Original Research—General Otolaryngology
The Association between Hearing Loss and Social Isolation in Older Adults Paul Mick, MD, MPH, FRCSC1, Ichiro Kawachi, MD, PhD2, and Frank R. Lin, MD, PhD3
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Abstract Objective. To determine if age-related hearing loss is associated with social isolation and whether factors such as age, gender, income, race, or hearing aid use moderated the association. Study Design. Cross-sectional. Setting. Randomly sampled United States communities. Subjects and Methods. Cross-sectional data on adults 60 to 84 years old from the 1999 to 2006 cycles of the National Health and Nutrition Examination Survey were analyzed. The dependent variable was social isolation, which was defined using the social isolation score (SIS), a 4-point composite index consisting of items pertaining to strength of social network and support. SIS scores 2 were considered indicative of social isolation. The independent (predictor) variable was the pure tone average of speech frequency (0.5-4 kHz) hearing thresholds in the better-hearing ear. Covariates included potential medical, demographic, and otologic confounders. We used multivariate logistic regression to evaluate the association between hearing loss and the odds of having social isolation. An exploratory analysis was performed to assess the strength of associations between hearing loss and individual items of the SIS scale. Results. Greater hearing loss was associated with increased odds of social isolation in women aged 60 to 69 years (odds ratio [OR], 3.49 per 25-dB of hearing loss; 95% confidence interval, 1.91, 6.39; P \ .001). Effect modification by gender was significant in this age group (P = .003). Hearing loss was not significantly associated with social isolation in other age and gender groups. Conclusions. Greater hearing loss is associated with increased odds of being social isolated in a nationally representative sample of women aged 60 to 69 years.
Keywords hearing loss, hearing disorders, presbycusis, aging, social isolation, social network, social support
Otolaryngology– Head and Neck Surgery 2014, Vol. 150(3) 378–384 Ó American Academy of Otolaryngology—Head and Neck Surgery Foundation 2013 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0194599813518021 http://otojournal.org
Received June 13, 2013; revised November 19, 2013; accepted December 4, 2013.
S
ocial networks refer to the web of interpersonal relationships that may provide an individual with various types of social support (eg, instrumental assistance, information, and emotional reinforcement).1 Persons who have a small social network or lack close relationships or sources of social support can be considered isolated. Social isolation has been shown to be a predictor of mortality, psychiatric illness, and cognitive and functional decline in elderly persons.2-7 Therefore, determining interventions and characteristics that promote social engagement in older persons is important. Previous research has implicated hearing loss in the development of social isolation. Chia et al, in a sample of older Australians, found that hearing impairment was associated with poorer scores in the social functioning and role limitation due to emotional problems domains of the Short Form– 36 scale (SF-36).8 Pronk et al9 reported that poorer selfreported hearing scores predicted deterioration in social support among a cohort of 900 senior citizens in Amsterdam. Population-based analyses have not been conducted in the United States. The current literature does not answer whether there are certain aspects of social networks and support more strongly associated with hearing impairment or if associations are modified by demographic factors or hearing aid use. In situations in which hearing is difficult, greater cognitive resources are needed to process auditory signals.10-12 Persons with even mild hearing loss have challenges communicating verbally in the presence of background noise. As a consequence, social gatherings may become difficult,
1
University of British Columbia, Vancouver, British Columbia, Canada Harvard School of Public Health, Boston, Massachusetts, USA 3 Departments of Otolaryngology–Head & Neck Surgery, Geriatric Medicine, Mental Health, and Epidemiology, Johns Hopkins University, Baltimore, Maryland, USA 2
This article was presented at the 2013 AAO-HNSF Annual Meeting & OTO EXPO; September 29–October 3, 2013; Vancouver, British Columbia, Canada. Corresponding Author: Paul Mick, MD, MPH, FRCSC, University of British Columbia, Vancouver, 202-3330 Richter St, Kelowna, BC, V1W 4H6, Canada. Email: [email protected]
Mick et al spontaneity might be reduced as events must be planned around acoustic environments, and relationships may suffer because of communication problems or because activities once enjoyed together are no longer pleasant to the person with hearing loss. Approximately 290 million people worldwide are hearing impaired, including 48% of Americans aged 60 or older.13,14 The prevalence is expected to increase in aging populations. Age-related hearing loss (ARHL) is a significant public health concern, compounded by underutilization of treatment options.15 The objectives of our study were 2-fold. First, we sought to determine if there was a cross-sectional association between hearing loss and social isolation in a nationally representative sample of Americans aged 60 to 69 and 70 to 84 years. Second, we sought to determine if the association was modified by gender, income, race, or hearing aid use.
Methods Data were publicly available and deidentified, and thus institutional ethics board approval was not required.
Study Cohort The National Health and Nutrition Examination Study (NHANES) is an annual cross-sectional survey designed to assess the health and nutrition of a sample of approximately 5000 non-institutionalized American adults and children. The study visits 15 randomly selected counties annually and employs a complex sampling procedure designed to oversample individuals who are older, younger, or members of minority groups. Each individual is assigned a sampling weight that accounts for the variable odds of selection and enables results to be generalized to the national population. Interviews, physical examinations, and laboratory tests are performed. For some items, only a subset of participants is tested. We used NHANES data from 1999 to 2006. Different age groups completed the social support questionnaire (SSQ) and audiometric testing in different years. Between 1999 and 2004, half of all participants between 20 and 69 years had their hearing tested, and all patients older than 60 years completed the SSQ. Therefore, 50% of all patients aged 60 to 69 years completed both and were included in the study. In the 2005-2006 NHANES cycle, only adults aged 70 years or older had hearing tests. They all completed the SSQ, and thus all 701-year-olds from the 2005-2006 cycle were included. The SSQ and audiometry were not administered concurrently in the same age group after 2006. NHANES classifies anyone older than 85 years as being ‘‘85 years old.’’ Patients who were 851 years old were thus excluded so that age could be used as a continuous variable in regression models.
Audiometric Assessment Trained examiners performed audiometric testing according to NHANES protocols.16 Air conduction thresholds were measured following a modified Hughson Westlake procedure.
379 Supra-aural earphones were used unless they caused external auditory canal collapse or there was hearing asymmetry, in which case they were replaced with insert earphones. Masking procedures were not performed. Speech-frequency pure tone averages (PTAs) were calculated for both ears as the mean of pure tone thresholds at 0.5, 1, 2, and 4 kHz. The lower PTA (ie, in the betterhearing ear) was defined as the independent variable and was analyzed as a continuous variable. All hearing thresholds in this article are reported as dB HL (American National Standards Institute, 2004).
Outcome Variable: Social Isolation The SSQ provides interview data on emotional, financial, and network support. Questions were selected from the Yale Health and Aging Study17 and the Social Network Index developed by Berkman et al in the Alameda County Study.3 Three items from the SSQ and marital status were combined into a summary measure of social isolation (the ‘‘social isolation score,’’ or SIS) that was tallied from 0 to 4, with higher scores indicating greater isolation. The first item asked whether the respondent was married or in a domestic partnership (0 = married or in partnership; 1 = not married or in partnership). The second item asked, ‘‘In general, how many close friends do you have?’’ (0 = at least 1 close friend; 1= no close friends). A third item characterized availability and adequacy of emotional support. A person scored 1 point if they had no one to provide emotional support (‘‘such as talking over problems or helping to make a difficult decision’’) or had at least one person to provide support, but ‘‘could have used more emotional support’’ in the past 12 months. The final item addressed the availability of financial support. The question asked, ‘‘If you need some extra help financially, could you count on anyone to help; for example, by paying any bills, housing costs, hospital visits, or providing you with food or clothes?’’ Those who answered ‘‘no’’ received 1 point. Participants were considered to be socially isolated if they were isolated according to 2 or more items (ie, SIS 2).
Other Study Variables Data on demographic variables (age, sex, race/ethnicity, education, annual household income), medical history (smoking status, cardiovascular disease, diabetes), noise exposure, and hearing aid use were obtained from interviews. Tests for multiplicative effect modification according to sex, income (low versus other), race (Mexican American or other Hispanic, and non-Hispanic black versus non-Hispanic white), and hearing aid use were planned a priori.
Statistical Methods Analyses were stratified according to age (60-69 and 70-84 years). The 1999-2004 and 2005-2006 NHANES data sets could not be merged because the ages of the hearing test participants during the 2 periods were mutually exclusive. We accounted for the complex sampling design in all multivariate analyses by weighting individual outcomes according to
380 National Center for Health Statistics guidelines. Nonweighted univariate assessments of the proportions with social isolation according to hearing status and other pertinent variables were performed using the x2 test. Multivariable logistic regression was used to determine odds ratios (ORs) for social isolation per 25-dB increment in PTA. The OR describes the multiplication of the odds of social isolation observed between persons with PTAs that differed by 25 dB. ORs .1 indicate increased odds of social isolation with hearing loss, while ORs \1 indicate lower odds with hearing loss. Multiplicative interaction terms between hearing loss and sex, income, race, and hearing aid use were added to the models and retained if their P values were \.05. An exploratory analysis was performed whereby each of the individual SSQ items were used as the dependent variable in multivariate models that were otherwise identical to the model used in the primary analysis. NHANES is a complex sample survey requiring mathematical variance approximation procedures to estimate sampling errors. The Taylor Series Linearization method was used as recommended by the National Centre for Health Statistics. NHANES participants eligible for hearing tests and the SSQ were excluded from the current study if they had any missing hearing, SSQ, or covariate data. A total of 18.4% of individuals were thus excluded. All analyses were performed with STATA 12.0 (StataCorp, College Station, TX), and a 2-sided P \ .05 was considered to be statistically significant.
Results Participant Characteristics There were 860 participants in the 60- to 69-year-old age group. Social isolation was prevalent in 19.8% of these individuals. In the 70- to 84-year-old age group, there were 593 participants, and social isolation was prevalent in 21.3%. Table 1 shows the raw (unweighted) characteristics of study participants. Table 2 displays the weighted proportions of socially isolated persons according to hearing status. In the 60- to 69-year-old group, 20.6% and 11.9% of persons with and without hearing loss, respectively, were isolated (P = .003). The corresponding proportions were 19.8% and 15.6% (respectively) in the 70- to 84-year-old group.
Multivariate Analysis Social isolation was independently associated with hearing loss in the 60- to 69-year-old group after adjusting for demographic, otologic, and medical covariates, and there was significant effect modification by gender (interaction term P value = .003). In women, the odds of social isolation increased 3.49 times with every 25-dB shift in PTA (95% confidence interval [CI]: 1.91-6.39, P \ .001). In contrast, the association in men was not significant (OR 1.11, 95% CI: 0.66-1.81, P = .68). In the 70- to 84-year-old age group, the association between hearing loss and social isolation was not significant, and there was no effect modification by gender (Table 3). Effect modification by history of hearing
Otolaryngology–Head and Neck Surgery 150(3) aid use, race, income, or education levels was not significant (data not shown).
Exploratory Analysis When the individual items were used as dependent variables in multivariate models, for 60- to 69-year-old women there were significant associations between hearing impairment and low emotional support and low financial support. The strongest association was with low emotional support. The P value for the association between hearing loss and being neither married nor in a domestic partnership was .059 (Table 4).
Discussion The primary analysis showed a strong association between hearing loss and social isolation in 60- to 69-year-old women but not in men or older individuals. In the exploratory analysis examining associations between hearing loss and components of the SIS, among 60- to 69-year-old women hearing loss was associated with reduced availability of emotional and financial support. This is the largest population-based study specifically investigating the association between hearing loss and social isolation. Other studies examined the impact of hearing loss on quality-of-life (QoL) metrics that contained measures of social and emotional function. Hearing loss was predictive of poorer scores in the social functioning and role limitation due to emotional problems domains of the SF-36 scale in a sample of older Australians. Effect modification by age and gender was not assessed, and regular hearing aid use did not have an impact on scores.8 A 2011 study reported a significant association between self-reported hearing loss and worsening emotional and social loneliness over a 4-year period among elderly Amsterdam residents.9 The extent to which men and women use verbal communication to obtain emotional support might, in part, explain the gender interaction in our data. Some sociologists have argued that talk is the primary vehicle by which women create and maintain intimacy and connectedness.18 In contrast, men may view dialogue as a way to accomplish instrumental tasks, convey information, and maintain autonomy in relationships.19 Burleson et al investigated differences in how men and women value communication skills among close friends and romantic partners. Women rated expressive skills such as ego support, conflict management, and comforting higher than men did, who gave greater weight to instrumental or interactional skills.20-23 The magnitude of the differences between genders in preferences for communication of emotional support is contested.24 The upshot is that if women rely more heavily on verbal communication, then hearing loss might impair their ability to receive emotional support to a greater degree than men. Other reasons may help explain the gender difference. Studies have shown that women are better at providing emotional support than men.22,25 Within heterosexual couples, men with hearing loss might therefore receive more
Mick et al
381
Table 1. Subject characteristics according to being socially isolateda or not isolated.b Social isolation Age 60-69 y Yes
PTA .25 dB in better earc Yes No P Sex Female Male P Race White Hispanic Black Other P Education High school incomplete High school graduate Some college College graduate P Annual household income \$20,000 $20,000-$44,999 $45,000 P History of heart disease Yes No P History of diabetes Yes No P History of smoking Yes No P Total
Age 70-84 y No
Yes
No
n
Col %
n
Col %
n
Col %
n
Col %
62 108 .014
36.5 63.5
186 504
27.0 73.0
85 41
67.5 32.5
290 177
62.1 37.9 .268
99 71 .050
58.2 41.8
344 346
50.1 49.9
57 69
45.2 54.8
206 261
44.1 55.9 .821
53 79 34 4 \.001
31.2 46.5 20.0 2.4
362 173 130 25
52.5 25.1 18.8 3.6
73 18 33 2
57.9 14.3 26.2 1.6
360 43 59 5
77.1 9.2 12.6 1.1 \.001
100 29 26 15 \.001
58.8 17.1 15.3 8.8
238 159 166 127
34.5 23.0 24.1 18.4
63 37 12 14
50.0 29.4 9.52 11.1
146 143 101 77
31.3 30.6 21.6 16.5 \.001
97 49 24
57.1 28.8 14.1
146 274 270 \.001
21.2 39.7 39.1
64 44 18
50.8 34.9 14.3
126 212 129
27.0 45.4 27.6 \.001
29 141
17.1 82.9
101 589 .435
14.6 85.4
38 88
30.2 69.8
105 362
22.5 77.5 .074
48 122
28.2 71.8
138 552 .019
20.0 80.0
35 91
27.8 72.2
93 374
19.9 80.1 .057
99 71
58.2 41.8
58.1 41.9
72 54
57.1 42.9
267 200
170
19.8
401 289 .981 690
80.2
126
21.3
467
57.2 42.8 .995 78.7
a
Social isolation as defined by a social isolation score 2. Source: 1999-2006 National Health and Nutrition Examination Survey. Values are not adjusted for survey weights. P values are inferred from x2 tests. c PTA = pure tone average (500, 1000, 2000, and 4000 Hz). b
support from their partners than women. Discussing emotional problems is perceived to be a feminine trait,22 and men may have been hesitant to admit they needed more social support when completing the SSQ. The onset of ARHL is approximately 7 years sooner in men than women,26 and men with ARHL may have adjusted to the
social impact of hearing loss before the age of 60 years. Longitudinal analysis is needed. Hearing loss was surprisingly not associated with social isolation in the older age group. Prevalence of social isolation was higher among the hearing impaired in both age groups (Table 1), but the relative difference was smaller in
382
Otolaryngology–Head and Neck Surgery 150(3)
Table 2. Prevalence of social isolationa among the hearing impairedb and non–hearing impaired.c Social isolation Age 60-69 y
PTA .25 dB in better-hearing ear Yes No P
Age 70-84 y
Yes Row %
No Row %
Yes Row %
No Row %
20.6 11.9
79.4 88.1 .003
19.8 15.6
80.4 84.4 .360
a
Social isolation as defined by a social isolation score 2. Pure tone average (500, 1000, 2000, and 4000 Hz) .25 dB in the better-hearing ear. c Values are adjusted to account for survey weighting and thus approximate the general population. P values are inferred from x2 tests. b
Table 3. Multivariate-adjusted odds ratios for social isolationa per 25-dB of hearing loss.b,c Odds of social isolation per 25-dB hearing loss Both sexes
Women
Men
Age, y
OR
SE
95% CI
P
OR
SE
95% CI
P
OR
SE
95% CI
P
60-69d 70-84
2.14 1.24
0.54 0.30
[1.29, 3.57] [0.75, 2.0]
.004 .371
3.49 1.39
1.05 0.54
[1.91, 6.39] [0.60, 3.16]
\.001 .415
1.11 1.06
0.29 0.36
[0.66, 1.88] [0.53, 2.12]
.679 .865
Abbreviations: CI, confidence interval; OR, odds ratio; SE, standard error. a Social isolation as defined by a social isolation score 2. b Pure tone average (500, 1000, 2000, and 4000 Hz) in the better-hearing ear. c Covariates in the models included: pure tone average in the better-hearing ear, gender, age, race, hearing aid use, noise exposure history, education, income, ever-smoking status, stroke history, heart disease history, hypertension history, and diabetes history. d Effect modification by gender was significant in the 60- to 69-year-old age group (P = .003) but not in the 70- to 84-year-old age group.
Table 4. Multivariate-adjusted odds ratios for each component item of the social isolation scale per 25-dB hearing loss.a,b Multiplication of odds of social isolation per 25-dB hearing loss Both sexes Item Age 60-69 y Low emotional support2 Low financial support Not married No close friends Age 70-84 y Low emotional support Low financial support Not marriedc No close friends
Women
Men
OR
SE
95% CI
P
OR
SE
95% CI
P
OR
SE
95% CI
P
1.32 1.17 1.62 1.05
0.23 0.13 0.41 0.15
[0.93, [0.94, [0.98, [0.78,
1.88] 1.45] 2.69] 1.41]
.111 .151 .060 .743
2.04 1.23 1.84 1.11
0.67 0.12 0.57 0.11
[1.04, [1.01, [0.98, [0.91,
3.97] 1.49] 3.45] 1.37]
.037 .040 .059 .302
0.70 0.86 1.30 0.69
0.21 0.32 0.44 0.36
[0.38, [0.41, [0.65, [0.24,
1.29] 1.82] 2.57] 1.99]
.247 .687 .449 .486
1.52 1.31 0.96 1.37
0.46 0.25 0.20 0.42
[0.80, [0.88, [0.62, [0.72,
2.89] 1.95] 1.49] 2.64]
.187 .170 .852 .315
2.21 1.75 1.28 1.15
1.11 0.77 0.37 0.43
[0.76, [0.69, [0.69, [0.53,
6.43] 4.45] 2.38] 2.53]
.135 .219 .405 .703
0.98 1.02 0.64 1.86
0.30 0.34 0.17 0.92
[0.51, [0.50, [0.36, [0.65,
1.90] 2.07] 1.11] 5.35]
.953 .950 .106 .226
Abbreviations: CI, confidence interval; OR, odds ratio; SE, standard error. a Pure tone average (500, 1000, 2000, 4000 Hz) in the better-hearing ear. b Covariates in the models included pure tone average in the better-hearing ear, age, gender, race, hearing aid use, noise exposure history, education, income, ever-smoking status, stroke history, heart disease history, hypertension history, and diabetes history. c Indicates P value for gender interaction term \.05.
Mick et al the older group. The difference might have reached statistical significance with a larger sample size. Persons with hearing loss and social isolation may die at younger ages than socially isolated persons without hearing loss. Persons may only experience increased social isolation around the time of hearing loss onset or during working years and adapt in the later years. Future research will be needed to corroborate the age interaction using metrics of social isolation validated for the elderly population and to address underlying reasons. Wu et al27 observed that among a sample of persons with bilateral hearing impairment, older adults (age 65 years) appeared to have lifestyles that placed fewer demands on hearing than younger adults (age \65 years). Their data are consistent with 2 hypotheses. First, age-related cognitive decline may reduce the ability of older individuals to compensate for hearing impairment, and thus they have a tendency to seek out quiet acoustic environments. Second, the shift to quieter environments is related more to lifestyle preferences that naturally change with age. Our data, while not disproving the first hypothesis, support the second. The more an individual is forced to alter his or her lifestyle from age-adjusted norms, the more likely one might expect there to be ensuing social isolation. This rationale might explain the null findings in the 70- to 84-year-old age group. Alternatively, the outcome metric may not have been sensitive enough to detect differences in social isolation in the elderly group. Aspects of social networks and support that are relevant to individuals may vary according to personality, age, sex, and culture. Glass et al28 found that with age, marital ties became less important for the conduit of emotional support. Bonds with children, close relatives, and confidantes assumed greater significance.28 It may be that hearing loss affects only certain aspects of the social network and support structure. As our knowledge about aging and social support improves, more refined scales may become available to measure isolation across different age groups. The consequences of an association between hearing impairment and social isolation, among any demographic, may be significant. In older adults, reduced social networks and support have been associated with functional decline, all-cause mortality, death following myocardial infarction, depression, and lower perceptions of health status. Isolated persons are generally thought to have reduced QoL. Social isolation has been associated with increased risk of dementia and cognitive decline in several prospective studies,29-31 and Wilson et al32 demonstrated that the risk of Alzheimer disease was more than doubled in lonely persons compared with nonlonely controls. Hearing loss has also been shown to be predictive of dementia, and it is possible that social isolation is a mediator in the pathway in some individuals. Social isolation might prevent hearing-impaired individuals from accessing adequate treatment and services. In the Blue Mountains study in Australia, hearing loss predicted greater reliance on formal (eg, community programs) and informal support services (eg, nonspouse family and friends providing assistance with daily activities such as cleaning or shopping).33 Isolation from others might restrict knowledge of and access to such supports. Effective mitigation of hearing impairment often requires
383 complicated assistive technologies, new communication strategies such as speech reading, environmental accommodations, and patience from others. These tasks can be daunting, even when one has a strong support network. Isolated persons may find the challenges extremely difficult, if not impossible to overcome, and therefore represent a vulnerable group. Our study had limitations. The NHANES SSQ questionnaire and the SIS have not been formally evaluated and standardized, and the composite scores might not have been weighted appropriately. Inferences are limited by the crosssectional nature of the data, and reverse causation cannot be excluded. However, it seems biologically unlikely that social isolation would cause hearing loss. Confounding by unmeasured covariates is possible. No information was available for persons younger than 60 years. Data were significantly more likely to be missing for persons who were African American, Hispanic, low income, or older. We conditioned on these factors in the multivariate analysis and therefore do not expect that differential nonresponse across levels of these variables would have biased results. We cannot exclude the possibility that unmeasured factors affected nonresponses; however, it is reasonable to expect that predictors of social isolation might also predict nonresponse. If so, bias would be to the null, and our estimates would underestimate the association between hearing loss and social isolation. Research is needed to confirm the association found between hearing loss and social isolation in 60- to 69-yearold women and reproduce the age and sex interaction. Questions remain regarding mechanisms that might mediate the observed association, and the consequences of social isolation in hearing-impaired individuals need to be fully elucidated. Studies are necessary to clarify which aspects of social networks and support might be most affected by hearing impairment and to determine the most appropriate outcome measures to use among different demographic groups. An emerging body of evidence suggests that ARHL might have more substantial effects on health and wellbeing than previously thought. The finding of an association between hearing impairment and social isolation in women aged 60 to 69 years adds to the evidence and calls for greater attention to the challenges faced by the millions of older individuals who are hearing impaired. As populations around the globe continue to age, the importance of understanding and mediating the consequences of age-related hearing loss is becoming increasingly clear. Author Contributions Paul Mick, conception and design, acquisition of data, analysis and interpretation of data, formulation of draft manuscript, editing of manuscript, manuscript submission; Ichiro Kawachi, design, analysis and interpretation of data, editing of manuscript; Frank R. Lin, conception and design, analysis and interpretation of data, editing of manuscript.
Disclosures Competing interests: Frank R. Lin, consultant for Cochlear, Ltd; Scientific Advisory Board for Pfizer and Autifony; speaker for Amplifon, Med El, and Cochlear.
384 Sponsorships: None. Funding source: None.
References 1. Berkman LF, Kawachi I. Social Epidemiology. New York, NY: Oxford University Press; 2000. 2. Gopinath B, Rochtchina E, Anstey KJ, et al. Living alone and risk of mortality in older, community-dwelling adults. JAMA Intern Med. 2013;173:320-321. 3. Berkman LF, Syme L. Social networks, host resistance and mortality: a nine-year follow up study of Alameda County residents. Am J Epidemiol. 1979;109:186-204. 4. Oxman TE, Berkman LF, Kasl S, et al. Social support and depressive symptoms in the elderly. Am J Epidemiol. 1992; 135:356-368. 5. Perissinotto CM, Stijacic Cenzer I, Covinsky KE. Loneliness in older persons: a predictor of functional decline and death. Arch Intern Med. 2012;172:1078-1083. 6. Cobb S. Social support as a moderator of life stress. Psychosom Med. 1976;38:300-314. 7. White AM, Philogene GS, Fine L, et al. Social support and self-reported health status of older adults in the United States. Am J Public Health. 2009;99:1872-1878. 8. Chia E, Wang JJ, Rochtchina E, et al. Hearing impairment and health related quality of life: the Blue Mountains hearing study. Ear Hear. 2007;28:187-195. 9. Pronk M, Deeg DJ, Smits C, et al. Prospective effects of hearing status on loneliness and depression in older persons: identification of subgroups. Int J Audiol. 2011;50:887-896. 10. Tun P, McCoy S, Wingfield A. Aging, hearing acuity and the attentional costs of effortful listening. Psychol Aging. 2009;24: 761-766. 11. Pichora-Fuller MK, Schneider BA, Daneman M. How young and old adults listen to and remember speech in noise. J Acoust Soc Am. 1995;97:593-608. 12. Wingfield A, Grossman M. Language and the aging brain: patterns of neural compensatoin revealed by functional brain imaging. J Neurophysiol. 2006;96:2830-2839. 13. Mathers C, Smith A, Concha M. Global Burden of Hearing Loss in the Year 2000. Geneva, Switzerland: World Health Organization; 2000. 14. National Health and Nutrition Examination Survey Data. Hyattsville, MD: Centers for Disease Control and Prevention, National Center for Health Statistics; 1999-2006. 15. Pacala JT, Yueh B. Hearing deficits in the older patient: ‘‘I didn’t notice anything.’’JAMA. 2012;307:1185-1194. 16. The National Health and Nutrition Examination Survey Questionnaire and Exam Protocol, 2013. Audiometry Manual 2005-06. Centers for Disease Control and Prevention National Centers for Health Statistics. 17. Seeman TE, Berkman LF. Structural characteristics of social networks and their relationship with social support in the elderly: who provides support. Soc Sci Med. 1988;36:737-749.
Otolaryngology–Head and Neck Surgery 150(3) 18. Maltz DN, Borker RA. A cultural approach to male-female miscommunication. In Gumperz JJ, ed, Language and Social Identity. Cambridge, UK, Cambridge University Press; 1982: 196-216. 19. Wood JT, Inman C. In a different mode: masculine styles of communicating closeness. J Appl Commun Res. 1993;21:279-295. 20. Burleson BR. A different voice on different cultures: illusion and reality in the study of sex differences in personal relationships. Personal Relationships. 1997;4:229-241. 21. Burleson BR, Gilstrap CM. Explaining sex differences in interaction goals in support situations: Some mediating effects of expressivity and instrumentality. Communication Reports. 2002;15:43-55. 22. Burleson BR, Kunkel AW, Samter W, et al. Men’s and women’s evaluations of communication skills in personal relationships: when sex differences make a difference—and when they don’t. Journal of Social and Personal Relationships. 1996;13:201-224. 23. Burleson BR, Samter W. Individual differences in the perception of comforting messages: an exploratory investigation. Central States Speech Journal. 1985;36:39-50. 24. Burleson BR. The experience and effects of emotional support: what the study of cultural and gender differences can tell us about close relationships, emotion and interpersonal communication. Personal Relationships. 2003;10:1-23. 25. MacGeorge EL, Clark RA, Gillihan SJ. Sex differences in the provision of skillful emotional support: the mediating role of self-efficacy. Communication Reports. 2002;15:17-28. 26. Cruikshanks KJ, Nondahl DM, Tweed TS, Wiley TL, Klein BEK. Education, occupation, noise exposure history and the 10-yr cumulative incidence of hearing impairment in older adults. Hear Res. 2010;264:3-9. 27. Wu YH, Bentler RA. Do older adults have social lifestyles that place fewer demands on hearing?J Am Acad Audiol. 2012; 23:697-711. 28. Glass TA, Mendes de Leon CF, Seeman TE, et al. Beyond single indicators of social networks: a Lisrel analysis of social ties among the elderly. Soc Sci Med. 1997;44:1503-1517. 29. Bassuk SS, Glass TA, Berkman LF. Social disengagement and incident cognitive decline in community-dwlling elderly persons. Ann Intern Med. 1999;131:165-173. 30. Cacioppo JT, Hawkley LC. Perceived social isolation and cognition. Trends Cogn Sci. 2009;13:447-454. 31. Wang HX, Karp A, Winbald B, et al. Late-life engagement in social and leisure activities is associated with a decreased risk of dementia. Am J Epidemiol. 2002;155:1081-1087. 32. Wilson RS, Krueger KR, Arnold SE, et al. Loneliness and risk of Alzheimer disease. Arch Gen Psychiatry. 2007;64:234-240. 33. Schneider J, Gopinath B, Karpa MJ, et al. Hearing loss impacts on the use of community and informal supports. Age Ageing. 2010;39:458-464.
The Association between Hearing Loss and Social Isolation in Older Adults Paul Mick, MD, MPH, FRCSC1, Ichiro Kawachi, MD, PhD2, and Frank R. Lin, MD, PhD3
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Abstract Objective. To determine if age-related hearing loss is associated with social isolation and whether factors such as age, gender, income, race, or hearing aid use moderated the association. Study Design. Cross-sectional. Setting. Randomly sampled United States communities. Subjects and Methods. Cross-sectional data on adults 60 to 84 years old from the 1999 to 2006 cycles of the National Health and Nutrition Examination Survey were analyzed. The dependent variable was social isolation, which was defined using the social isolation score (SIS), a 4-point composite index consisting of items pertaining to strength of social network and support. SIS scores 2 were considered indicative of social isolation. The independent (predictor) variable was the pure tone average of speech frequency (0.5-4 kHz) hearing thresholds in the better-hearing ear. Covariates included potential medical, demographic, and otologic confounders. We used multivariate logistic regression to evaluate the association between hearing loss and the odds of having social isolation. An exploratory analysis was performed to assess the strength of associations between hearing loss and individual items of the SIS scale. Results. Greater hearing loss was associated with increased odds of social isolation in women aged 60 to 69 years (odds ratio [OR], 3.49 per 25-dB of hearing loss; 95% confidence interval, 1.91, 6.39; P \ .001). Effect modification by gender was significant in this age group (P = .003). Hearing loss was not significantly associated with social isolation in other age and gender groups. Conclusions. Greater hearing loss is associated with increased odds of being social isolated in a nationally representative sample of women aged 60 to 69 years.
Keywords hearing loss, hearing disorders, presbycusis, aging, social isolation, social network, social support
Otolaryngology– Head and Neck Surgery 2014, Vol. 150(3) 378–384 Ó American Academy of Otolaryngology—Head and Neck Surgery Foundation 2013 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0194599813518021 http://otojournal.org
Received June 13, 2013; revised November 19, 2013; accepted December 4, 2013.
S
ocial networks refer to the web of interpersonal relationships that may provide an individual with various types of social support (eg, instrumental assistance, information, and emotional reinforcement).1 Persons who have a small social network or lack close relationships or sources of social support can be considered isolated. Social isolation has been shown to be a predictor of mortality, psychiatric illness, and cognitive and functional decline in elderly persons.2-7 Therefore, determining interventions and characteristics that promote social engagement in older persons is important. Previous research has implicated hearing loss in the development of social isolation. Chia et al, in a sample of older Australians, found that hearing impairment was associated with poorer scores in the social functioning and role limitation due to emotional problems domains of the Short Form– 36 scale (SF-36).8 Pronk et al9 reported that poorer selfreported hearing scores predicted deterioration in social support among a cohort of 900 senior citizens in Amsterdam. Population-based analyses have not been conducted in the United States. The current literature does not answer whether there are certain aspects of social networks and support more strongly associated with hearing impairment or if associations are modified by demographic factors or hearing aid use. In situations in which hearing is difficult, greater cognitive resources are needed to process auditory signals.10-12 Persons with even mild hearing loss have challenges communicating verbally in the presence of background noise. As a consequence, social gatherings may become difficult,
1
University of British Columbia, Vancouver, British Columbia, Canada Harvard School of Public Health, Boston, Massachusetts, USA 3 Departments of Otolaryngology–Head & Neck Surgery, Geriatric Medicine, Mental Health, and Epidemiology, Johns Hopkins University, Baltimore, Maryland, USA 2
This article was presented at the 2013 AAO-HNSF Annual Meeting & OTO EXPO; September 29–October 3, 2013; Vancouver, British Columbia, Canada. Corresponding Author: Paul Mick, MD, MPH, FRCSC, University of British Columbia, Vancouver, 202-3330 Richter St, Kelowna, BC, V1W 4H6, Canada. Email: [email protected]
Mick et al spontaneity might be reduced as events must be planned around acoustic environments, and relationships may suffer because of communication problems or because activities once enjoyed together are no longer pleasant to the person with hearing loss. Approximately 290 million people worldwide are hearing impaired, including 48% of Americans aged 60 or older.13,14 The prevalence is expected to increase in aging populations. Age-related hearing loss (ARHL) is a significant public health concern, compounded by underutilization of treatment options.15 The objectives of our study were 2-fold. First, we sought to determine if there was a cross-sectional association between hearing loss and social isolation in a nationally representative sample of Americans aged 60 to 69 and 70 to 84 years. Second, we sought to determine if the association was modified by gender, income, race, or hearing aid use.
Methods Data were publicly available and deidentified, and thus institutional ethics board approval was not required.
Study Cohort The National Health and Nutrition Examination Study (NHANES) is an annual cross-sectional survey designed to assess the health and nutrition of a sample of approximately 5000 non-institutionalized American adults and children. The study visits 15 randomly selected counties annually and employs a complex sampling procedure designed to oversample individuals who are older, younger, or members of minority groups. Each individual is assigned a sampling weight that accounts for the variable odds of selection and enables results to be generalized to the national population. Interviews, physical examinations, and laboratory tests are performed. For some items, only a subset of participants is tested. We used NHANES data from 1999 to 2006. Different age groups completed the social support questionnaire (SSQ) and audiometric testing in different years. Between 1999 and 2004, half of all participants between 20 and 69 years had their hearing tested, and all patients older than 60 years completed the SSQ. Therefore, 50% of all patients aged 60 to 69 years completed both and were included in the study. In the 2005-2006 NHANES cycle, only adults aged 70 years or older had hearing tests. They all completed the SSQ, and thus all 701-year-olds from the 2005-2006 cycle were included. The SSQ and audiometry were not administered concurrently in the same age group after 2006. NHANES classifies anyone older than 85 years as being ‘‘85 years old.’’ Patients who were 851 years old were thus excluded so that age could be used as a continuous variable in regression models.
Audiometric Assessment Trained examiners performed audiometric testing according to NHANES protocols.16 Air conduction thresholds were measured following a modified Hughson Westlake procedure.
379 Supra-aural earphones were used unless they caused external auditory canal collapse or there was hearing asymmetry, in which case they were replaced with insert earphones. Masking procedures were not performed. Speech-frequency pure tone averages (PTAs) were calculated for both ears as the mean of pure tone thresholds at 0.5, 1, 2, and 4 kHz. The lower PTA (ie, in the betterhearing ear) was defined as the independent variable and was analyzed as a continuous variable. All hearing thresholds in this article are reported as dB HL (American National Standards Institute, 2004).
Outcome Variable: Social Isolation The SSQ provides interview data on emotional, financial, and network support. Questions were selected from the Yale Health and Aging Study17 and the Social Network Index developed by Berkman et al in the Alameda County Study.3 Three items from the SSQ and marital status were combined into a summary measure of social isolation (the ‘‘social isolation score,’’ or SIS) that was tallied from 0 to 4, with higher scores indicating greater isolation. The first item asked whether the respondent was married or in a domestic partnership (0 = married or in partnership; 1 = not married or in partnership). The second item asked, ‘‘In general, how many close friends do you have?’’ (0 = at least 1 close friend; 1= no close friends). A third item characterized availability and adequacy of emotional support. A person scored 1 point if they had no one to provide emotional support (‘‘such as talking over problems or helping to make a difficult decision’’) or had at least one person to provide support, but ‘‘could have used more emotional support’’ in the past 12 months. The final item addressed the availability of financial support. The question asked, ‘‘If you need some extra help financially, could you count on anyone to help; for example, by paying any bills, housing costs, hospital visits, or providing you with food or clothes?’’ Those who answered ‘‘no’’ received 1 point. Participants were considered to be socially isolated if they were isolated according to 2 or more items (ie, SIS 2).
Other Study Variables Data on demographic variables (age, sex, race/ethnicity, education, annual household income), medical history (smoking status, cardiovascular disease, diabetes), noise exposure, and hearing aid use were obtained from interviews. Tests for multiplicative effect modification according to sex, income (low versus other), race (Mexican American or other Hispanic, and non-Hispanic black versus non-Hispanic white), and hearing aid use were planned a priori.
Statistical Methods Analyses were stratified according to age (60-69 and 70-84 years). The 1999-2004 and 2005-2006 NHANES data sets could not be merged because the ages of the hearing test participants during the 2 periods were mutually exclusive. We accounted for the complex sampling design in all multivariate analyses by weighting individual outcomes according to
380 National Center for Health Statistics guidelines. Nonweighted univariate assessments of the proportions with social isolation according to hearing status and other pertinent variables were performed using the x2 test. Multivariable logistic regression was used to determine odds ratios (ORs) for social isolation per 25-dB increment in PTA. The OR describes the multiplication of the odds of social isolation observed between persons with PTAs that differed by 25 dB. ORs .1 indicate increased odds of social isolation with hearing loss, while ORs \1 indicate lower odds with hearing loss. Multiplicative interaction terms between hearing loss and sex, income, race, and hearing aid use were added to the models and retained if their P values were \.05. An exploratory analysis was performed whereby each of the individual SSQ items were used as the dependent variable in multivariate models that were otherwise identical to the model used in the primary analysis. NHANES is a complex sample survey requiring mathematical variance approximation procedures to estimate sampling errors. The Taylor Series Linearization method was used as recommended by the National Centre for Health Statistics. NHANES participants eligible for hearing tests and the SSQ were excluded from the current study if they had any missing hearing, SSQ, or covariate data. A total of 18.4% of individuals were thus excluded. All analyses were performed with STATA 12.0 (StataCorp, College Station, TX), and a 2-sided P \ .05 was considered to be statistically significant.
Results Participant Characteristics There were 860 participants in the 60- to 69-year-old age group. Social isolation was prevalent in 19.8% of these individuals. In the 70- to 84-year-old age group, there were 593 participants, and social isolation was prevalent in 21.3%. Table 1 shows the raw (unweighted) characteristics of study participants. Table 2 displays the weighted proportions of socially isolated persons according to hearing status. In the 60- to 69-year-old group, 20.6% and 11.9% of persons with and without hearing loss, respectively, were isolated (P = .003). The corresponding proportions were 19.8% and 15.6% (respectively) in the 70- to 84-year-old group.
Multivariate Analysis Social isolation was independently associated with hearing loss in the 60- to 69-year-old group after adjusting for demographic, otologic, and medical covariates, and there was significant effect modification by gender (interaction term P value = .003). In women, the odds of social isolation increased 3.49 times with every 25-dB shift in PTA (95% confidence interval [CI]: 1.91-6.39, P \ .001). In contrast, the association in men was not significant (OR 1.11, 95% CI: 0.66-1.81, P = .68). In the 70- to 84-year-old age group, the association between hearing loss and social isolation was not significant, and there was no effect modification by gender (Table 3). Effect modification by history of hearing
Otolaryngology–Head and Neck Surgery 150(3) aid use, race, income, or education levels was not significant (data not shown).
Exploratory Analysis When the individual items were used as dependent variables in multivariate models, for 60- to 69-year-old women there were significant associations between hearing impairment and low emotional support and low financial support. The strongest association was with low emotional support. The P value for the association between hearing loss and being neither married nor in a domestic partnership was .059 (Table 4).
Discussion The primary analysis showed a strong association between hearing loss and social isolation in 60- to 69-year-old women but not in men or older individuals. In the exploratory analysis examining associations between hearing loss and components of the SIS, among 60- to 69-year-old women hearing loss was associated with reduced availability of emotional and financial support. This is the largest population-based study specifically investigating the association between hearing loss and social isolation. Other studies examined the impact of hearing loss on quality-of-life (QoL) metrics that contained measures of social and emotional function. Hearing loss was predictive of poorer scores in the social functioning and role limitation due to emotional problems domains of the SF-36 scale in a sample of older Australians. Effect modification by age and gender was not assessed, and regular hearing aid use did not have an impact on scores.8 A 2011 study reported a significant association between self-reported hearing loss and worsening emotional and social loneliness over a 4-year period among elderly Amsterdam residents.9 The extent to which men and women use verbal communication to obtain emotional support might, in part, explain the gender interaction in our data. Some sociologists have argued that talk is the primary vehicle by which women create and maintain intimacy and connectedness.18 In contrast, men may view dialogue as a way to accomplish instrumental tasks, convey information, and maintain autonomy in relationships.19 Burleson et al investigated differences in how men and women value communication skills among close friends and romantic partners. Women rated expressive skills such as ego support, conflict management, and comforting higher than men did, who gave greater weight to instrumental or interactional skills.20-23 The magnitude of the differences between genders in preferences for communication of emotional support is contested.24 The upshot is that if women rely more heavily on verbal communication, then hearing loss might impair their ability to receive emotional support to a greater degree than men. Other reasons may help explain the gender difference. Studies have shown that women are better at providing emotional support than men.22,25 Within heterosexual couples, men with hearing loss might therefore receive more
Mick et al
381
Table 1. Subject characteristics according to being socially isolateda or not isolated.b Social isolation Age 60-69 y Yes
PTA .25 dB in better earc Yes No P Sex Female Male P Race White Hispanic Black Other P Education High school incomplete High school graduate Some college College graduate P Annual household income \$20,000 $20,000-$44,999 $45,000 P History of heart disease Yes No P History of diabetes Yes No P History of smoking Yes No P Total
Age 70-84 y No
Yes
No
n
Col %
n
Col %
n
Col %
n
Col %
62 108 .014
36.5 63.5
186 504
27.0 73.0
85 41
67.5 32.5
290 177
62.1 37.9 .268
99 71 .050
58.2 41.8
344 346
50.1 49.9
57 69
45.2 54.8
206 261
44.1 55.9 .821
53 79 34 4 \.001
31.2 46.5 20.0 2.4
362 173 130 25
52.5 25.1 18.8 3.6
73 18 33 2
57.9 14.3 26.2 1.6
360 43 59 5
77.1 9.2 12.6 1.1 \.001
100 29 26 15 \.001
58.8 17.1 15.3 8.8
238 159 166 127
34.5 23.0 24.1 18.4
63 37 12 14
50.0 29.4 9.52 11.1
146 143 101 77
31.3 30.6 21.6 16.5 \.001
97 49 24
57.1 28.8 14.1
146 274 270 \.001
21.2 39.7 39.1
64 44 18
50.8 34.9 14.3
126 212 129
27.0 45.4 27.6 \.001
29 141
17.1 82.9
101 589 .435
14.6 85.4
38 88
30.2 69.8
105 362
22.5 77.5 .074
48 122
28.2 71.8
138 552 .019
20.0 80.0
35 91
27.8 72.2
93 374
19.9 80.1 .057
99 71
58.2 41.8
58.1 41.9
72 54
57.1 42.9
267 200
170
19.8
401 289 .981 690
80.2
126
21.3
467
57.2 42.8 .995 78.7
a
Social isolation as defined by a social isolation score 2. Source: 1999-2006 National Health and Nutrition Examination Survey. Values are not adjusted for survey weights. P values are inferred from x2 tests. c PTA = pure tone average (500, 1000, 2000, and 4000 Hz). b
support from their partners than women. Discussing emotional problems is perceived to be a feminine trait,22 and men may have been hesitant to admit they needed more social support when completing the SSQ. The onset of ARHL is approximately 7 years sooner in men than women,26 and men with ARHL may have adjusted to the
social impact of hearing loss before the age of 60 years. Longitudinal analysis is needed. Hearing loss was surprisingly not associated with social isolation in the older age group. Prevalence of social isolation was higher among the hearing impaired in both age groups (Table 1), but the relative difference was smaller in
382
Otolaryngology–Head and Neck Surgery 150(3)
Table 2. Prevalence of social isolationa among the hearing impairedb and non–hearing impaired.c Social isolation Age 60-69 y
PTA .25 dB in better-hearing ear Yes No P
Age 70-84 y
Yes Row %
No Row %
Yes Row %
No Row %
20.6 11.9
79.4 88.1 .003
19.8 15.6
80.4 84.4 .360
a
Social isolation as defined by a social isolation score 2. Pure tone average (500, 1000, 2000, and 4000 Hz) .25 dB in the better-hearing ear. c Values are adjusted to account for survey weighting and thus approximate the general population. P values are inferred from x2 tests. b
Table 3. Multivariate-adjusted odds ratios for social isolationa per 25-dB of hearing loss.b,c Odds of social isolation per 25-dB hearing loss Both sexes
Women
Men
Age, y
OR
SE
95% CI
P
OR
SE
95% CI
P
OR
SE
95% CI
P
60-69d 70-84
2.14 1.24
0.54 0.30
[1.29, 3.57] [0.75, 2.0]
.004 .371
3.49 1.39
1.05 0.54
[1.91, 6.39] [0.60, 3.16]
\.001 .415
1.11 1.06
0.29 0.36
[0.66, 1.88] [0.53, 2.12]
.679 .865
Abbreviations: CI, confidence interval; OR, odds ratio; SE, standard error. a Social isolation as defined by a social isolation score 2. b Pure tone average (500, 1000, 2000, and 4000 Hz) in the better-hearing ear. c Covariates in the models included: pure tone average in the better-hearing ear, gender, age, race, hearing aid use, noise exposure history, education, income, ever-smoking status, stroke history, heart disease history, hypertension history, and diabetes history. d Effect modification by gender was significant in the 60- to 69-year-old age group (P = .003) but not in the 70- to 84-year-old age group.
Table 4. Multivariate-adjusted odds ratios for each component item of the social isolation scale per 25-dB hearing loss.a,b Multiplication of odds of social isolation per 25-dB hearing loss Both sexes Item Age 60-69 y Low emotional support2 Low financial support Not married No close friends Age 70-84 y Low emotional support Low financial support Not marriedc No close friends
Women
Men
OR
SE
95% CI
P
OR
SE
95% CI
P
OR
SE
95% CI
P
1.32 1.17 1.62 1.05
0.23 0.13 0.41 0.15
[0.93, [0.94, [0.98, [0.78,
1.88] 1.45] 2.69] 1.41]
.111 .151 .060 .743
2.04 1.23 1.84 1.11
0.67 0.12 0.57 0.11
[1.04, [1.01, [0.98, [0.91,
3.97] 1.49] 3.45] 1.37]
.037 .040 .059 .302
0.70 0.86 1.30 0.69
0.21 0.32 0.44 0.36
[0.38, [0.41, [0.65, [0.24,
1.29] 1.82] 2.57] 1.99]
.247 .687 .449 .486
1.52 1.31 0.96 1.37
0.46 0.25 0.20 0.42
[0.80, [0.88, [0.62, [0.72,
2.89] 1.95] 1.49] 2.64]
.187 .170 .852 .315
2.21 1.75 1.28 1.15
1.11 0.77 0.37 0.43
[0.76, [0.69, [0.69, [0.53,
6.43] 4.45] 2.38] 2.53]
.135 .219 .405 .703
0.98 1.02 0.64 1.86
0.30 0.34 0.17 0.92
[0.51, [0.50, [0.36, [0.65,
1.90] 2.07] 1.11] 5.35]
.953 .950 .106 .226
Abbreviations: CI, confidence interval; OR, odds ratio; SE, standard error. a Pure tone average (500, 1000, 2000, 4000 Hz) in the better-hearing ear. b Covariates in the models included pure tone average in the better-hearing ear, age, gender, race, hearing aid use, noise exposure history, education, income, ever-smoking status, stroke history, heart disease history, hypertension history, and diabetes history. c Indicates P value for gender interaction term \.05.
Mick et al the older group. The difference might have reached statistical significance with a larger sample size. Persons with hearing loss and social isolation may die at younger ages than socially isolated persons without hearing loss. Persons may only experience increased social isolation around the time of hearing loss onset or during working years and adapt in the later years. Future research will be needed to corroborate the age interaction using metrics of social isolation validated for the elderly population and to address underlying reasons. Wu et al27 observed that among a sample of persons with bilateral hearing impairment, older adults (age 65 years) appeared to have lifestyles that placed fewer demands on hearing than younger adults (age \65 years). Their data are consistent with 2 hypotheses. First, age-related cognitive decline may reduce the ability of older individuals to compensate for hearing impairment, and thus they have a tendency to seek out quiet acoustic environments. Second, the shift to quieter environments is related more to lifestyle preferences that naturally change with age. Our data, while not disproving the first hypothesis, support the second. The more an individual is forced to alter his or her lifestyle from age-adjusted norms, the more likely one might expect there to be ensuing social isolation. This rationale might explain the null findings in the 70- to 84-year-old age group. Alternatively, the outcome metric may not have been sensitive enough to detect differences in social isolation in the elderly group. Aspects of social networks and support that are relevant to individuals may vary according to personality, age, sex, and culture. Glass et al28 found that with age, marital ties became less important for the conduit of emotional support. Bonds with children, close relatives, and confidantes assumed greater significance.28 It may be that hearing loss affects only certain aspects of the social network and support structure. As our knowledge about aging and social support improves, more refined scales may become available to measure isolation across different age groups. The consequences of an association between hearing impairment and social isolation, among any demographic, may be significant. In older adults, reduced social networks and support have been associated with functional decline, all-cause mortality, death following myocardial infarction, depression, and lower perceptions of health status. Isolated persons are generally thought to have reduced QoL. Social isolation has been associated with increased risk of dementia and cognitive decline in several prospective studies,29-31 and Wilson et al32 demonstrated that the risk of Alzheimer disease was more than doubled in lonely persons compared with nonlonely controls. Hearing loss has also been shown to be predictive of dementia, and it is possible that social isolation is a mediator in the pathway in some individuals. Social isolation might prevent hearing-impaired individuals from accessing adequate treatment and services. In the Blue Mountains study in Australia, hearing loss predicted greater reliance on formal (eg, community programs) and informal support services (eg, nonspouse family and friends providing assistance with daily activities such as cleaning or shopping).33 Isolation from others might restrict knowledge of and access to such supports. Effective mitigation of hearing impairment often requires
383 complicated assistive technologies, new communication strategies such as speech reading, environmental accommodations, and patience from others. These tasks can be daunting, even when one has a strong support network. Isolated persons may find the challenges extremely difficult, if not impossible to overcome, and therefore represent a vulnerable group. Our study had limitations. The NHANES SSQ questionnaire and the SIS have not been formally evaluated and standardized, and the composite scores might not have been weighted appropriately. Inferences are limited by the crosssectional nature of the data, and reverse causation cannot be excluded. However, it seems biologically unlikely that social isolation would cause hearing loss. Confounding by unmeasured covariates is possible. No information was available for persons younger than 60 years. Data were significantly more likely to be missing for persons who were African American, Hispanic, low income, or older. We conditioned on these factors in the multivariate analysis and therefore do not expect that differential nonresponse across levels of these variables would have biased results. We cannot exclude the possibility that unmeasured factors affected nonresponses; however, it is reasonable to expect that predictors of social isolation might also predict nonresponse. If so, bias would be to the null, and our estimates would underestimate the association between hearing loss and social isolation. Research is needed to confirm the association found between hearing loss and social isolation in 60- to 69-yearold women and reproduce the age and sex interaction. Questions remain regarding mechanisms that might mediate the observed association, and the consequences of social isolation in hearing-impaired individuals need to be fully elucidated. Studies are necessary to clarify which aspects of social networks and support might be most affected by hearing impairment and to determine the most appropriate outcome measures to use among different demographic groups. An emerging body of evidence suggests that ARHL might have more substantial effects on health and wellbeing than previously thought. The finding of an association between hearing impairment and social isolation in women aged 60 to 69 years adds to the evidence and calls for greater attention to the challenges faced by the millions of older individuals who are hearing impaired. As populations around the globe continue to age, the importance of understanding and mediating the consequences of age-related hearing loss is becoming increasingly clear. Author Contributions Paul Mick, conception and design, acquisition of data, analysis and interpretation of data, formulation of draft manuscript, editing of manuscript, manuscript submission; Ichiro Kawachi, design, analysis and interpretation of data, editing of manuscript; Frank R. Lin, conception and design, analysis and interpretation of data, editing of manuscript.
Disclosures Competing interests: Frank R. Lin, consultant for Cochlear, Ltd; Scientific Advisory Board for Pfizer and Autifony; speaker for Amplifon, Med El, and Cochlear.
384 Sponsorships: None. Funding source: None.
References 1. Berkman LF, Kawachi I. Social Epidemiology. New York, NY: Oxford University Press; 2000. 2. Gopinath B, Rochtchina E, Anstey KJ, et al. Living alone and risk of mortality in older, community-dwelling adults. JAMA Intern Med. 2013;173:320-321. 3. Berkman LF, Syme L. Social networks, host resistance and mortality: a nine-year follow up study of Alameda County residents. Am J Epidemiol. 1979;109:186-204. 4. Oxman TE, Berkman LF, Kasl S, et al. Social support and depressive symptoms in the elderly. Am J Epidemiol. 1992; 135:356-368. 5. Perissinotto CM, Stijacic Cenzer I, Covinsky KE. Loneliness in older persons: a predictor of functional decline and death. Arch Intern Med. 2012;172:1078-1083. 6. Cobb S. Social support as a moderator of life stress. Psychosom Med. 1976;38:300-314. 7. White AM, Philogene GS, Fine L, et al. Social support and self-reported health status of older adults in the United States. Am J Public Health. 2009;99:1872-1878. 8. Chia E, Wang JJ, Rochtchina E, et al. Hearing impairment and health related quality of life: the Blue Mountains hearing study. Ear Hear. 2007;28:187-195. 9. Pronk M, Deeg DJ, Smits C, et al. Prospective effects of hearing status on loneliness and depression in older persons: identification of subgroups. Int J Audiol. 2011;50:887-896. 10. Tun P, McCoy S, Wingfield A. Aging, hearing acuity and the attentional costs of effortful listening. Psychol Aging. 2009;24: 761-766. 11. Pichora-Fuller MK, Schneider BA, Daneman M. How young and old adults listen to and remember speech in noise. J Acoust Soc Am. 1995;97:593-608. 12. Wingfield A, Grossman M. Language and the aging brain: patterns of neural compensatoin revealed by functional brain imaging. J Neurophysiol. 2006;96:2830-2839. 13. Mathers C, Smith A, Concha M. Global Burden of Hearing Loss in the Year 2000. Geneva, Switzerland: World Health Organization; 2000. 14. National Health and Nutrition Examination Survey Data. Hyattsville, MD: Centers for Disease Control and Prevention, National Center for Health Statistics; 1999-2006. 15. Pacala JT, Yueh B. Hearing deficits in the older patient: ‘‘I didn’t notice anything.’’JAMA. 2012;307:1185-1194. 16. The National Health and Nutrition Examination Survey Questionnaire and Exam Protocol, 2013. Audiometry Manual 2005-06. Centers for Disease Control and Prevention National Centers for Health Statistics. 17. Seeman TE, Berkman LF. Structural characteristics of social networks and their relationship with social support in the elderly: who provides support. Soc Sci Med. 1988;36:737-749.
Otolaryngology–Head and Neck Surgery 150(3) 18. Maltz DN, Borker RA. A cultural approach to male-female miscommunication. In Gumperz JJ, ed, Language and Social Identity. Cambridge, UK, Cambridge University Press; 1982: 196-216. 19. Wood JT, Inman C. In a different mode: masculine styles of communicating closeness. J Appl Commun Res. 1993;21:279-295. 20. Burleson BR. A different voice on different cultures: illusion and reality in the study of sex differences in personal relationships. Personal Relationships. 1997;4:229-241. 21. Burleson BR, Gilstrap CM. Explaining sex differences in interaction goals in support situations: Some mediating effects of expressivity and instrumentality. Communication Reports. 2002;15:43-55. 22. Burleson BR, Kunkel AW, Samter W, et al. Men’s and women’s evaluations of communication skills in personal relationships: when sex differences make a difference—and when they don’t. Journal of Social and Personal Relationships. 1996;13:201-224. 23. Burleson BR, Samter W. Individual differences in the perception of comforting messages: an exploratory investigation. Central States Speech Journal. 1985;36:39-50. 24. Burleson BR. The experience and effects of emotional support: what the study of cultural and gender differences can tell us about close relationships, emotion and interpersonal communication. Personal Relationships. 2003;10:1-23. 25. MacGeorge EL, Clark RA, Gillihan SJ. Sex differences in the provision of skillful emotional support: the mediating role of self-efficacy. Communication Reports. 2002;15:17-28. 26. Cruikshanks KJ, Nondahl DM, Tweed TS, Wiley TL, Klein BEK. Education, occupation, noise exposure history and the 10-yr cumulative incidence of hearing impairment in older adults. Hear Res. 2010;264:3-9. 27. Wu YH, Bentler RA. Do older adults have social lifestyles that place fewer demands on hearing?J Am Acad Audiol. 2012; 23:697-711. 28. Glass TA, Mendes de Leon CF, Seeman TE, et al. Beyond single indicators of social networks: a Lisrel analysis of social ties among the elderly. Soc Sci Med. 1997;44:1503-1517. 29. Bassuk SS, Glass TA, Berkman LF. Social disengagement and incident cognitive decline in community-dwlling elderly persons. Ann Intern Med. 1999;131:165-173. 30. Cacioppo JT, Hawkley LC. Perceived social isolation and cognition. Trends Cogn Sci. 2009;13:447-454. 31. Wang HX, Karp A, Winbald B, et al. Late-life engagement in social and leisure activities is associated with a decreased risk of dementia. Am J Epidemiol. 2002;155:1081-1087. 32. Wilson RS, Krueger KR, Arnold SE, et al. Loneliness and risk of Alzheimer disease. Arch Gen Psychiatry. 2007;64:234-240. 33. Schneider J, Gopinath B, Karpa MJ, et al. Hearing loss impacts on the use of community and informal supports. Age Ageing. 2010;39:458-464.
