of “normal hearing” as it pertains
to pure-tone thresholds, which varies across clinical settings and publications. The
American National Standards Institute standards define audiological zero (0 dB HL)
for each frequency as the average threshold level at that frequency for normal-hearing
young adults. For frequencies measured in a typical audiological evaluation (octave
intervals 250–8000 Hz), the SD from the mean (i.e., 0 dB HL) is 3–5 dB, depending
on the frequency ([Wilber et al, 1988]; ANSI, 2010). Although early descriptions defined the normal range as 3 SDs from
audiological zero, today many clinicians and researchers do not use this definition.
It is common for contemporary definitions to be based on a three-frequency (500, 1000,
2000) or four-frequency (500, 1000, 2000, 4000) PTA. A PTA, even if it is calculated
from thresholds at each octave test frequency from 250 to 8000 Hz, allows for normal-hearing
individuals to have some elevated thresholds. More stringent definitions of normal
hearing place an upper limit on thresholds at each frequency. An upper limit (applied
to each octave frequency) of 25 dB HL was chosen for this study because it is common
in clinical protocols, which ensures that results are applicable to clinical settings
([Glorig et al, 1956]; [Goodman, 1965]; [Glorig, 1966]). It should be noted that an upper limit of 25 dB HL to define normal hearing allows
for a large range of “normal” variation that may be problematic when considering functional
deficits and hearing difficulty complaints in patients seen in the clinic.
The prevalence of Veterans with normal hearing thresholds visiting VA audiology clinics
in the DALC audiological data repository was 10.12%. Other studies have reported a
prevalence of around 5% normal-hearing individuals (excluding middle ear pathologies)
in civilian clinical populations ([Saunders and Haggard, 1989]; [Higson et al, 1994]). We suspect that the prevalence reported herein may underestimate patient visits
within the VA system. Requirements for entry of information into the repository vary
both by clinic site and clinician as some enter all data into the DALC repository,
whereas others only use it to track hearing aid orders. In an informal survey of VA
audiologists that we conducted, it was reported that when hearing aids were not ordered,
the test results were not consistently entered into the repository at every site or
by every audiologist. The magnitude of this effect is unknown, but would result in
an underestimation of the prevalence of normal audiograms (i.e., fewer normal records
would be entered than records that reflect hearing loss). Furthermore, the ratio of
reported data included in this analysis has changed over the years depending on the
clinic and clinician.
It is not known what portion of these normal-hearing individuals report hearing difficulty.
Future studies could examine visit notes or diagnosis/treatment codes to determine
the reason for visits as well as noise exposure history and other comorbid conditions.
Medication effects may also be important to consider. Medications might be prescribed
for Veterans with various comorbid conditions (e.g., post-traumatic stress disorder)
who may also report hearing difficulties. It has been noted that these medications
should be used with caution for some patients as it may cause confusion and deterioration
of cognitive performance ([VA/DoD, 2010]). A deterioration of cognitive performance may exacerbate existing difficulties
with hearing, especially in the presence of background noise. All of these factors
are important considering that Veterans come to the clinic with different medical
histories than the average civilian.
Although Veterans with normal hearing visit audiology clinics for a variety of reasons,
including referral from case manager, compensation and pension evaluation, employment
screenings, and problems with tinnitus or balance (rather than sensitivity or clarity
of hearing), we assume that some of them are seeking care for hearing and listening
difficulties. It is unknown what the exact causes of hearing problems are among individuals
with normal hearing thresholds. In an attempt to begin to understand this issue, the
second goal of this study was to report abnormalities found among the results from
other audiological tests often included in a comprehensive testing battery. Among
Veterans with normal pure-tone thresholds, the 41.10% prevalence of abnormal test
results illustrates the importance of considering the complete battery of audiological
tests. It is known that normal hearing thresholds do not necessarily indicate a normal
auditory system, nor do they necessarily predict how an individual will perform in
challenging listening situations. The high rate of abnormal test results among these
individuals presents an argument for more rigorous testing, even for individuals who
present with normal hearing. This is especially true if Veterans come complaining
of hearing difficulty and that difficulty is not adequately captured by their audiogram.
For example, abnormal physiological test results may be related to middle-ear or reflex-arc
problems that disrupt the normal encoding of information, abnormal speech-testing
results may indicate distortion in auditory processing, and the presence of elevated
thresholds and notches, even within the normal range, may decrease speech-in-noise
understanding. The greater depth and prevalence of left-ear notches are difficult
to explain without information about noise exposure. Given the nature of the Veteran
patient population, it is conceivable that these notches are a result of firearm use
or blast exposure. The notch prevalence and depth results from this study are consistent
with previous findings ([Wilson, 2011]; [Wilson and McArdle, 2013]). The nontrivial number of 3000 and 6000 Hz notches found in this population whose
octave thresholds all fall within normal limits speaks to the necessity of testing
interoctave frequencies because these are potential abnormalities that might otherwise
be hidden in an audiogram that only examines octave frequencies. Although more comprehensive
testing can provide additional clarity, it must be noted that 58.9% of the patients
with normal audiograms also had normal results on all other audiological tests we
examined despite the likelihood that some of them came into the clinic with hearing
complaints. For these Veterans, conventional diagnostic methods and treatments do
not provide adequate solutions.
In addition to diagnostic considerations, it is challenging for audiologists to develop
treatment plans for these normal-hearing individuals who present with complaints of
hearing difficulty. These challenges present themselves regardless of whether other
audiological test results are abnormal. The presence of a normal audiogram does not
necessarily mean there is no damage to the auditory system. The audiogram, on which
we often base our judgment of a normal or nonnormal auditory system, only represents
near-threshold testing. The existence of many diverse etiologies that may result in
hearing difficulties further contributes to the lack of consensus regarding treatment
for these patients. The utilization of the three treatment options for these patients
(low- to mild-gain amplification, FM systems, and auditory training) may or may not
be appropriate depending on each Veteran's specific needs. It is noteworthy that normal
pure-tone threshold prevalence in the Veteran population was twice that found in the
civilian population.
CONCLUSIONS
This large-sample retrospective analysis of Veterans establishes an estimate of the
prevalence of Veterans visiting VA audiology clinics with normal pure-tone thresholds
at 10%; in addition, estimates of the audiological abnormalities on other tests (tympanometry,
acoustic reflexes, word recognition, and bone conduction) in this population are detailed.
These findings may inform future studies that investigate why people with normal pure-tone
thresholds sometimes report difficulties hearing, and they suggest that clinicians
should think beyond the audiogram when considering what constitutes “normal hearing.”
Abbreviations
ART:
acoustic reflex threshold
ABG:
air-bone gap
DALC:
Denver Acquisition and Logistics Center
PTA:
pure tone average
SD:
standard deviation
VA:
Veterans Affairs
