Hearing Loss in Older People with Schizophrenia: Audiologic Characteristics and Association with Psychosocial Functioning

Abstract

Objective:

The severity and impact of hearing deficits among adults with schizophrenia spectrum disorders may become increasingly relevant with advancing age. This study evaluated hearing ability and associated psychosocial functioning among older adults aged 50–70.

Design:

Cross-sectional analysis.

Setting:

Four outpatient psychiatry clinics in New York City.

Participants:

Individuals aged 50–70 years with diagnoses of schizophrenia or schizoaffective disorder.

Measurements:

Unaided pure tone air conduction audiometry conducted using a portable audiometry system determined the pure tone average (PTA) hearing threshold across four frequencies: 500, 1k, 2k, and 4k Hz. Better ear PTA defined the hearing threshold. Audiometry data retrieved from the U.S. National Health and Nutrition Examination Survey aided interpretation of sample hearing loss rates. Standard measures evaluated psychiatric symptoms, perceived impact of hearing impairment, loneliness, and quality of life.

Results:

Among audiometry completers (N=40), 35% (n=14) demonstrated subclinical hearing loss (16–25 dB) and 35% (n=14) had mild or worse hearing loss (≥26 dB). Rates were higher than expected based on age-based population data. Those who perceived hearing handicap rated it moderate (12.2%) or severe (7.3%); those who perceived tinnitus rated the impact as mild to moderate (12.2%) or catastrophic (2.4%). Neither psychiatric symptoms nor interviewer-rated quality of life was associated with hearing ability. Greater loneliness was significantly correlated with worse audiologic performance (r=.475, p<0.01) and greater perceived hearing handicap (r=.480, p<.01).

Conclusions:

Identifying the need for hearing loss treatment among aging adults with schizophrenia spectrum disorders is important given the potential implications for social functioning, cognitive and mental health.

INTRODUCTION

Age-related hearing loss (HL) is the third most common health condition affecting older adults. The prevalence of HL increases steeply with age, from 3% among adults 20–29 to 49% of adults ages 60–69 and greater than 80% in those over 85 years.¹ While historically considered a benign effect of aging, it is associated with significant morbidity, including social isolation, depression, cognitive decline, and dementia.^2–7 HL is the fifth leading cause of years lived with disability worldwide.^8,9

There is strong preliminary evidence that HL is a potentially unrecognized and therefore unmet service need among people with schizophrenia. A recent audiometry study of 84 outpatients ages 21–50 with schizophrenia, previously undiagnosed with any hearing impairment, found higher hearing thresholds and higher rates of mild hearing loss (23.81%) relative to an age-matched non-psychiatric control group (6.17%).¹⁰ Incidence of mild HL in the sample was also higher than expected based on age-based general population data.¹ The reasons for this increased HL are unclear. Antipsychotic medications are not known to be ototoxic, known environmental exposures in this sample were insignificant, and there was no association with concurrent psychiatric symptoms which might have affected performance. Hearing threshold in this sample was inversely and significantly correlated with a measure of global cognitive ability, a relationship that was reduced but remained significant after controlling for age.¹¹ Thus, incident HL was not only higher than expected but was also associated with greater cognitive impairment. These findings have generated interest regarding both the nature and impact of hearing deficits in people with schizophrenia, which may become increasingly relevant as adults with a schizophrenia spectrum diagnosis advance in age. While there is no consensus on the age that defines an “older adult,” evidence for accelerated aging and premature mortality within this population suggests it is prudent to consider the incidence and impact of such health conditions at ages 50 and older.^12,13

The current study thus sought to extend the previous study findings of increased HL incidence to an older sample of adults ages 50 to 70, while overcoming the limitations of the prior audiometric evaluation by assessing hearing across a broader frequency range and using methods of measurement comparable to those of epidemiological studies of HL in the general U.S. population. The primary aim was to evaluate whether, relative to historical data from U.S. adults of the same age, individuals with a schizophrenia spectrum diagnosis evidence higher rates of HL defined by standard cut-points. A second aim of this study was to evaluate a pathway towards routine screening by using a portable audiometry system to collect study data within community clinic settings where adults with schizophrenia spectrum disorder are likely to receive behavioral health services. To this end, we report on preliminary feasibility and reliability data for community clinic-based audiometry evaluation. A third aim was to explore both objectively and subjectively measured hearing ability in relation to psychosocial functioning. If hearing impairment, detected with clinic-based screening measures, is associated with levels of disability, there may be opportunity to impact the recovery trajectory through systematic identification and appropriate HL treatment.

METHODS

Participants

Recruitment took place at four outpatient clinics in the New York City metropolitan area. Participants were ages 50–70 years old with a DSM-5 diagnosis of schizophrenia or schizoaffective disorder, and stable on any psychotropic medication for the past 21 days. Individuals had no documented retrocochlear pathology or organic brain lesion (e.g., acoustic neuroma) responsible for hearing loss, Alzheimer’s disease, vascular dementia, or Parkinson’s disease, severe substance use disorder, documented intellectual disability, or premorbid IQ estimate <70. All participants were determined to be capable of providing informed consent and following study procedures.

Procedures

This study was conducted with oversight from the New York State Psychiatric Institute (NYSPI) IRB. Participants were recruited through advertisements posted in clinic common areas and through research referrals following the completion of a separate NYSPI IRB approved protocol. For all potential participants, a written and oral description of the study was provided, and written informed consent was obtained. Participants who met eligibility criteria were assessed by a trained research assistant for sociodemographic data, current medications and medical history (confirmed by chart review), audiological functioning, psychiatric symptoms, and psychosocial functioning. All subjects were assessed using the ShoeBox¹⁴ portable audiometry system at their outpatient behavioral health clinic. A subset of participants completed a second audiology assessment in the same location, at least 1 week apart, to examine reliability.

Assessment Measures

Screening

The Test of Premorbid Functioning (TOPF¹⁵) is a brief single-word reading test used to estimate premorbid intellectual functioning. Inclusion required an estimated Full-Scale IQ of ≥ 70. The Mini Mental State Exam (MMSE¹⁶) is a commonly used screening tool that evaluates orientation, memory, attention, and language abilities. To ensure participants could comprehend and follow task procedures, the inclusion criteria for this study specified a MMSE score of 20 or higher.¹⁷

Audiologic Functioning

The Shoebox Portable Audiometer System assessment followed existing protocols used in NIH-funded longitudinal studies.¹⁸ Equipment includes an iPad with ShoeBox Audiometry software and RadioEar DD450¹⁹ headphones. All testing was performed in a quiet room with ambient sound monitored continuously via the built-in sound meter. Testing was paused or repeated with noise interruptions above 45 dB. Unaided pure tone air conduction audiometry was performed to determine hearing thresholds at frequencies across a broad spectrum of human hearing (500 Hz to 8 kHz). Pure tone signals (beeps) of varying intensities (measured as decibels, dB) were presented to each ear. The participant responded to the signal by raising their hand. The audiometric threshold for the left and right ears was defined as the lowest intensity at which the participant was able to detect the signal. The pure tone average (PTA) was established for each ear across 4 frequencies: 500, 1k, 2k, and 4k Hz. Consistent with prior epidemiologic studies,^4,20,21 better ear PTA defined the hearing threshold, where higher threshold (dB) indicates worse hearing. Severity of hearing loss was classified categorically using clinical and research cut points including subclinical (16–25 dB), mild (26–40 dB), moderate-severe (41–70 dB), and profound (>70 dB) hearing loss.^4,20–22

Additional aspects of perceived audiologic functioning were assessed via self-report. Participants were asked if they experience tinnitus, and the Tinnitus Handicap Inventory (THI²³) provided a self-report evaluation of difficulties they may experience because of tinnitus. The THI total score ranges from 0–100, with higher scores indicating greater negative impact or perceived handicap due to tinnitus. The Hearing Handicap Inventory for the Elderly Screening Version (HHIE-S²⁴) is a brief questionnaire that assesses how an individual perceives the social and emotional effects of any experienced HL. The HHIE-S total score ranges from 0–40, with higher scores indicating greater perceived handicap or negative impact of hearing loss on daily functioning.

Population-based Audiometry

To aid in the interpretation of results, HL data from the U.S. general population of adults was derived from the National Health and Nutrition Examination Survey (NHANES²⁵). Five data cycles from 2005 through 2018 were merged for analysis, resulting in 3,360 total participants in the age range of 50–70 (inclusive). In NHANES, hearing is assessed with pure tone audiometry. We calculated the better ear PTA and defined HL as described above. Sample weighting, per NHANES protocol, was used so that HL prevalence is representative of the U.S. population. Additionally, direct standardization was performed so that the NHANES cohort matched the study cohort for race, ethnicity, sex, and age group.

Psychosocial Functioning

The Heinrichs-Carpenter Quality of Life Scale (QoL²⁶) is an interviewer-rated measure of functioning in four domains: interpersonal functioning, intrapsychic foundations, instrumental role function, and common objects/activities. The abbreviated 7-item version of the QoL was used to rate psychosocial functioning.^27,28 Participants rated perceived loneliness using the 20-item UCLA Loneliness Scale version 3 (UCLA-LS²⁹), a widely used self-report measure of loneliness.

Psychiatric Symptoms

The Structured Clinical Interview for the Positive and Negative Syndrome Scale (SCI-PANSS³⁰) assessed positive, negative, and general psychopathology symptom severity in the past 2 weeks. Positive and negative symptom subscale scores each range from 7 to 49 and general symptom subscale scores range from 16 to 112. Higher scores indicate greater symptom severity.

Analyses

Hearing ability was characterized continuously by threshold and categorically by HL severity. Feasibility of community clinic-based audiometry was characterized by the percentage of participants who completed the evaluation with <45 dB ambient noise. Test-retest reliability was examined in a subset of participants via Pearson’s correlations for left and right ear PTA. Descriptive statistics characterized sample mean symptom severity, the frequency of medical co-morbidities associated with HL risk (diabetes/prediabetes, hypertension, cardiovascular disease, cerebrovascular disease), tinnitus, and the perceived impact of tinnitus or HL as reported on the THI and HHIE-S, respectively. A bivariate correlation tested the specific relationship between positive symptom severity and hearing performance. The relationship between audiologic and psychosocial functioning as rated on the QoL and reported on the UCLA-LS was examined with bivariate Pearson’s correlations.

RESULTS

Sample Characteristics

The total sample consisted of N=41 participants with a diagnosis of schizophrenia and schizoaffective disorder, with 23 participants aged 50–59 years and 18 participants aged 60–70 years. The mean (SD) age was 59.8 (5.6) years. Sociodemographic and clinical characteristics are presented in Table 1. All participants were prescribed at least 1 antipsychotic medication. HL and medical conditions associated with HL risk co-occurred in 55% of the sample, including diabetes/prediabetes (50%), hypertension (52.5%), and cardiovascular disease (7.5%).

Table 1.

Sociodemographic and Clinical Characteristics

	Ages 50–59 n = 23	Ages 60–70 n = 18	Total Sample N = 41
Sex n (%)
Female	8 (34.8)	6 (33.3)	14 (34.1)
Male	14 (60.9)	12 (66.7)	26 (63.4)
Other	1 (4.3)	-	1 (2.4)
Ethnicity n (%)
Hispanic/Latinx	5 (21.7)	3 (16.7)	8 (19.5)
Not Hispanic/Latinx	18 (78.3)	15 (83.3)	33 (80.5)
Race n (%)
Asian	-	1 (5.6)	1 (2.4)
Black	11 (47.8)	13 (72.2)	24 (58.5)
White	8 (34.8)	1 (5.6)	9 (22.0)
More than 1 race	3 (13.0)	3 (16.7)	6 (14.6)
Unknown	1 (4.3)	-	1 (2.4)
Years of Education
mean (SD)	12.13 (2.47)	11.67 (3.2)	11.93 (2.79)
FSIQa Estimate
mean (SD)	99.96 (18.63)	100.78 (16.37)	100.32 (17.46)
PANSSb Subscale Scores
Positive mean (SD)	13.2 (6.3)	10.9 (4.2)	12.2 (5.5)
Negative mean (SD)	12.4 (4.9)	10.9 (3.9)	11.7 (4.5)
General mean (SD)	25.8 (6.5)	24.5 (4.4)	25.2 (5.7)

^aFull Scale IQ (FSIQ) Estimate was derived from the Test of Premorbid Functioning

^bPositive and Negative Syndrome Scale

Audiological Functioning

Of 41 evaluated, 97.56% of participants were able to complete the audiology assessment (N=40). All assessments were completed with <45 dB ambient noise. The mean (SD) hearing threshold (pure tone average; PTA) for the full sample (N=40) was 24.48 (12.66) dB. In participants ages 50–59, the mean (SD) hearing threshold was 23.48 (12.93) dB, versus 25.82 (12.54) dB in participants ages 60–70. Presence of HL (PTA>15 dB) was distributed across the range of ages, with 70% demonstrating some degree of HL, including 35% who demonstrated subclinical HL and 35% who demonstrated mild or worse HL. Table 2 presents the severity of hearing loss in this sample relative to a demographically matched sample drawn from the general adult U.S. population. Test-retest audiometry data (N=11) yielded large Pearson’s correlation coefficients for left (r(9)=0.96) and right (r(9)=0.93) ears.

Table 2.

Rates of Hearing Loss by Severity Category in the Study Sample and U.S. General Population Age 50–70

Hearing Loss Category	Sample Rate n (%)a	U.S. Rate unweighted n (weighted %)b
Normal Hearing (≤ 15 dB)	12 (30.0)	1,818 (58.8)
Subthreshold HL (16–25 dB)	14 (35.0)	940 (26.8)
Mild HL (26–40 dB)	8 (20.0)	484 (11.8)
Moderate-Severe HL (41–70 dB)	5 (12.5)	113 (2.5)
Profound HL (>70 dB)	1 (2.5)	5 (0.08)

^aSample rates represent the rates of hearing loss within the sample of study participants (N = 40)

^bU.S. rates represent those in the general population ages 50–70 drawn from the National Health and Nutrition Examination Survey. n represents unweighted numbers. % represents weighted prevalence adjusted for the sociodemographic makeup of the United States and additionally matched to the sample cohort for age, sex, and race/ethnicity (N=3,360).

All participants (N=41) completed self-report measures of audiological functioning. Participants were first asked whether they experienced hearing loss or tinnitus; 6 participants (14.6%) reported some hearing loss, and 9 participants (22%) self-reported tinnitus. The majority perceived no handicap associated with hearing ability (80.5%) or tinnitus (85.4%). Those who did perceive some hearing handicap rated it moderate (12.2%) or severe (7.3%); those who perceived handicap due to tinnitus rated it as mild to moderate (12.2%) or catastrophic (2.4%).

Association with Psychiatric Symptoms

Positive, negative, and general psychiatric symptoms were overall minimal to mild (see Table 1). The bivariate correlation between hearing threshold and the PANSS positive symptom subscale was not significant (r(38)=0.07, p=0.68).

Psychosocial Impact

The results of the bivariate correlation analysis are presented in Table 3. Most notably, greater loneliness (UCLA-LS) was significantly correlated with poorer audiological performance (higher PTA; r(38)=.475, p<0.01) and greater perceived hearing handicap (HHIE-S; r(38)=.480, p<.01).

Table 3.

Bivariate Correlations Between Audiologic and Psychosocial Functioning (N=40)

	PTA	HHIE-S	UCLA-LS	QoL
PTA	-	.432**	.475**	−.044
HHIE-S	.432**	-	.480**	.020
UCLA-LS	.475**	.480**	-	−.303
QoL	−.044	.020	−.303	-

^**p < .01

PTA = Pure Tone Average, HHIE-S = Hearing Handicap Inventory for the Elderly Screening Version, UCLA-LS = UCLA Loneliness Scale, QoL = abbreviated Heinrichs-Carpenter Quality of Life Scale

CONCLUSIONS

This study evaluated older adults with schizophrenia in order to replicate and extend prior research indicating higher rates of hearing loss (HL) in younger people with schizophrenia than would be expected given age-based rates in the U.S. general population. While aging increases risk for HL, schizophrenia may put people at additional risk. Given that the average life expectancy in schizophrenia is under 65 years, people with schizophrenia in the 50–70 year age group are considered older,^12,13 and were the focus of this study. The clinical significance of HL was evaluated by including measures of perceived functional impact, loneliness, and clinician-rated quality of life. By bringing methods of measurement comparable to those of epidemiological studies of HL to the community clinic, this study also sought to explore the feasibility of conducting brief assessments of hearing ability in routine practice settings.

Consistent with recent findings¹⁰, people with schizophrenia spectrum disorder aged 50–70 evidenced elevated HL rates relative to U.S. adults of the same age, sex, race, and ethnicity. Hearing threshold was significantly associated with subjective report of hearing-related impairment, though few participants rated the functional impact as being significant. Nonetheless, those with poorer hearing, whether measured objectively or subjectively, reported greater experienced loneliness. While interviewer-rated quality of life was not associated with hearing ability in this sample, the significant association between loneliness and hearing ability aligns with previous studies highlighting the strong link between hearing loss and social isolation among older adults.^2,5 Notably, even subthreshold levels of HL may significantly impact an individual’s ability to communicate and engage with others, leading to feelings of loneliness and social exclusion.⁵

Given the potential impact of any hearing impairment, it is critical to develop scalable methods of assessment. Mental health providers have a unique opportunity to incorporate hearing loss screening into routine health assessments and make appropriate referrals for further evaluation and treatment.^31,32 Though the Shoebox portable audiometer system has been validated for use in the general population,³³ it was not previously known whether its use could be feasibly implemented to reliably measure hearing ability in people with schizophrenia. Indeed, assessment with the portable audiometer system was feasible, with a high percentage of participants able to complete the evaluation in a community clinic setting without interference. Further, although the re-test sample was small, high correlations between pure tone average thresholds provide preliminary evidence that the assessment is reliable.

Study results suggest self-reports of hearing loss will underestimate prevalence. Whereas 35% of the participants were classified as having mild or worse hearing loss on audiometry, only about 15% reported having hearing loss. Since self-report measures may be biased due to social desirability or lack of awareness of hearing loss, audiometry assessment should be considered a more robust measure for identifying hearing impairment in community settings. Audiometry systems, such as Shoebox, may hold promise for routinely implementing screening in settings where people with schizophrenia spectrum disorder are likely to receive other behavioral healthcare services. Screening results can then be used to make appropriate referrals for further assessment and intervention.

The reasons for elevated hearing thresholds and rates of HL in this sample are unknown. Prior research suggests that psychiatric symptoms are not responsible for poorer performance on audiology assessments, and data in this study would suggest the same; positive symptoms (e.g., hallucinations, conceptual disorganization) were not associated with audiometry performance. Notably, medical comorbidities associated with HL risk were present in this sample, though causality cannot be inferred. Hearing may be one of several physical health outcomes that require routine assessment and monitoring. Epidemiological research has emphasized the study of modifiable risk factors in order to develop HL prevention strategies.^34,35 Nonetheless, hearing-related impairment is modifiable through HL screening and treatment. Identifying the need for treatment is important given the implications of HL for other social and mental health outcomes, including social isolation and depression. In addition, schizophrenia spectrum disorder is already associated with cognitive impairment, but given results of previous research^3,4,11 there is concern that HL could further contribute to early cognitive decline.³⁶

Several limitations should be taken into consideration. The findings may only be applicable to the population and environments in which the study was carried out. Replication in a larger sample with diverse representation is required. The re-test assessment used a small sample and relatively brief time between assessments, which may have impacted reliability estimates. While traditional audiometry evaluations have provided validation of the ShoeBox portable audiometry in nonpsychiatric populations, those “gold standard” evaluations were not performed with this sample. Future research can replicate and build upon these findings by using larger and more diverse samples, employing longitudinal designs, and investigating the possible advantages of interventions like hearing aids,³⁷ or auditory cognitive training programs^38–40 in enhancing the psychosocial well-being of this population.

Investigative focus on etiology and risk factors will facilitate better understanding of the increased prevalence of HL in schizophrenia and inform service needs, from risk assessment and mitigation, to screening, and treatment. As the results of this study indicate, HL in people with schizophrenia is associated with experienced loneliness. Enhancing our understanding of hearing ability in this diagnostic population has the potential to improve individuals’ overall well-being and quality of life.

HIGHLIGHTS.

• This study evaluated whether older adults with a schizophrenia spectrum diagnosis evidence higher rates of hearing loss relative to historical data from U.S. adults of the same age.

• People with schizophrenia spectrum disorder aged 50–70 evidenced elevated hearing loss rates. Hearing threshold was significantly associated with subjective report of hearing-related impairment and greater experienced loneliness.

• Evidence of increased prevalence of hearing loss in older adults with schizophrenia, which was associated with greater experienced loneliness, indicates need for routine risk assessment and mitigation, screening, and treatment with the goal of improving well-being and quality of life.

DATA STATEMENT

Part of this work was presented at the Schizophrenia International Research Society in Toronto, Ontario, Canada on May 13, 2023.