Annals of Otology, Rhinology & Laryngology 121(12):771-775.

© 2012 Annals Publishing Company. All rights reserved.

Association of Hearing Loss With Decreased Employment and

Income Among Adults in the United States
David Jung, MD, PhD; Neil Bhattacharyya, MD

Objectives: We evaluated the association of hearing loss with employment and income in adults.
Methods: Patients with a coded diagnosis of hearing loss were identified from the 2006 and 2008 Medical Expenditure
Panel Survey linked household and medical conditions files and compared to patients without hearing loss. Differences
in employment, wage income, and Supplemental Security Income were evaluated with multivariate regression models
after adjustment for several demographic and Charlson comorbidity variables.
Results: An estimated 933,921 ± 88,474 adults were identified with hearing loss (54.7% of whom were male; mean age
for all, 51.0 years). Patients with hearing loss were more likely to be unemployed or partly unemployed than those with-
out hearing loss (adjusted odds ratio, 2.2; p < 0.001). Similarly, adults with hearing loss were less likely to have any wage
income than those without hearing loss (adjusted odds ratio, 2.5; p < 0.001). The population with hearing loss earned a
mean wage of $23,481 ± $3,366, versus $31,272 ± $517 for the population without hearing loss (difference in wages,
$7,791; p < 0.001). The association between hearing loss and receiving Supplemental Security Income was not signifi-
cant (p = 0.109).
Conclusions: Adults with hearing loss are more likely to be unemployed and on average earn significantly less wage
income than adults without hearing loss. Further work is needed to determine the potential impact of treatment on these
differences.
Key Words: employment, hearing loss, income, wage.

INTRODUCTION particular advantage of these data is that survey in-

Hearing loss can profoundly affect a person’s terview information is supplemented by information
ability to interact with his or her surroundings. This collected directly from medical providers caring
is likely to have consequences both in and out of for the survey participants. Furthermore, insurance
the workplace and may ultimately have economic data are collected both from survey participants and
implications for people affected with hearing loss from additional surveys given to US employers re-
and their families. Furthermore, hearing-impaired garding provision of health-care benefits. As such,
workers experience higher levels of stress, expend MEPS data have been used to evaluate costs associ-
increased effort in listening at work, and tend to ated with chronic conditions such as asthma, diabe-
take more sick days as a result of stress-related com- tes, hypertension, chronic rhinosinusitis, and aller-
plaints.1,2 However, no study to date has directly ex- gic rhinitis.5-9 In this study, we examined the effect
amined the effect of hearing loss on employment of hearing loss on wages for adults in the United
and personal income for working-age adults. Such States. We hypothesized that hearing loss in adults
an analysis is especially timely given the increasing would be associated with a decreased ability to earn
prevalence of hearing loss among adolescents, who income, as based on employment status and magni-
themselves represent future wage-earning adults in tude of wage income. Secondarily, we sought to de-
the United States.3 termine whether hearing loss was associated with an
increased likelihood of an adult’s receiving Supple-
The Medical Expenditure Panel Survey (MEPS) mental Security Income, a form of disability income
is a representative sample of the noninstitutional- from the US Social Security Administration.
ized civilian population of the United States and
collects data on approximately 15,000 families and METHODS
40,000 individuals over a 2-year panel cycle.4 One The data source for this study was the MEPS for
From the Department of Otology and Laryngology, Harvard Medical School (both authors), the Massachusetts Eye and Ear Infirmary/
Harvard Otolaryngology Training Program (Jung), and the Division of Otolaryngology, Brigham and Women’s Hospital (Bhattacha-
ryya), Boston, Massachusetts. Dr Bhattacharyya serves as a consultant for Intersect-ENT, Inc, and Entellus, Inc, in unrelated enter-
prises.
Correspondence: David Jung, MD, PhD, Massachusetts Eye and Ear Infirmary, 243 Charles St, Boston, MA 02114.

771

Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 10, 2015
772 Jung & Bhattacharyya, Decreased Employment & Income Among Hearing Loss Patients 772
calendar years 2006 and 2008 as administered by
the Agency for Healthcare Research and Quality un-
der the US Department of Health and Human Ser-
vices. This study was reviewed by our hospital’s
committee on clinical investigations and designat-
ed as exempt from review. In accordance with data
access agreements, precautions were taken to avoid
individual patient identification. Using SPSS ver-
sion 17.0 (Chicago, Illinois), we merged the 2006
and 2008 medical conditions file components of the
MEPS with the corresponding consolidated expen-
ditures files according to the appropriate linkage
variables. To obtain the most complete evaluation
of our patient population, we allowed as many as 34
medical condition variables per patient after the data
file merge. We restricted the data to adult patients
only (ages 18 to 64 years) for the remainder of the
analysis, as this adult group was the most likely to
be employed and earn wages over the study period.
Next, we identified adult patients as having hear-
ing loss by examining the ICD9CODX data field for
each listed medical condition. This field consists of
the 3-digit International Classification of Diseases
version 9 code for a given medical condition based
on an assessment of medical conditions recorded
as verbatim text and then coded by a professional
coder. Patients for whom any one of the 34 medical
condition codes corresponded to “389.x” were con-
sidered as reporting hearing loss, and the remainder
were considered as not reporting hearing loss.
We compiled standard demographic information,
including age and sex distribution, for patients who
reported hearing loss and those who did not. The
additional demographic information extracted and
compared included race, ethnicity, level of educa-
tion attained, insurance status, geographic region,
and marital status. Additionally, we calculated the
D’Hoore et al10 adaptation of the Charlson comor-
bidity index for the comorbid medical conditions of
each subject. This index adjusts for such comorbidi-
ties as myocardial infarction, congestive heart fail-
ure, peripheral vascular disease, dementia, cerebro-
vascular disease, asthma, chronic obstructive pul-
monary disease, peptic ulcer disease, and liver dis-
ease and allows patients to be compared with “like”
patients with similar comorbidities.9
First, the association of hearing loss with em-
ployment versus unemployment status and with the
presence or absence of earned wage income was
determined with a multivariate logistic regression
adjusting for age, sex, race, city, educational level
achieved, insurance status, region, marital status,
and Charlson comorbidity index score with statisti-
cal significance (p value) set at 0.05. Next, selecting
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 10, 2015
only adults with positive wage income (ie, actively
employed), we then compared differences in wage
income between adult patients with and without
hearing loss by using a general multivariate linear
model adjusting for age, sex, race, ethnicity, educa-
tion, insurance, region, marital status, and Charlson
comorbidity index score. Wage income was loga-
rithmically transformed to fulfill normal distribution
criteria for multivariate regression.
Data are reported as mean or proportion plus or
minus the standard error of the estimate. Because
the sample design of the MEPS survey includes
stratification, clustering, multiple stages of selec-
tion, and disproportionate sampling, we utilized ap-
propriate complex survey design statistical methods
to determine nationally representative means and
standard errors. Estimates were considered reliable
if the relative standard error of the estimate was less
than 30%.
RESULTS
As shown in the Table, an estimated 933,921 ±
88,474 adults in the United States were identified
with a diagnosis of hearing loss; they had an aver-
age age of 51.0 ± 1.0 years and a slight male pre-
dominance (54.7% ± 4.4%). This constitutes ap-
proximately 0.5% of the US population between the
ages of 18 and 64 years. The Table further summa-
rizes the demographic differences between patients
who reported hearing loss and those who did not;
these demographic and Charlson comorbidity index
differences were used for the adjusted multivariate
analysis.
With respect to employment, patients with hear-
ing loss were more likely to be either unemployed
or partly employed than adults without hearing loss,
with an adjusted odds ratio of 2.2 (95% confidence
interval, 1.4 to 3.4; p < 0.001). Similarly, adults with
hearing loss were less likely to have any form of
wage income than those without hearing loss, with
an adjusted odds ratio of 2.5 (95% confidence inter-
val, 1.6 to 3.9; p < 0.001). With respect to wage in-
come, adults with hearing loss earned a mean wage
of $23,481 ± $3,366, versus $31,272 ± $517 for
the population without hearing loss (net difference
in wages, $7,791; p < 0.001). Finally, there was no
statistically significant association between hearing
loss and having Supplemental Security Income (p =
0.109).
DISCUSSION
This study analyzed household-based, nationally
representative MEPS data sets to examine the rela-
tionship between hearing loss and income for adults
773 Jung & Bhattacharyya, Decreased Employment & Income Among Hearing Loss Patients 773

Comparison of Demographic Characteristics Between Patients With and Without Hearing Loss
Patients With Patients With
Hearing Loss
No Hearing Loss Univariate
Characteristic Value SE Value SE p Value
Population size 933,921 88,474 186,704,373 3,105,481 NA
Mean age (y) 51.0 1.0 40.4 0.1 0.109
Gender (%) 0.235
  Male 54.7 4.4 49.4 0.2
  Female 45.3 4.4 50.6 0.2
Race (%) 0.092
  White 86.6 2.5 80.3 0.6
  Black 8.6 2.1 12.3 0.5
  Other 4.9 1.4 7.3 0.3
Ethnicity (%) <0.001
  Hispanic 4.6 1.3 14.9 0.6
  Non-Hispanic 95.4 1.3 85.1 0.6
Education (%) 0.002
  No degree 11.0 2.6 15.2 0.3
  High-school equivalency test (GED) 4.5 1.5 4.1 0.1
  High-school diploma 37.8 4.3 45.1 0.5
  Bachelor’s degree 14.9 3.0 18.2 0.4
  Master’s degree 15.4 4.1 6.9 0.2
  Doctorate 2.7 1.2 1.9 0.1
  Other degree 13.8 3.1 8.5 0.2
Insurance (%) 0.030
  Uninsured 11.1 2.6 17.9 0.4
  Private 75.5 3.4 72.7 0.4
  Public 13.4 2.6 9.4 0.3
Region (%) 0.632
  Northeast 21.2 3.8 18.3 0.6
  Midwest 20.9 4.0 22.0 0.7
  South 31.5 4.6 36.2 0.8
  West 26.4 3.9 23.5 0.7
Marital status (%) 0.030
  Presently married 64.0 4.3 54.2 0.4
  Presently not married 36.0 4.3 45.8 0.4
Mean Charlson comorbidity index score 0.82 0.12 0.30 0.01 <0.001
Analysis limited to adults 18 to 64 years of age.
SE — standard error of the estimate; NA — not applicable.

in the United States. After adjusting for multiple de-
mographic variables and the Charlson comorbid-
ity score, we found that patients with a diagnosis of
hearing loss are more likely to be unemployed or
partly employed and earn, on average, significant-
ly lower wages than patients without hearing loss.
The difference in wage income, which approaches
$8,000 per year, represents a nearly 25% decrease
in wage income relative to patients without hearing
loss. Perhaps the most striking finding was the odds
ratio (2.2) for unemployment for working-age adults
with hearing loss.
MEPS data, which have commonly been used to
evaluate health-care costs associated with various
diagnoses, are also useful for evaluating demograph-
ic factors associated with different chronic medical
conditions. Given the inaccuracy with which the
typical person who is not a medical professional
may describe his or her medical history, a particular
advantage of the MEPS data is that diagnostic cod-
ing is done by trained health-care providers. MEPS
data are considered the gold standard for examining
health-care costs, insurance costs, and income fac-
tors associated with medical conditions in the Unit-
ed States. We, and others, have used MEPS data to
quantify expenditures for various illnesses in the
United States.5-9 Other investigators have also used
MEPS data to look at the influence of chronic dis-
ease, such as rheumatoid arthritis, on wage-earning

Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 10, 2015
774 Jung & Bhattacharyya, Decreased Employment & Income Among Hearing Loss Patients 774
capacity.11 In that study, rheumatoid arthritis was as-
sociated with an expected loss of $8,957 in annual
earnings.
Previous studies have suggested an associa-
tion between hearing loss and decreased income.
A study of Swedish workers found that nearly one
third of this population described problems with tin-
nitus, hearing, or both on self-reported surveys. In
this study, respondents who endorsed difficulties
with hearing also tended to self-report lower socio-
economic status.12 Similarly, a comparison between
two large cross-sectional studies carried out in Nor-
way demonstrated that lower socioeconomic status,
as measured by education level and type of work,
is associated with hearing loss.13 Finally, a popula-
tion-based cohort study of a small US community
showed that progression of hearing loss over a 10-
year follow-up period was significantly correlated
with lower socioeconomic status, although, interest-
ingly, it was not significantly correlated with a his-
tory of noise exposure at work.14 Our study extends
these previous findings to the entire US population
by demonstrating and quantifying a significant asso-
ciation between hearing loss in adults and decreased
wages. To our knowledge, this is the first study to
broadly examine this association across the United
States, which as a nation manifests great econom-
ic diversity and a lack of centralized, government-
sponsored medical care and records.
These data, although broadly based and likely
comprehensive, do have several limitations. First,
our method of identifying patients with hearing
loss via ICD-9 code, although efficient, does not al-
low for distinction between unilateral and bilateral
hearing losses; this is important, because a bilater-
al hearing loss would be expected to have greater
consequences at work. In addition, our data do not
provide information regarding the severity of hear-
ing loss for any given patient or group. Further, the
subtypes of hearing loss, which are indicated in the
fourth and fifth digits of the ICD9 code, were not
REFERENCES
1. Hasson D, Theorell T, Wallén MB, Leineweber C, Can-
lon B. Stress and prevalence of hearing problems in the Swedish
working population. BMC Public Health 2011;11:130.
2. Kramer SE, Kapteyn TS, Houtgast T. Occupational per-
formance: comparing normally-hearing and hearing-impaired
employees using the Amsterdam Checklist for Hearing and
Work. Int J Audiol 2006;45:503-12.
3. Shargorodsky J, Curhan SG, Curhan GC, Eavey R.
Change in prevalence of hearing loss in US adolescents. JAMA
2010;304:772-8.
4. Cohen JW, Cohen SB, Banthin JS. The Medical Expen-
diture Panel Survey: a national information resource to support
healthcare cost research and inform policy and practice. Med
Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 10, 2015
possible to discern, because the MEPS removes the
fourth and fifth digits to preserve patient confiden-
tiality. Finally, we could not identify which patients
with reported hearing loss were being treated with
assistive devices (eg, hearing aids or cochlear im-
plants) or had received other treatments in the past
(eg, tympanostomy tube placement or ossiculoplasty
for conductive hearing loss). Successful hearing re-
habilitation would be expected, however, to render
our findings more significant, given that hearing re-
habilitation would likely offset the wage impact to
some degree.
Given these limitations, it is certainly premature
to suggest changes to health-care policy or insur-
ance coverage based on the present findings. How-
ever, further studies are warranted, particularly giv-
en the unemployment differentials and the magni-
tude of the decrease in wage income for adults with
hearing loss. These data highlight the need to further
characterize the type of hearing loss, as well as the
actual diagnosis rate and treatment rate for hearing
loss, in wage-earning adults. Identifying which sub-
groups of patients are most disadvantaged by a di-
agnosis of hearing loss could be the goal of more fo-
cused studies. Such studies may also provide insight
into which types of hearing loss are the most rele­
vant with respect to employment and wages, there-
by highlighting specific treatment strategies. More-
over, these data raise the question of whether young
adults with hearing loss, if such loss is treatable by
hearing aids or surgery, might realize significant im-
provements in wage income over their subsequent
lifetimes.
CONCLUSIONS
In the United States, hearing loss is statistically
significantly associated with a higher likelihood of
unemployment and decreased wage income in adults.
Future studies are needed to determine which treat-
ments may best address these discrepancies, given
their significant indirect costs to society.
Care 2009;47(suppl 1):S44-S50.
5. Brown JB, Nichols GA, Glauber HS, Bakst AW. Type 2
diabetes: incremental medical care costs during the first 8 years
after diagnosis. Diabetes Care 1999;22:1116-24.
6. Balu S, Thomas J III. Incremental expenditure of treating
hypertension in the United States. Am J Hypertens 2006;19:810-
7.
7. Bhattacharyya N. Incremental healthcare utilization and
expenditures for allergic rhinitis in the United States. Laryngo-
scope 2011;121:1830-3.
8. Bhattacharyya N. Incremental health care utilization and
expenditures for chronic rhinosinusitis in the United States. Ann
Otol Rhinol Laryngol 2011;120:423-7.
775 Jung & Bhattacharyya, Decreased Employment & Income Among Hearing Loss Patients 775

9. Kamble S, Bharmal M. Incremental direct expenditure of
treating asthma in the United States. J Asthma 2009;46:73-80.
10. D’Hoore W, Bouckaert A, Tilquin C. Practical consider-
ations on the use of the Charlson comorbidity index with admin-
istrative data bases. J Clin Epidemiol 1996;49:1429-33.
11. Sullivan PW, Ghushchyan V, Huang XY, Globe DR. In-
fluence of rheumatoid arthritis on employment, function, and
productivity in a nationally representative sample in the United
States. J Rheumatol 2010;37:544-9.
12. Hasson D, Theorell T, Westerlund H, Canlon B. Preva-
lence and characteristics of hearing problems in a working and
non-working Swedish population. J Epidemiol Community
Health 2010;64:453-60.
13. Helvik AS, Krokstad S, Tambs K. Socioeconomic in-
equalities in hearing loss in a healthy population sample: The
HUNT Study. Am J Public Health 2009;99:1376-8.
14. Cruickshanks KJ, Nondahl DM, Tweed TS, et al. Educa-
tion, occupation, noise exposure history and the 10-yr cumula-
tive incidence of hearing impairment in older adults. Hear Res
2010;264:3-9.

Downloaded from aor.sagepub.com at UCSF LIBRARY & CKM on April 10, 2015