Journal Logo

Research Article

Occupational Noise Exposure and Longitudinal Hearing Changes in Post-9/11 US Military Personnel During an Initial Period of Military Service

Reavis, Kelly M.1,2; McMillan, Garnett P.1; Carlson, Kathleen F.1,2,3; Joseph, Antony R.4; Snowden, Jonathan M.2; Griest, Susan1,5; Henry, James A.1,5

Author Information
doi: 10.1097/AUD.0000000000001008

Abstract

INTRODUCTION

This article presents the estimates of the annual rate of hearing threshold change by audiometric test frequency during military service as a consequence of military occupational noise exposure ranking. This issue is important because hearing loss is often an unfortunate consequence of military service and is an important concern for the US Departments of Defense (DoD) and Veterans Affairs (VA). For Active-Duty Service members, hearing loss makes it difficult to engage in war fighting and to perform job tasks, thereby impacting fitness for duty (Abel 2005; Yong & Wang 2015). It is also associated with many adverse health outcomes including unintentional injuries, falls, social isolation, depression, cognitive decline, hospital readmissions, and increased health care costs (Mulrow et al. 1990; Carabellese et al. 1993; Appollonio et al. 1996; Cacciatore et al. 1999; Dalton et al. 2003; Saunders & Griest 2009; Lin 2011; Lin et al. 2011; Lin & Ferrucci 2012; Genther et al. 2013,2015; Reed et al. 2019). As of September 2019, over 1.3 million Veterans were service-connected for hearing loss, meaning their hearing loss was officially attributed to their military service and they received a disability award from the Veterans Benefits Administration. In addition to rehabilitation costs, some of these Veterans receive financial compensation for their hearing loss placing an economic burden on the VA health care system (Department of Veterans Affairs 2018). Hearing loss acquired during military service can lead to a lifetime of consequences for Service members and Veterans. Understanding the longitudinal trends in hearing thresholds among military personnel may inform prevention and rehabilitation strategies, leading to increased quality of life.

Noise exposure is the primary cause of hearing loss among military personnel. Prevention of hearing loss is clearly important to the DoD, which engages in audiometric surveillance to abate and mitigate auditory injury (Department of Defense 2004,2010,2019). The development of noise-induced hearing loss in the military has been described primarily on the basis of a calculation referred to as significant threshold shift (STS), a clinically meaningful change in hearing thresholds in the speech frequency range compared with a baseline hearing test (Occupational Safety and Health Administration 1983; Department of Defense 2004). Specifically, STS in the military is defined as a change in hearing by 10 dB or more in the unweighted average of hearing thresholds at test frequencies 2, 3, and 4 kHz. On an individual level, the detection of STS activates a workplace intervention to prevent further loss of hearing sensitivity. On a population level, STSs are an important metric for measuring the effectiveness of hearing conservation programs (HCPs). For instance, a report by the DoD Hearing Center of Excellence noted the rate of STS in 2018 among Service members enrolled in an HCP was 11% and had remained relatively constant from 2012 to 2018 (Department of Defense Hearing Conservation Working Group 2020). However, due to the emphasis placed on a shift in the average hearing threshold across the speech frequency range, STS are likely to miss early hearing threshold changes and are not suitable for estimating hearing change over time within an individual. To date, very few studies have examined the rate of hearing threshold changes at either individual or population levels in relation to military noise exposure (Collée et al. 2019). Rate of early career hearing threshold change might be especially relevant, given the link between early noise exposures and accelerated age-related hearing loss later in life (Gates et al. 2000; Kujawa & Liberman 2006).

The military consistently strives to improve its hearing conservation efforts. In general, the DoD is shifting away from audiometric surveillance of only individuals deemed “at risk” for noise exposure to a broader population-health approach, encouraging the audiometric surveillance of hearing among all military personnel (Nelson et al. 2017; Department of Defense 2019). In part, this programmatic change is warranted due to the notion that all Service members will be excessively exposed to hazardous noise at some point during their military career (Yankaskas 2013). However, there remains inconsistencies in various military strategies between service branches for HCP implementation and oversight, audiometric surveillance, and hearing protection (US Government Accountability Office 2011). For example, it has been suggested that hearing protection compliance and enforcement may vary based on noise exposure level, thus placing military personnel at risk for hearing threshold changes (US Government Accountability Office 2011).

The aim of this study was to estimate the average annual rate of hearing threshold change during military service for individuals enrolled in the Noise Outcomes in Service members Epidemiology (NOISE) Study and who served after September 2001 (Henry et al. 2020). It is important to note that the analysis was designed to assess the difference in the average hearing threshold change by occupational noise exposure. A deeper understanding of the rate of hearing threshold change, and the occupational factors that influence the development of hearing threshold change, are important to clarify the impact of military service on hearing, guide prevention strategies, and administer subsequent care.

MATERIALS AND METHODS

Data Sources

To estimate the annual rate of hearing change during military service, we require a dataset with details about military occupational noise exposure and audiometric thresholds collected over time. Such a dataset does not exist but can be constructed through the joining of two disparate data sources: (1) The NOISE Study; and (2) Defense Occupational and Environmental Health Readiness System-Hearing Conservation (DOEHRS-HC) data repository.

NOISE Study

The NOISE Study is a longitudinal investigation into the effects of military and nonmilitary exposures on auditory functioning among post-9/11 Service members and Veterans. Individuals are eligible to participate in the NOISE Study if they are Active-Duty or within about 2.5 years from separation. Two sites are enrolling study participants, the National Center for Rehabilitative Auditory Research (NCRAR), located at the VA Portland Health Care System in Portland, Oregon and the DoD Hearing Center of Excellence on Joint Base San Antonio, in San Antonio, TX; in the present study, only NCRAR-based participant data were used. Potential study participants are excluded if they did not serve post-9/11 or have been separated for more than 2.5 years from military service. Study participants undergo a comprehensive audiologic assessment and complete numerous questionnaires measuring demographics and relevant military and nonmilitary exposures.

The military occupational history of study participants served as the basis for assigning a noise exposure category (defined later). Individuals were asked to note each military occupation held during service, and the timing and duration of each job held. Only military service occupations occurring within the first “continuous service period” were considered. Continuous service period is defined as having no breaks greater than 1 month during service. For example, if an individual had enlisted for 6 years, left the military, and then re-enlisted 2 years later, only data from the first 6 years of that individual’s service period were used for analysis. This focused our analyses on military personnel’s initial service periods because we were unable to account for noise exposures that may have occurred between service periods. For military personnel still enlisted, the end of service is defined as the date of enrollment into the NOISE Study. The study was reviewed and approved by the Institutional Review Board of the VA Portland Health Care System. Informed consent is obtained from all individuals before data collection, and study participants are compensated for their effort.

Defense Occupational and Environmental Health Readiness System-Hearing Conservation

In the military, audiometric surveillance is conducted using the DOEHRS-HC (Department of Defense 2019). The DOEHRS-HC system, which was implemented in 2000, collects, maintains, compares, and reports audiometric thresholds for DoD personnel. Currently, each branch of service (Air Force, Army, Marine Corps, and Navy) and component (Active Duty, National Guard, and Reserves) utilizes DOEHRS-HC to monitor the hearing health of its personnel and is an integral component of the DoD HCP. Generally, audiometric surveillance occurs when military personnel are routinely exposed to hazardous levels of noise, just before and after combat deployments, and when separating from the service.

Over the time period examined (2001 to 2017), the DoD HCP underwent changes. Originally cast as a risk-based conservation program, military personnel were only enrolled if they were deemed noise-exposed (i.e., at or above 85 dB A-weighted as an 8-hr time weighted average; impulse noise of ≥140 dB peak sound pressure level; or ultrasonic exposures; Department of Defense 2004). In addition to audiometric surveillance, military personnel in the HCP are required to receive hearing protection fitting and hearing loss prevention education. In 2006, the Army enforced a population-level HCP, which required all Service members be enrolled and receive annual audiometric surveillance; the Marine Corps followed suit in 2012 (US Government Accountability Office 2011; Nelson et al. 2017). The Navy and Air Force have yet to implement these programmatic changes. Additional policy differences may exist between branches. As such, the total number of audiograms per individual was expected to vary based on their military service, including branch, component, occupational specialty, and whether a hearing threshold shift was observed.

Analytic Sample

Through approved data use agreements between the VA and DoD, we requested DOEHRS-HC audiograms for the first 367 NOISE Study participants enrolled at the NCRAR site. The audiograms were linked using name and social security number with NOISE Study participant data. We restricted the analysis to individuals who served only after September 2001 because this date aligns with the inception date of the DOEHRS-HC data repository (2000) and the beginning of the most recent series of military conflicts. These restrictions reduced our available sample to 246 study participants. These 246 participants provided data from 2615 audiograms between October 2001 and June 2017. Formation of our analytic sample is displayed in Figure 1.

Fig. 1.
Fig. 1.:
Flowchart depicting entry into the analytic sample. *Coast Guard personnel are not included in DOEHRS-HC. DOEHRS-HC indicates Defense Occupational and Environmental Health Readiness System-Hearing Conservation; NOISE, Noise Outcomes in Service members Epidemiology.

Measurements

Cumulative Occupational Exposure to Military Noise

Capturing military noise exposure information is especially challenging because military personnel work in large and acoustically diverse environments and perform a wide spectrum of operations and tasks, resulting in complex noise exposure patterns. Furthermore, quantitative dosimetry measurements are not easily accessible, which forces epidemiologic exposure assessment to be ascertained by self-report or through expert opinion. For measurable hearing threshold change, the noise exposure has already occurred. It is thus necessary to estimate past exposures, where individual-level noise exposure was not measured. Retrospective assessment of exposures for epidemiologic purposes adds another layer of measurement complexity. Self-reported noise exposures create study validity problems to the extent that individuals with symptoms such as hearing difficulties and/or tinnitus, or knowledge of hearing loss in the absence of symptoms, may overestimate their noise exposure leading to inflated estimates of exposure-outcome relationships, a type of recall bias. Because of these inherent problems, retrospective noise exposure assessment via a job exposure matrix (JEM) may be the most accurate way to reconstruct past noise exposures.

A surrogate measure of occupational noise exposure is given by a JEM which ideally combines qualitative (e.g., service branch and occupation classification) and quantitative (e.g., workplace dosimetry) exposure information. The JEM used in the VA is the “Duty Military Occupational Specialty Noise Exposure Listing.” The VA JEM categorizes military occupations within each service branch as having a low, moderate, or high “probability of exposure to hazardous noise.” The language used by the VA is “probability” and this corresponds to expert opinion alone. In other occupational settings, expert opinion has been successful at rank ordering exposure levels even though the expert opinion estimates of exposure correlate poorly with quantitative measures (Teschke et al. 2002). The JEM noise exposure rank was assigned to each military occupation held by each of the 246 study participants during their military service. Military personnel can be assigned more than one rank if they held more than one occupation. Using the relation between military occupation and noise exposure ranking (low, moderate, high), the number of months in a military occupation becomes a proxy measure of overall noise exposure.

Cumulative noise exposure to an occupation was determined by each individual’s audiogram date. As of each audiogram date, we computed the total number of months held in an occupation with low, moderate, and high noise exposure ranking. Basic training, which is completed by all Service members, was considered its own exposure with time exposed based on military service branch (Army = 10 weeks; Marine Corp = 13 weeks; Navy = 8 weeks; Air Force = 9 weeks; Powers 2011). Cumulative exposure is the sum of the number of months exposed to basic training, to an occupation, or to both between service entry and the first audiogram date. Each subsequent audiogram had the same number of months exposed as the earlier audiogram plus the number of months that occurred between audiograms. An example of how cumulative noise exposure was determined for a single individual is shown in Figure 2. The x axis is time since service entry (in months) and the y axis is the audiogram index, ordered from 1 = first audiogram, 2 = second audiogram, etc. The shaded regions of each horizontal bar represent the accumulated months of exposure to basic training (dark gray), an occupation with a high noise exposure ranking (medium gray), or an occupation with a moderate noise exposure ranking (light gray). An open circle indicates the audiogram date. In this example, the individual received an audiogram at service entry and therefore had accrued 0 months of military noise exposure. At the second audiogram, approximately 1 year after service entry, this individual had accrued 3 months of exposure to basic training and 9.3 months to a military occupation with a high noise exposure ranking. They then switched to an occupation with a moderate noise exposure ranking just before their sixth audiogram. At this time, cumulative exposure was calculated as 3 months of basic training, 60 months in a military occupation with a high noise exposure ranking, and 2.7 months in an occupation with moderate noise exposure ranking. The key point is that for each audiogram date, the effects of military occupational noise exposure are given by all the exposures up to that date.

Fig. 2.
Fig. 2.:
Example of cumulative noise exposure determination. Each bar represents the accumulation of exposure to an occupation by the date of the audiogram. The varying shades of gray denote varying occupations as determined using the VA job exposure matrix. In this example, the individual was exposed to basic training noise, an occupation with a high noise ranking, and an occupation with a moderate noise ranking. Furthermore, they had an audiogram at service entry (audiogram number 1) and a termination audiogram (audiogram number 8) 5 mo before separating. At the time of the termination audiogram (audiogram number 8), this individual had accumulated noise exposure during 3 mo of basic training, 60 mo at an occupation with a high noise exposure ranking, and 27.5 mo at an occupation with a moderate noise exposure ranking. Exposures occurring between the termination audiogram and the end of service were not included. VA indicates Veterans Affairs.

Estimating the association between noise exposure and hearing threshold change depends on the quality of the noise exposure assessment. The estimate can be biased if military personnel are misclassified with respect to their noise exposure status. Incorrect exposure measurements can dilute or exaggerate the relevant associations. To examine the potential for misclassification, VA JEM rankings were compared with participants’ self-reported occupational noise exposure, obtained retrospectively during NOISE Study participants’ baseline data collection. Study participants were asked how often they were exposed to loud noise during their occupation (response scale: never, several times a year, several times a month, several times a week, and daily). Concordance between the JEM ranking and self-reported loud noise exposure was examined in tabular format using counts and proportions; this comparison is displayed in Table 1. There were 299 total occupations among the 246 study participants. Table 1 shows general concordance between the VA JEM noise rankings and self-reported noise exposure. Military personnel in occupations with a high noise exposure ranking generally reported daily and weekly exposure to loud noise; whereas personnel in occupations with low noise exposure rankings generally reported yearly or monthly exposure to loud noise. In addition, no military personnel reported never having loud noise exposures. All occupations were reported to have loud noise exposures at least several times a year. The broadly similar results between the two different exposure methods increases our confidence that the noise exposure categories were accurately assigned.

TABLE 1. - Self-reported frequency (N, row %) of exposure to loud noise by job classification noise exposure ranking among 299 military occupations for 246 NOISE Study participants
Occupation
Noise Exposure Ranking
Self-Reported Frequency of Loud Noise Exposure
Never (%) Several Times a Year (%) Several Times a Month (%) Several Times a Week (%) Daily Total
Low 0 (0) 22 (29.7) 20 (27.0) 16 (21.6) 16 (21.6) 74
Moderate 0 (0) 12 (15.8) 12 (15.8) 24 (31.6) 28 (36.8) 76
High 0 (0) 18 (12.1) 28 (18.8) 33 (22.2) 70 (46.0) 149
Total 0 52 60 73 114 299
NOISE, Noise Outcomes in Service members Epidemiology.

Pure-Tone Hearing Threshold Outcome

The audiogram is the outcome. The audiogram is composed of air-conducted hearing thresholds recorded in dB HL in both the right and left ears at 6 test frequencies: 500, 1000, 2000, 3000, 4000, 6000 Hz. Currently, 8000 Hz is not routinely tested for baseline and monitoring audiograms within the military HCP; however, it is tested by audiologists when individuals are referred for further evaluation. Pure-tone thresholds were typically obtained and recorded using a computer-automated audiometer, in a certified sound booth or a quiet space, by a trained audiometric technician under the supervision of a licensed audiologist. Ambient noise levels were measured annually for stationary sound booths. For testing conducted outside the sound booth, ambient noise levels were surveyed before and after testing (Department of Defense 2004,2010). Despite these best efforts, the intrusion of ambient noise was possible. Audiometers were functionally checked daily and acoustically calibrated annually according to ANSI S3.6 standards (American National Standards Institute 1996, 2004,2010). Pure-tone thresholds were obtained in both ears in 5-dB steps ranging from −10 to 90 dBHL using a modified Hughson-Westlake procedure; threshold was defined as 2 of 3 positive responses during ascending presentation (Carhart & Jerger 1959). For the purposes of this study, hearing threshold values were not averaged across frequencies or ears.

Occasionally, an individual had repeat audiograms administered on the same day to confirm suspected hearing threshold shifts. When this occurred, only the last audiogram of the day was used. Audiograms associated with self-reported ear, nose, or throat problems were excluded from the analytic sample (n = 34 audiograms from n = 15 individuals). One individual had hearing threshold responses that exceeded the limits of the audiometer at 6000 Hz during 1 visit. This single value was set to missing.

As with any administrative data source, there is the potential for errors. Currently, no research standards exist for utilizing pure-tone thresholds established within a military HCP. Extant literature in nonmilitary industrial workplace settings suggests that audiograms be reviewed for the influence of ambient noise, improbable threshold values, large interaural differences, and missing threshold values (Masterson et al. 2013). As a sensitivity analysis, audiograms with potential errors were removed from the final dataset and the statistical analysis was repeated.

Covariates

Characteristics of the sample were obtained from both the DOEHRS-HC data repository and the NOISE Study questionnaires administered at the time of study enrollment. Age and service branch (Army, Marine Corps, Navy, Air Force) were collected at the time of the audiogram and obtained from the DOEHRS-HC data repository; these variables were used in our statistical model (described later). To characterize our analytic sample, we used age and service branch information obtained at the time of enrollment in the NOISE Study. Gender (male/female), race/ethnicity (White/other than White), enlistment duration, service component (Active, Reserve, National Guard), number of deployments (0, 1, 2, 3 or more), and number of military occupations were also derived from the NOISE Study.

Statistical Analysis

The overall goal of analysis was to estimate the average annual rate of hearing threshold change during military service among post-9/11 Veterans by military occupational noise exposure ranking. The unit of analysis was the audiogram (n = 2615). Pure-tone thresholds are expected to naturally vary among individuals, between ears within the same individual, and across frequency within ears of the same individual. We estimated the average annual rate of hearing threshold change using results of a hierarchical linear model, accounting for the natural variability in pure-tone thresholds. In addition, this analytic approach accounts for unequal intervals and missing observations (Singer & Willett 2003). Thus, all observations (including military personnel with only one audiogram) are retained in the analysis, contributing to the estimation of regression parameters at the time for which participants contributed data.

We modeled the cumulative probability of noise exposure based on military occupation (in months) and frequency as an interaction term and fit our regression model with both frequency and frequency2 terms. The model also included age at the time of the audiogram. Two-way interactions between age and frequency, and between age and frequency2 were included to allow frequency changes to vary by age. Service branch at the time of the audiometric visit was included; three-way interactions among service branch, frequency, and frequency2, and cumulative noise exposure were included to allow the longitudinal patterns of change to vary with service branch and to allow the rates of changes in threshold to vary with different occupations and frequencies. Finally, to account for natural variation in the rates of change, subject-specific frequency effects along with a subject-ear random intercept were included. To summarize, the final model contained 70 fixed effects (age, frequency, frequency2, service branch, cumulative noise exposure by military occupation, and interaction terms defined earlier) and two random effects (subject-specific frequency and subject-ear). Further details of the statistical approach are provided in the Supplemental Digital Content 1, http://links.lww.com/EANDH/A747.

Our aim was to estimate the average annual rate of pure-tone hearing threshold change as a consequence of accumulated exposure to military occupational noise. To do this, we modeled the effects of cumulative exposure to a military occupation on the average pure-tone threshold, so that the average change with 1 year of exposure to a military occupation is the difference between the model-based average pure-tone threshold with 1 year of exposure to a military occupation minus the model-based average threshold with no exposure. We estimated the 95% confidence intervals (CI) around the estimates of the average annual hearing threshold change. A 95% CI that includes zero is consistent with the null hypothesis of no pure-tone hearing threshold change. Estimates of hearing change are provided by individual frequency (500, 1000, 2000, 3000, 4000, 6000 Hz), service branch (Navy, Marine Corps, Army, Air Force), and occupational noise exposure ranking (low, moderate, high). A positive change indicates hearing sensitivity worsened and a negative change indicates hearing sensitivity improved.

RESULTS

Among the 246 study participants, most were men (87%), non-Hispanic White (68%), and young (median age 20 years at the time of enlistment, Table 2). In addition, over half of the analytic sample were in the Army (58%). The median duration of enlistment was 6 years and most, during that time, had 1 military occupational classification (82%). The median number of audiograms across 246 study participants was 5 (range 1 to 17). Seven individuals had only 1 audiogram, while 11 had 10 or more audiograms (Table 3). The total number of audiograms varied by service branch and year (Table S1 in Supplemental Digital Content 1, http://links.lww.com/EANDH/A747). Median number of audiograms among Army, Navy, Marine Corps, and Air Force Service members were 6 (range 2 to 17), 4 (range 2 to 14), 4 (range 1 to 10), and 4 (range 1 to 15), respectively. The median time between tests was 377 days but ranged from 1 to 3402 days.

TABLE 2. - Participant-level characteristics of analytic sample
Characteristic n = 246
Gender, n (%)
 Male 214 (87)
 Female 32 (13)
Age at service entry, yrs (median, range) 20 (17–40)
Race/ethnicity, n (%)
 Non-Hispanic White 168 (68)
 Other 72 (29)
 Declined to report/missing 6 (2)
Enlistment suration, yrs (median, range) 6.0 (1.5–15)
Branch of service, n (%)
 Army 142 (58)
 Marine 50 (20)
 Navy 35 (14)
 Air force 19 (8)
Service component, n (%)
 Active 204 (83)
 National guard 28 (11)
 Reserves 14 (6)
Number of deployments, n (%)
 0 23 (9)
 1 144 (59)
 2 40 (16)
 3 or more 39 (16)
Number of military occupations, n (%)
 1 202 (82)
 2 36 (15)
 3 or more 8 (3)
Number of audiograms, median (range) 5 (1–17)

TABLE 3. - Distribution of the number of audiograms for 246 military personnel
Number of Audiograms n (%) of Military Personnel
1 7 (2.9)
2 19 (7.7)
3 32 (13.0)
4 44 (17.9)
5 37 (15.0)
6 34 (13.8)
7 31 (12.6)
8 16 (6.5)
9 15 (6.1)
10 3 (1.2)
11 2 (0.8)
12 1 (0.4)
13 2 (0.8)
14 1 (0.4)
15 1 (0.4)
17 1 (0.4)

Figure 3 shows the estimated average annual rate of hearing threshold change by audiometric test frequency for 500 to 6000 Hz for each service branch; again, positive values suggest hearing sensitivity worsened and negative values suggest hearing sensitivity improved. Average annual rates of hearing threshold change are plotted by military occupational noise exposure rank and varied between −0.5 dB per year of exposure to 1.1 dB per year of exposure. The data along the x axis within each panel reports the estimated average hearing threshold change in dB/year of exposure for each service branch by test frequency. There were too few Air Force Service members in occupations with low or moderate noise exposure ranking to provide stables estimates of the average annual rate of hearing threshold change.

Fig. 3.
Fig. 3.:
Annual rate of hearing threshold changes by audiometric test frequency for each service branch and occupational noise exposure ranking. Filled circles, squares, and triangles represent exposure to an occupation with low, moderate, and high probability of noise exposure, respectively. Positive values indicate hearing declined and negative values indicate hearing improved. Vertical bars represent 95% confidence intervals. *Statistical significance.

A monotonic dose-response pattern in which increasing noise exposure ranking is associated with increasing hearing threshold change did not emerge. However, some patterns within service branches are notable. Among Marine Corps personnel, the annual rate of hearing threshold change was the greatest for individuals with military occupations with a high noise exposure ranking. For 1000 to 6000 Hz, the average annual rate of hearing threshold change increased from 0.4 (95% CI: 0.01 to 0.7) to 1.1 (95% CI: 0.4 to 1.8) dB per year. In the Army, the average annual rate of hearing threshold change was the greatest among personnel with exposure to a military occupation with a moderate noise exposure ranking. On average, Army personnel experienced hearing threshold change from 0.3 (95% CI: 0.05 to 0.6) to 0.6 (95% CI: 0.1 to 1.0) dB per year from 3000 to 6000 Hz. In addition, Army personnel with a high noise exposure ranking demonstrated elevated hearing threshold changes at 500 Hz (0.4 dB per year; 95% CI: 0.2 to 0.6). Navy personnel with a high noise exposure ranking demonstrated a threshold change of −0.5 (95% CI: −1.0 to −0.03) dB per year at 2000 Hz, an improvement in thresholds over time. No statistically significant (all CIs include zero) hearing threshold changes were observed at any test frequency among Air Force study participants.

Sensitivity Analysis

Across all audiograms (n = 2615), only 1 (0.04%) was missing data. One audiogram was missing a single pure-tone threshold value at 6000 Hz and this missing value is very unlikely to affect the results. Improbable pure-tone thresholds, defined as values that differed by 50 dB in comparison to adjacent threshold values, only occurred in 0.3% of audiograms (9/2615) across 4 study participants. These audiograms were compared over time within an individual and in all cases the unlikely threshold value was observed on more than one occasion. Furthermore, improbable thresholds were only found at 4000 and 6000 Hz, which suggests these pure-tone thresholds were not testing errors. Interaural differences ≥40 dB at the same frequency were also flagged for review because it is presumed that noise exposures rarely cause strongly sided asymmetric hearing loss (Masterson et al. 2013). While this may be true for nonmilitary workplace noise exposures, this pattern of hearing loss could emerge with exposure to military noise (Joseph et al. 2020). We observed interaural differences in 1.6% of audiograms (43/2615) across 15 study participants. Review of these audiograms revealed asymmetries that were often repeatedly observed over time within a study participant, suggesting they too were not testing errors. Finally, we reviewed audiograms flagged with a rising slope between 500 and 2000 Hz, which can be an indication of excessive ambient background noise. 2.8% of audiograms (73/2615) across 44 study participants were flagged as having a rising slope. We removed these audiograms from our analysis and re-ran the model. The results did not appreciably change (data not shown).

DISCUSSION

This analysis describes the longitudinal progression of hearing threshold change for 246 individuals enrolled in the NOISE Study using audiometric data extracted from the DOEHRS-HC data repository. Joining the NOISE Study data with the DOEHRS-HC data provided an opportunity to examine associations between military occupational noise exposure and hearing over time during an initial period of service. Individuals were followed during the course of their initial service period which ranged from 1.5 to 15 years (median service duration was 6 years).The average annual rate of hearing threshold change varied by service branch, military occupational noise exposure ranking (low, moderate, and high), and audiometric frequency. Determining the rate of hearing threshold change and potential risk factors that affect the rate of change is important to elucidate the impact of military service on hearing and to guide prevention strategies and subsequent care. To our knowledge, this unique application of DOEHRS-HC data is the first analysis of hearing threshold change over time in US Service members and contributes to the limited literature on longitudinal effects.

We estimated that the average rate of hearing threshold change varied between −0.5 and 1.1 dB per year when stratified by service branch, military noise exposure ranking, and audiometric test frequency. Generally, the higher test frequencies demonstrated the greatest average annual rates of hearing threshold change consistent with noise exposure. However, linear dose-response associations were not observed. Army occupations with a moderate noise ranking, and Marine Corps occupations with a high noise ranking, demonstrated the highest average rates of hearing threshold change—the latter more so than the former. On average, Navy and Air Force personnel did not display meaningful hearing threshold changes; the exception being Navy personnel with a high noise exposure rank demonstrating an improvement at 2000 Hz. The difference in hearing threshold degradations across the military branches is likely indicative of the varying types of noise these branches are exposed to and, it is important to note, the varying policies on hearing protection device use and enforcement (US Government Accountability Office 2011). The observed improvement in hearing thresholds over time among Navy personnel was an unexpected finding. Further longitudinal research examining hearing among this service branch is warranted.

There is a limited literature base with which we can compare our results. Only one study in Active-Duty military personnel estimated the annual rate of hearing threshold change. Collée et al. (2019) found, among Belgian military personnel, hearing threshold levels averaged at 3000, 4000, and 6000 Hz increased slowly over a 6-year period (0.08 dB/year) and noted some additional increases in the annual rate of change due to moderate and high probability of noise exposure. No longitudinal study of US military personnel currently reports the average annual rate of hearing threshold change by frequency for comparison. One longitudinal study conducted in Veterans is available for comparison: The VA Normative Aging Study (NAS; Echt et al. 2010). The VA NAS enrolled healthy Veteran men, free of known chronic medical conditions, between 1961 and 1970. In this study, Veterans aged 21 to 81 at enrollment were followed from 1962 to 1996 and their average annual rate of change was reported by frequency and age group. The average rate of hearing threshold change in the 30 to 39 year age group (the closest age group to our study that was reported) ranged from about 0.1 dB/year to just under 0.7 dB/year for frequencies 500 to 6000 Hz (Echt et al. 2010). Generally, our estimates of hearing threshold change across frequencies, service branches, and military occupational noise exposure rankings were less than what was reported by The VA NAS. Since 1970, the end of study enrollment into the VA NAS, concerted efforts have attempted to improve the DoD HCP. The difference between the VA NAS and our estimates of the average annual rate of hearing threshold change could be the product of improved hearing conservation efforts between study periods. However, despite potential improvements in hearing conservation, the estimated average annual rates of hearing threshold change among Marine Corps personnel exposed to military occupations with a high noise exposure ranking enrolled in the NOISE Study exceeded the published estimates of the 30- to 39-year-old group from the NAS. Our estimates of annual change are closely aligned with those among Veterans aged 50 to 59 in the VA NAS. This suggests that the hearing thresholds of Marine Corps personnel in an occupation with a high noise exposure ranking is declining at a faster rate than what would be expected for their age group.

The lack of a dose response may not be surprising when considering hearing threshold changes in the presence of hearing protection device use. A report by the US Government Accountability Office (2011) noted that inconsistencies in various military strategies for hearing protection and the lack of adequate surveillance and oversight placed military personnel unnecessarily at risk for hearing loss. Since 2006, the Army has enrolled all of their military personnel into the HCP, which includes annual audiometric surveillance and hearing protection device fitting. Thus, Army personnel exposed to occupations with a moderate noise ranking are being surveilled and fit with hearing protection who otherwise might not have been enrolled in an HCP based on high noise dosimetry measures (i.e., ≥85 dBA).

Service members who are exposed to moderate amounts of noise may feel that hearing protection devices provide excessive attenuation, leading to a decrease in their ability to communicate and perform job duties. Thus, compliance with hearing protection may vary based on noise exposure and, at moderate noise levels, hearing protection use may not be enforced. This patterned outcome may be expressed in the population-level surveillance currently used by the Army. At least one study of a nonmilitary, industrial occupational cohort indicated the lack of a dose response was likely related to varying hearing protection use among the different occupational noise levels and noted that hearing protection use may not be required or enforced by management at moderate levels (Rabinowitz et al. 2007). By comparison, those in occupations with high noise exposure rankings may find that use of hearing protection results in less fatigue, pain, tinnitus, and hearing threshold shifts and thus may be more compliant and consistent with hearing protection use. This suggests hearing protection fitting strategies may need to be altered to provide adequate protection without excessive attenuation and that continued efforts should be made to improve on-site earplug selection, fit-testing, and training in the use of individually fitted protection.

Our results must be interpreted against the backdrop of a changing HCP and auditory surveillance strategies by service branches. During the time period examined (2001 to 2017), the vast majority (>90%, Table S1 in Supplemental Digital Content 1, http://links.lww.com/EANDH/A747) of Army audiograms occurred after 2006 when the Army decided to enroll all Service members in the HCP. Thus, the Army results are more likely to represent population-level hearing health behaviors and hearing threshold changes over time compared with the other service branches. While the Marine Corps launched their population-level audiometric surveillance in 2012, many audiograms examined in this study were collected before this policy change (67%, Table S1 in Supplemental Digital Content 1, http://links.lww.com/EANDH/A747). Consequently, the results from the Marine Corps are likely reflecting more of the risk-based auditory monitoring strategy employed by the Armed Services for many years. Differing surveillance/monitoring strategies thus reflect different populations of interest and may highlight differences in exposure and hearing protection use patterns, helping explain the lack of a dose-response association.

Strengths, Limitations, and Future Directions

This study has several strengths. First, the study used repeat audiograms over time and the retrospective assignment of noise exposure rankings to aid in our understanding of the impacts of occupational noise on hearing during military service. Second, in analyzing the rate of hearing threshold change by frequency, we have expanded our understanding of how slowly or rapidly hearing deteriorates at a population-level and within different service branches. Therefore, a deeper understanding of the time course and rate of hearing threshold changes, and the factors that influence hearing threshold changes, will aid the DoD in designing more effective hearing loss prevention interventions. Finally, the hierarchical linear modeling approach allowed us to utilize all available data.

Our study is not without limitation. The main limitation with job exposure matrices is that specific risk factors cannot be clearly identified. For example, Service members are exposed to noise, but they may also be exposed to solvents, inhalants, medications, and other stressors in the workplace. In addition, if only some personnel within a military occupation are exposed to noise while others are not, the heterogeneity within the group may mask the ability to observe associations between noise and hearing. With these limitations in mind, an observed elevated risk in hearing threshold change in a military occupation only suggests risks from noise exposure and more precise etiologic questions relating to hearing outcomes remain. Taken together, our study results may not be generalizable to the entire Armed Services. Our results are also subject to residual confounding, namely the potential for high levels of recreational noise exposure to travel with military occupations with high noise rankings (Flamme et al. 2012); although, this is an unlikely explanation for our results given only some service branches demonstrated hearing declines and not always among occupations with the highest noise exposure ranking.

A core strength of this study was the longitudinal assessment of audiometric data. However, despite review by an audiologist, the DOEHRS-HC audiograms could have errors. We examined potentially problematic audiograms, removed them from the dataset and repeated our analysis to which our results and conclusions did not change suggesting our findings are robust to administrative data errors.

In the future, attention should be given to the development of a more robust military JEM that adjusts for hearing protection use and incorporates noise dosimetry measures when available. This would strengthen the ability to epidemiologically assess the association between military occupational noise and hearing, which would be instrumental to uncovering how early noise exposures, including noise exposures that induce permanent hearing shifts but remain within the normal range, as well as hearing loss, impact hearing later in life. It would also be prudent for future investigations to include measures of recreational noise exposures outside of the military and adjust for these additional exposures. Moreover, future study should include a larger and more diverse group of military personnel, including military personnel with low noise exposures.

CONCLUSIONS

This study used the audiometric data of military personnel collected as part of the DoD HCP to determine the average change in audiometric thresholds by frequency over time and estimated how military occupational noise exposure altered that trajectory. This was, to our knowledge, the first analysis of hearing threshold changes over time using such data and adds to the limited literature on frequency-specific hearing outcomes during military service. Based on this analysis, we estimate that some military personnel are at increased risk for hearing threshold changes due to occupational noise exposure and hearing sensitivity may be declining at faster rates than would be expected for their age group. Estimates of the amount of hearing threshold change by frequency, and risk factors that impact hearing inform DoD efforts to protect Service members’ hearing during their military service.

ACKNOWLEDGMENTS

This study was supported by a Department of Defense Congressionally Directed Medical Research Program Investigator-Initiated Research Award (PR121146), a Joint Warfighter Medical Research Program Award (JW160036), and a U.S. Department of Veterans Affairs (VA) Rehabilitation Research and Development (RR&D) Research Career Scientist Award (1 IK6 RX002990-01). This material is the result of work supported with resources and the use of facilities at the VA RR&D National Center for Rehabilitative Auditory Research (VA RR&D NCRAR Center Award; C9230C) at the VA Portland Health Care System in Portland, Oregon, as well as the Department of Defense, Hearing Center of Excellence in San Antonio, TX.

REFERENCES

Abel S. M. Hearing loss in military aviation and other trades: Investigation of prevalence and risk factors. Aviat Space Environ Med, (2005). 76, 1128–1135.
American National Standards Institute. Specification for Audiometers, S3.6-1996.1996).
American National Standards Institute. Specification for Audiometers, S3.6-2004.2004).
American National Standards Institute. Specification for Audiometers, S3.6-2010.2010).
Appollonio I., Carabellese C., Frattola L., Trabucchi M. Effects of sensory aids on the quality of life and mortality of elderly people: A multivariate analysis. Age Ageing, (1996). 25, 89–96.
Cacciatore F., Napoli C., Abete P., Marciano E., Triassi M., Rengo F. Quality of life determinants and hearing function in an elderly population: Osservatorio Geriatrico Campano Study Group. Gerontology, (1999). 45, 323–328.
Carabellese C., Appollonio I., Rozzini R., Bianchetti A., Frisoni G. B., Frattola L., Trabucchi M. Sensory impairment and quality of life in a community elderly population. J Am Geriatr Soc, (1993). 41, 401–407.
Carhart R., Jerger J. Preferred method for clinical determination of pure-tone thresholds. J Speech Hear Disorders, (1959). 24, 330–345.
Collée A., Watelet J. B., Vanmaele H., Van Thielen J., Clarys P. Longitudinal changes in hearing threshold levels for noise-exposed military personnel. Int Arch Occup Environ Health, (2019). 92, 219–226.
Dalton D. S., Cruickshanks K. J., Klein B. E., Klein R., Wiley T. L., Nondahl D. M. The impact of hearing loss on quality of life in older adults. Gerontologist, (2003). 43, 661–668.
Department of Defense. DoD Instruction 6055.12 “DoD Hearing Conservation Program (HCP)”. (2004). https://biotech.law.lsu.edu/blaw/dodd/corres/pdf/i605512_030504/i605512p.pdf. Accessed October 28, 2019.
Department of Defense. DoD Instruction 6055.12 “DoD Hearing Conservation Program (HCP)”. (2010). https://www.med.navy.mil/sites/nmcphc/Documents/oem/dodi-6055-12.pdf. Accessed October 28, 2018.
Department of Defense. DoD Instruction 6055.12 “DoD Hearing Conservation Program (HCP)”. (2019, August 14). https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodi/605512p.pdf?ver=2019-08-14-073309-537. Accessed October 28, 2019.
Department of Defense Hearing Conservation Working Group. Hearing Health Surveillance Data Review Military Hearing Conservation - CY18. (2020). Department of Defense Hearing Center of Excellence.
Department of Veterans Affairs (DVA). Veterans Benefits Administration Annual Benefits Report Fiscal Year 2018. (2018). https://benefits.va.gov/REPORTS/abr/docs/2018-abr.pdf.
Echt K. V., Smith S. L., Burridge A. B., Spiro A. 3rd. Longitudinal changes in hearing sensitivity among men: The Veterans Affairs Normative Aging Study. J Acoust Soc Am, (2010). 128, 1992–2002.
Flamme G. A., Stephenson M. R., Deiters K., Tatro A., van Gessel D., Geda K., Wyllys K., McGregor K. Typical noise exposure in daily life. Int J Audiol, (2012). 51(Suppl 1), S3–11.
Gates G. A., Schmid P., Kujawa S. G., Nam B., D’Agostino R. Longitudinal threshold changes in older men with audiometric notches. Hear Res, (2000). 141, 220–228.
Genther D. J., Betz J., Pratt S., Martin K. R., Harris T. B., Satterfield S., Bauer D. C., Newman A. B., Simonsick E. M., Lin F. RHealth, Aging and Body Composition Study. Association between hearing impairment and risk of hospitalization in older adults. J Am Geriatr Soc, (2015). 63, 1146–1152.
Genther D. J., Frick K. D., Chen D., Betz J., Lin F. R. Association of hearing loss with hospitalization and burden of disease in older adults. JAMA, (2013). 309, 2322–2324.
Henry J. A., Griest S., Reavis K. M., Grush L., Theodoroff S. M., Young S., Thielman E. J., Carlson K. F. Noise Outcomes in Servicemembers Epidemiology (NOISE) Study: Design, methods, and baseline results [published online ahead of print December 16, 2020]. Ear Hear. (2020). doi:10.1097/AUD.0000000000000974.
Joseph A. R., Shaw J. L., Clouser M. C., MacGregor A. J., Dougherty A. L., Galarneau M. R. Clinical audiometric patterns of hearing loss following blast-related injury in U.S. military personnel. Int J Audiol, (2020). 59, 772–779.
Kujawa S. G., Liberman M. C. Acceleration of age-related hearing loss by early noise exposure: Evidence of a misspent youth. J Neurosci, (2006). 26, 2115–2123.
Lin F. R. Hearing loss and cognition among older adults in the United States. J Gerontol A Biol Sci Med Sci, (2011). 66, 1131–1136.
Lin F. R., Ferrucci L. Hearing loss and falls among older adults in the United States. Arch Intern Med, (2012). 172, 369–371.
Lin F. R., Ferrucci L., Metter E. J., An Y., Zonderman A. B., Resnick S. M. Hearing loss and cognition in the Baltimore Longitudinal Study of Aging. Neuropsychology, (2011). 25, 763–770.
Masterson E. A., Tak S., Themann C. L., Wall D. K., Groenewold M. R., Deddens J. A., Calvert G. M. Prevalence of hearing loss in the United States by industry. Am J Ind Med, (2013). 56, 670–681.
Mulrow C. D., Aguilar C., Endicott J. E., Tuley M. R., Velez R., Charlip W. S., Rhodes M. C., Hill J. A., DeNino L. A. Quality-of-life changes and hearing impairment. A randomized trial. Ann Intern Med, (1990). 113, 188–194.
Nelson J. T., Swan A. A., Swiger B., Packer M., Pugh M. J. Hearing testing in the U.S. Department of Defense: Potential impact on Veterans Affairs hearing loss disability awards. Hear Res, (2017). 349, 13–20.
Occupational Safety and Health Administration. Occupational noise exposure standard (29 CFR 1910.95). (1983). https://www.osha.gov/laws-regs/regulations/standardnumber/1910/1910.95. Accessed October 28, 2018.
Powers R. Basic Training for Dummies. (2011). Wiley Publishing.
Rabinowitz P. M., Galusha D., Dixon-Ernst C., Slade M. D., Cullen M. R. Do ambient noise exposure levels predict hearing loss in a modern industrial cohort? Occup Environ Med, (2007). 64, 53–59.
Reed N. S., Altan A., Deal J. A., Yeh C., Kravetz A. D., Wallhagen M., Lin F. R. Trends in health care costs and utilization associated with untreated hearing loss over 10 years. JAMA Otolaryngol Head Neck Surg, (2019). 145, 27–34.
Saunders G. H., Griest S. E. Hearing loss in veterans and the need for hearing loss prevention programs. Noise Health, (2009). 11, 14–21.
Singer J. D., Willett J. B. Applied Longitudinal Data Analysis: Modeling Change and Event Occurence. (2003). Oxford University Press.
Teschke K., Olshan A. F., Daniels J. L., De Roos A. J., Parks C. G., Schulz M., Vaughan T. L. Occupational exposure assessment in case-control studies: opportunities for improvement. Occup Environ Med, (2002). 59, 575–593.
US Government Accountability Office. Hearing loss prevention: Improvements to DOD hearing conservation programs could lead to better outcomes. (2011). https://www.gao.gov/new.items/d11114.pdf.
Yankaskas K. Prelude: Noise-induced tinnitus and hearing loss in the military. Hear Res, (2013). 295, 3–8.
Yong J. S., Wang D. Y. Impact of noise on hearing in the military. Mil Med Res, (2015). 2, 6.
Keywords:

Audiometry; Longitudinal studies; Military personnel; Noise-induced hearing loss; Occupational noise; Pure-tone thresholds

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved