Impact of hearing impairment on spousal mental health: the Nord-Trøndelag Health Study

Abstract

Background: Previous studies indicate that hearing loss have negative emotional implications also on spouses of the hearing impaired persons. We sought to assess the relationship between hearing impairment and spousal mental health in the general population. Methods: Pure tone audiometry and questionnaires were administered to the adult population of Nord–Trøndelag County, Norway (1996–97). In the age group between 20 and 44 years, the number of cases with hearing impairment was very low; thus, this age group was excluded from analyses. In total, 8607 couples with women over 44 years and 9530 couples with men over 44 years were identified. Associations between measured and self-reported hearing impairment and spousal self-reported symptoms of anxiety and depression, and subjective well-being were estimated. Stratified by sex and adjusting for several covariates, mental health in spouses of persons with hearing impairment was compared with that of spouses of persons with normal hearing using the general linear model. Results: Audiometrically measured hearing was not significantly associated with spousal mental health. Moderate relations between self-reported hearing and spousal mental health were observed. Conclusion: Contrary to previous results based on self-reported hearing loss, our results based on audiometry did not indicate severe loss of mental health among spouses of persons with impaired hearing.

Introduction

Hearing impairment is one of the most common chronic health conditions in the world.¹ The point prevalence among adults has been estimated at 16–17% in western countries,² increasing with age.³ With an ageing global population, the number of people with hearing impairment is rising. It is widely recognized that hearing loss can have negative implications for individuals experiencing it. These include depression, life dissatisfaction and reduced well-being.^4–6 Hearing loss might also be difficult to deal with for the spouses of individuals with hearing impairment.

Hearing loss reduces the quality and quantity of couple communication.^7,8 Considering the centrality of communication between spouses, a person's hearing loss may negatively affect the relationship,⁹ even when the loss is classified as ‘mild’.¹⁰ Communication problems manifest themselves through constant repetitions and misunderstandings and a decrease in intimate talk and joking.⁸ Other problematic consequences of hearing loss reported by spouses are exposure to raised TV and radio volume levels, always having to answer the phone and having to act as an interpreter when with other persons.⁹ Through this, the spouses become important communication aids and possible caregivers to the hearing impaired persons. Some studies have used a caregiver stress model¹⁰ to explain the negative consequences for spouses of hearing impaired persons.¹¹ Communication difficulties, and possible care giving for the hearing impaired, might represent a stressor for the spouse and affect mental health and subjective well-being.

To date, there are few studies investigating the association between hearing loss and spousal mental health and well-being.¹² Studies have predominately involved younger female spouses of workers affected by occupational, noise-induced hearing loss.^7,13–16 The majority of the studies used qualitative approaches that described living with a hearing impaired spouse as a demanding task.¹⁴ Interviews and group discussions revealed a number of emotional consequences experienced by the spouses: irritation, guilt, sadness, anger, depression, stress, embarrassment, social isolation and anxiety.^{5,7,9,16–19}

Many researchers conclude that living with hearing impaired persons affects spousal mental health,^9,17,19 although few studies actually have assessed mental health. We found only one longitudinal study that assessed mental health using a validated measure;²⁰ the study demonstrated that in 418 older couples, hearing loss was associated with spouses’ lower mental, physical and social well-being 5 years later.²⁰ A limitation of this study is the use of a self-reported measure of hearing loss.

The aim of the present study is to assess the extent to which spouses of persons suffering from hearing impairment experience reduced subjective well-being (SWB) and poorer mental health compared to the remaining population. To our knowledge, there have been no population-based investigations aimed at estimating these associations using validated measures of mental health and pure tone audiometry. A second aim is to assess whether previously published results, based on self-reported hearing, might have been inflated by self-report bias, comparing results based on measured hearing with results obtained using self-reported hearing.

Methods

Sample

From 1995 to 1997, the adult population of Nord-Trøndelag County, Norway, was invited to take part in a health-screening survey, the Nord-Trøndelag Health Study (HUNT 2). This was a follow-up of HUNT 1 (1984–86). The Nord-Trøndelag Hearing Loss Study (1996–98) was a part of HUNT 2. Seventeen of the 24 municipalities in the county, including 21 496 married couples, were invited to participate. A description of the sample is available elsewhere.²¹ For the purpose of the present study, only individuals from couples with complete data were included. The governmental statistics agency, Statistics Norway, used the 11-digit personal identification number to identify registered couples. Complete valid data from the hearing loss study were collected from 13 678 couples (63.6%). Age ranged from 20 to 99 years (mean, 51.9; SD, 13.9). The data were organized as a ‘double entry’ file, with all respondents included twice, both as index person (person whose hearing level is used as an independent variable) and as spouse. All analyses were therefore run stratified by sex. Since few persons younger than 45 years had severe hearing impairment, pairs with index persons in this age group was excluded. There were 9530 couples with male index persons and 8607 couples with female index persons aged 45 years or older.

Measures

Audiometric hearing loss

Pure tone audiometry was completed while participants were seated in semi-portable, dismountable sound attenuation booths (Tegnér T-booth 95 × 105 × 210 cm), without hearing aid. The results met the recommended International Organization for Standardization standard for audiometric test administration.^21,22 The pure tone hearing threshold examinations for each ear included the test frequencies 0.5, 1, 2 and 4 kHz. Data from 100 randomly drawn retested participants showed high test–retest correlations.²² World Health Organization's classification of hearing impairment was used to compute pure tone averages in the better hearing ear. Categories were defined as ‘no impairment’ [≤25 decibels hearing level (dBHL)], ‘slight impairment’ (26–40 dBHL) and ‘disabling impairment’ (≥41 dBHL) according to the threshold level.¹

Self-reported hearing loss

Participants completed a questionnaire at home before the audiometric examination. One section pertained to perceived functional disability due to different impairments, phrased ‘Do you suffer from any long-term illness or injury of a physical or psychological nature that impairs your functioning in your everyday life?’ (yes, no). If ‘yes’ the informants were asked to complete a number of items on specific types of impairment, one of which was hearing impairment. Response categories were ‘slight’, ‘moderate’ and ‘severe’, and hearing impairment was scored 0 (not impaired) to 3 (severe). While waiting for the hearing examination, the participants completed a one-page questionnaire. Included here was an item about acknowledged hearing loss: ‘Do you have a hearing loss of which you are aware?’ (‘no’ or ‘yes’ scored 0 and 1). The two questionnaire items were summed to generate a new variable, self-reported hearing loss (SHL) (scored 0–4), which in turn was recoded into two categories: 0–1 = no SHL and 2–4 = SHL.

Symptoms of anxiety and depression

These symptoms were measured using 10 items from the Hopkins Symptom Checklist (SCL-25).²³ SCL-25 taps symptoms of anxiety with 10 items and symptoms of depression with 15 items, while the abbreviated version (SCL-10) includes 4 questions tapping anxiety and 6 tapping depression. The 10 items about being ‘bothered or distressed’ during the last 2 weeks were phrased as follows: ‘suddenly scared for no reason’, ‘feeling fearful’, ‘faintness, dizziness, or weakness’, ‘feeling tense or keyed up’, ‘blaming yourself for things’, ‘difficulty in falling asleep or staying asleep’, ‘feeling blue’, ‘feeling of worthlessness’, ‘feeling everything is an effort’ and ‘feeling hopeless about future’. Response categories ranged from ‘1’ (not at all) to ‘4’ (extremely). Separate summative scores were computed for anxiety and depression. An available sample with 6000 subjects has been used to compare the original (SCL-25) and the abbreviated (SCL-10) scale.²⁴ The correlation between the SCL-10 and the SCL-25 anxiety score was 0.91, and the correlation between the corresponding depression scores was 0.96. The correlation between the SCL-25 anxiety and depression scores was 0.73,²⁴ and the correlation between the corresponding SCL-10 scores in our sample was 0.63. Cronbach's alpha for the four-item anxiety score was 0.64 for men and 0.70 for women, while the alpha for the six-item depression score was 0.81 for both men and women. Of the respondents, 17.5% had missing values on one or more of the ten SCL items. We used SPSS missing value analysis (MVA), expectation maximization for imputation of missing values in respondents with valid data for at least 5 of the 10 items. Of the records, 8.6% were imputed and the remaining 8.9% were treated as missing data.

Subjective well-being

An additional questionnaire (Q2) was handed out during the examination, and 81.8% of the participants returned this questionnaire by mail. Three items constituted a scale measuring subjective well-being (SWB). The questions were answered on seven categories ranging from highly positive to highly negative (scored 1–7), and phrased as follows: ‘When you think about your life at the moment, would you say that you are by and large satisfied with life, or are you mostly dissatisfied?’; ‘Would you say you are usually cheerful or dejected?’ and ‘Do you mostly feel strong and fit, or tired and worn out?’. The measure has been used in previous publications.²⁵ Cronbach's alpha was 0.79 for men and 0.89 for women. The majority of participants from HUNT1, a health-screening study of the same population 11 years earlier than HUNT2, completed the well-being items on two different questionnaires, with a typical time lag of 1–2 weeks. The test–retest reliability was 0.68.²¹ Of the respondents (Q2) 4.7% had missing values on one or two of the three SWB-items. Of the records (with only one item missing) 3.3% was imputed using MVA, while the remaining 1.4% was treated as missing data.

The anxiety and depression scores were highly skewed, and were ln-transformed to approximate normal distributions. All dependent variables were standardized before inclusion in the analyses. The unstandardized regression coefficients (b) therefore show adjusted group mean differences in fractions of standard deviations (SD) for the dependent variables.

Confounders

Low socio-economic status, indicated by poor education, could be a risk factor for both hearing loss and spousal mental health. Educational level was therefore entered as a covariate in the analyses. Educational level was scored as one of five categories ranging from ‘primary school’ (7–10 years of schooling) to ‘four years or more at college/university’. Educational levels were summed within couples, resulting in a scale ranging from 2 to 10, and recoded into four categories: 2, 3–4, 5–6 and higher than 6. Spousal age was also included as a covariate.

An individual's SHL may be associated with self-reported mental health (correlated self-report bias), which, in turn, may be correlated with spousal self-reported mental health, causing an artificially high correlation. To control for such possible confounding index, person's own mental health was included as a covariate.

Statistical analyses

Multivariate ANOVA [SPSS General Linear Models (GLM), Unianova] was conducted for each of the three outcome measures separately. Audiometric hearing loss (AHL) was entered as a factor with three groups, adjusting for the effect of spousal age, spousal AHL and couple education. All analyses were run stratified by sex. One set of analyses estimated the association between AHL among all men and mental health in their female spouses, while the other estimated the relation between female AHL and male spousal mental health.

The stress associated with having a hearing impaired partner could partly depend on spouses’ own hearing levels. An interaction term between index persons’ AHL and spousal AHL was therefore tested.

Another set of analyses included SHL as an independent variable instead of AHL. Supplementary analyses of SHL were run with index persons’ own mental health as a covariate. These analyses test for spurious effects of an association between own hearing loss and mental health combined with a spouse correlation for mental health.

Results

Descriptive statistics

The mean age was 59.6 (SD = 10.35) among male index persons and 56.21 (SD = 10.47) among their spouses. Among female index persons, the mean age was 58.0 (SD = 9.47), and among their spouses, it was 60.8 (SD = 10.02). Table 1 presents the distributions of couples both based on male and female index persons. The table shows higher prevalence of AHL and SHL among men than that among women. The polychoric correlation between AHL and SHL was 0.70 (CI = 0.69–0.71).

Measured hearing impairment

ANOVA was run consecutively with three outcome variables in both sex strata, controlling for spousal age, spousal AHL and couple education. Index persons’ AHL had no significant main effect on spousal mental health ratings neither for male nor for female spouses. Adjusted differences between the groups in fractions of SDs (b) are presented in Table 2. Specifying an interaction term between the index person's and the spouse's AHL revealed no significant interaction effects. Testing an interaction term between AHL and age also gave no significant results.

Self-reported hearing loss

Adjusted for spousal age, spousal SHL and couple education, index persons’ SHL showed four significant main effects in the expected direction on adjusted means of spousal mental health and SWB (Table 3). Among male spouses, adjusted means of SWB differed significantly dependent on index person's SHL. Among female spouses, there were significant differences in the adjusted means of depression, anxiety and SWB. No adjusted differences between spouses of hearing impaired and normally hearing persons exceeded 0.13 SD.

Testing for interaction between the index person's and the spouse's SHL showed no significant results. Testing an interaction term between index SHL and age also gave no significant result. Rerunning the GLM with the index person's own mental health as a covariate, three of the four significant associations disappeared. The remaining significant relation was between SHL and symptoms of anxiety in female spouses, who reported 0.08 SD higher symptoms of anxiety (CI = 0.01–0.12, P = 0.023). Other adjusted differences (b-values) ranged from 0.00 to 0.04.

Discussion

The results showed no significant associations of AHL on spousal symptoms of anxiety and depression and subjective well-being. Results based on SHL included four significant results, three among female spouses. Three of the four disappeared when controlling for the index person's mental health.

Contrary to earlier studies, our results indicate that mental health does not differ much as a consequence of partners’ AHL. This is not the first time that conclusions based on clinical research fail to be replicated in larger quantitative studies in the literature of hearing loss. A review found that studies based on validated questionnaire instruments did not demonstrate associations repeatedly reported by clinical studies regarding negative implications of hearing loss on couple relationships.⁷

Earlier studies reporting negative spousal consequences of hearing impairment are mostly based on small samples drawn on the basis of attendance at audiology rehabilitation clinics rather than those drawn randomly. There may be characteristics of persons attending the clinics that differ systematically from hearing impaired persons in the rest of the population. People seeking help for their hearing problems are usually highly aware of these problems, and marital and communication problems may lead to their clinic attendance in the first place. Furthermore, there may be an elevated frequency of comorbidity in clinical studies—not only in the individual but also in the couple. Having diseases or impairments other than hearing loss in the couple might increase their own inclination, as well as family pressure, to seek out a clinic and might thus explain the larger burden experienced.

Patients in clinical interviews are highly aware of the fact that hearing loss and its consequences are the focus of the interview, which may cause over-reporting of adverse effects. In HUNT, there was contextual neutrality in the sense that the participants were not aware of the purpose of the study. Selection and reporting bias may have affected previous results, but limitations in our own study might also have biased our results. The estimates may be somewhat attenuated because of imperfect reliability of some of the mental health measures.

One aim of the study was to assess whether previously published results based on self-reported hearing might differ systematically from results based on measured hearing due to self-report bias. The number of statistically significant tests in our study increased from zero to four when changing the predictor from AHL to SHL. There were substantially more subjects reporting SHL than those with AHL. This difference primarily reflects stricter criteria for AHL than those for SHL. The discrepancy between the two measures implies that the conventional audiometric criteria for disabling hearing impairment used here appear to be stricter than what most people actually perceive as hearing impairment.

The somewhat stronger effects of SHL than of AHL on spousal mental health may partly depend on what the hearing measures actually reflect. SHL may reflect a real functional disability that is not fully tapped by audiometry. People seeking help for their hearing problems would probably be more likely to report SHL, which might explain why results using SHL are more in accordance with earlier studies than the results of audiometry. But the difference may also depend on other factors, perhaps related to reporting bias. People who are generally depressed and discontented tend to report most things, including their hearing, as being more negative than do happy people.²¹ In fact, inspection of our own data-file showed some covariation between SHL and other types of self-reported functional impairments, even after adjustment for age and sex (e.g. vision, results nor reported). Since there is no obvious reason to expect a substantial correlation between various types of age-adjusted impairments, this covariation is most likely due to correlated measurement errors, a tendency to perceive or report most things as positive or as negative. People who resemble one another on certain traits tend to mate.²⁶ If there exists a partner correlation for such response bias, correlated measurement errors between spouses might cause artificial spouse correlations between most characteristics reported by spouses. Such correlated bias could also cause a correlation between SHL in one spouse and mental health in the other. This could partly explain why our and previous studies find effects of SHL, but no effect of AHL. Since we do not know to what extent the result regarding SHL is real or just an artefact, self-report data do not appear to be well suited for examining the relationship between hearing loss and spousal mental health.

Although there are significant associations between SHL and spousal mental health, the fractions of SD explained are small, ranging from 0.05 to 0.13, and the confidence intervals almost approach zero. On the other hand, all the estimated associations, even the majority of those based on audiometry, are in the expected direction. This trend indicates that some minor relations may exist, or perhaps a few spouses living under special conditions may be severely bothered by their spouses’ hearing loss.

Conclusions

Our study contributes to the field by estimating the relation in a large epidemiological study. Considering the discrepancy between the present and previous results, more research is needed to explore the extent to which hearing loss affects spousal mental health. One of the previous studies²⁰ that showed a spouse effect was a longitudinal study while our study is cross-sectional, with the limitations regarding causal direction innate in this type of design. A longitudinal analysis would be a good next step to further examine whether such a design gives systematically different results. However, while we cannot empirically decide whether hearing loss affects spousal mental health or spousal mental health affects hearing, the former would undoubtedly be more likely than the latter.

Although our results imply that strong associations between hearing loss and spousal mental health and well-being are unlikely, future research should examine the factors that make hearing loss in the family hard to handle. Our findings indicate that studies based on audiometry and based on SHL may give somewhat different results.

Funding

The Nord-Trøndelag Hearing Loss Study, which is part of Nord–Trøndelag Health Study (HUNT), was funded by the National Institute on Deafness and Other Communication Disorders (NIDCD), research contract No. N01-DC-6-2104, and by the Norwegian Research Council.

Conflicts of interest: None declared.

Acknowledgements

HUNT was conducted in collaboration between the National Institute of Public Health, the National Health Screening Service, Nord-Trøndelag County Council, and the Norwegian University of Technology and Science. Drs H. J. Hoffman, NICDC, H. M Borchgrevink, and J. Holmen participated in the planning and administration of the study. The Nord-Trøndelag County Health Officer and the Community Health Officer in Levanger and in other municipalities provided organizational and other practical support. We want to thank the audiometry team for their diligence.

References