 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a Centre for Community Child Health
h Clinical Epidemiology and Biostatistics Unit
i Centre for Adolescent Health, Royal Children's Hospital, Parkville, Australia
b Murdoch Childrens Research Institute, Parkville, Australia
c Speech Language and Hearing Sciences, University of Arizona, Tucson, Arizona
d Hearing Research Group
e Department of Philosophy
f Department of Otolaryngology
g Faculty of Education, University of Melbourne, Melbourne, Australia
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
METHODS. A cross-sectional, cluster-sample survey of 6581 children (response: 85%; grade 1: n = 3367; grade 5: n = 3214) in 89 schools in Melbourne, Australia, was performed. Slight/mild bilateral sensorineural hearing loss was defined as a low-frequency pure-tone average across 0.5, 1, and 2 kHz and/or a high-frequency pure-tone average across 3, 4, and 6 kHz of 16 to 40 dB hearing level in the better ear, with air/bone-conduction gaps of <10 dB. Parents reported children's health-related quality of life and behavior. Each child with slight/mild bilateral sensorineural hearing loss, matched to 2 normally hearing children (low-frequency pure-tone average and high-frequency pure-tone average of 
15 dB hearing level in both ears), completed standardized assessments. Whole-sample comparisons were adjusted for type of school, grade level, and gender, and matched-sample comparisons were adjusted for nonverbal IQ scores.
RESULTS. Fifty-five children (0.88%) had slight/mild bilateral sensorineural hearing loss. Children with and without sensorineural hearing loss scored similarly in language (mean: 97.2 vs 99.7), reading (101.1 vs 102.8), behavior (8.4 vs 7.0), and parent- and child-reported child health-related quality of life (77.6 vs 80.0 and 76.1 vs 77.0, respectively), but phonologic short-term memory was poorer (91.0 vs 102.8) in the sensorineural hearing loss group.
CONCLUSIONS. The prevalence of slight/mild bilateral sensorineural hearing loss was lower than reported in previous studies. There was no strong evidence that slight/mild bilateral sensorineural hearing loss affects adversely language, reading, behavior, or health-related quality of life in children who are otherwise healthy and of normal intelligence.
Key Words: child development • language • hearing loss • prevalence • outcomes
Abbreviations: SNHL—sensorineural hearing loss • LPTA—low-frequency pure-tone average • HPTA—high-frequency pure-tone average • HRQoL—health-related quality of life • CI—confidence interval • NHANES—National Health and Nutrition Examination Survey • HHANES—Hispanic Health and Nutrition Examination Survey • UNHS—universal newborn hearing screening • PTA—pure-tone average • HL—hearing level
Slight/mild bilateral sensorineural hearing loss (SNHL) is reported to be a prevalent childhood condition that may result in significant language and academic deficits.1–5 In US National Health and Nutrition Examination Survey (NHANES) III,6 1.5% of school-aged children had bilateral low-frequency hearing loss and >3% bilateral high-frequency loss of slight/mild degree. In the earlier NHANES II and the parallel US Hispanic Health and Nutrition Examination Survey (HHANES), low-frequency losses averaging >15 dB hearing level (HL) in the better ear affected 1.5% to 6.8% of children 6 to 19 years of age.7
There is debate regarding whether systematic efforts should be made to detect these children and for what (if any) services they should be eligible. These questions are especially pertinent with universal newborn hearing screening (UNHS) now being standard care in much of the world. Because typically UNHS targets moderate or worse bilateral congenital SNHL, some view mild losses as unwanted false-positive results of UNHS.8 Others take the opposing view that mild losses are numerous and often are not detected by current UNHS programs. Recently, Johnson et al9 called for programs to consider including mild losses within their target group, despite the fact that the technical program changes needed to achieve this higher sensitivity would reduce program specificity and greatly increase referral and false-positive rates. Hearing screening programs offered near the time of school entry also may detect mild SNHL; however, in the many regions that have discontinued such programs, mild SNHL may remain undetected well into the school years.
Addressing these issues requires accurate, population-based knowledge of the prevalence and effect of slight/mild SNHL. Unfortunately, existing population studies probably overestimated prevalence, either by combining conductive and sensorineural losses in their estimates6,7 or by relying on response rates of <50%, increasing the possibility of ascertainment bias.1 Several small outcome studies suggested that slight/mild hearing loss may not be associated with adverse outcomes10–12 or children may experience early difficulties that resolve.4 Findings on educational attainment and grade retention are inconsistent,1 and no studies have examined health-related quality of life (HRQoL), an important component for understanding the burden of any condition. Although otitis media with effusion has different pathophysiologic features, its associated hearing loss is typically in the slight to mild range (mean: 25 dB HL) and may be prolonged. Associations between persistent early otitis media with effusion and later speech and language outcomes in otherwise normally developing children are now known to be either weak or absent,13 and reversal of otitis media with effusion with tympanostomy tubes does not confer developmental or behavioral benefit.14
This article reports on a population study designed to ascertain the prevalence of slight/mild bilateral SNHL in a large community sample of elementary school-aged children and its associations with language, academic achievement, behavior, and HRQoL. Unlike preschool-aged children, school-aged children can cooperate with pure-tone audiometry; therefore, ascertainment of even the mildest losses should be accurate, literacy and academic achievement can be measured, and the perspectives of the children themselves can be included.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
6000 children, identified a target sample of 7784 grade 1 and grade 5 children in 89 elementary schools. Assuming a true prevalence of slight/mild SNHL of 3% (on the basis of the NHANES data), the anticipated sample size of 6000 would be large enough to estimate this with a SE of 0.22%. Grades 1 and 5 represent the second and sixth years of formal schooling for children in the state of Victoria, during which standardized, uniform, academic achievement results are available for all students. Parents were contacted via an information package sent home from school, and they returned a completed questionnaire with their written consent. All children with consent underwent otoscopy and screening audiometry (see "Measures"). In phase 2, participants were drawn from the base sample for more in-depth study of the language, academic, social, and HRQoL outcomes of slight/mild bilateral SNHL. All children who met the study criteria for slight/mild bilateral SNHL (see "Measures") were invited to participate in additional assessments, and those who consented were matched to 2 normally hearing children with respect to age, gender, grade level, and school attended. Exclusion criteria were known intellectual disability or major medical disorder reported by the parent, known syndromic cause for hearing loss, or hearing loss attributable solely to otitis media with effusion. We did not exclude children with language impairment, because the relationship of language abilities to hearing impairment was of specific interest in this study. Each child underwent a 3-hour assessment with pairs of research assistants, 4 of whom were blinded with respect to hearing status; the fifth assistant was blinded for some but not all assessments.
Measures
Audiometry was conducted under soundproof conditions by 8 audiologists certified by the Audiological Society of Australia. Each screen consisted of 3 presentations to each ear of each test frequency (0.5, 1, 2, 3, 4, 6, and 8 kHz) at 15 dB HL; a pass at a given frequency was noted if the child responded correctly to 2 of 3 presentations at 15 dB HL. If any frequency was not passed, then the audiologist proceeded to a full audiometric evaluation, including pure-tone air-conduction and bone-conduction threshold tests and tympanometry. The air-conduction pure-tone thresholds were averaged at 0.5, 1, and 2 kHz to indicate the low-frequency pure-tone average (LPTA) and at 3, 4, and 6 kHz to indicate the high-frequency pure-tone average (HPTA), classified as described by Niskar et al.6 We defined slight/mild SNHL as LPTA and/or HPTA of 16 to 40 dB HL in the better ear, with air/bone-conduction gaps of <10 dB. Normal hearing was defined as LPTA and HPTA of 15 dB HL in both ears.
Parents provided sociodemographic details and completed standardized measures of child behavior and HRQoL for the entire cohort (Table 1). Table 1 shows the language and psychological assessments and the child-completed questionnaire measures used for the phase 2 subsample. Grade 5 children also reported on their perceptions of how their hearing affected their daily lives.
|
Finally, we conducted 2 sets of exploratory analyses. Firstly, we looked for correlations between severity of hearing loss within the slight/mild range and outcomes, by using Pearson's correlation coefficient. This required that we assign a pure-tone average (PTA) value to each of the 55 children with slight/mild bilateral SNHL. Generally, this was the better-ear LPTA or HPTA or the 4-frequency (0.5, 1, 2, and 4 kHz) PTA for those with "flat" loss (ie, slight/mild SNHL at both high and low frequencies), as appropriate. Secondly, we looked for evidence that the outcome differences between the slight/mild SNHL and normal-hearing groups were greater or smaller in older children than younger children, by using tests of group-grade level interaction effects on the outcomes.
Data were analyzed by using SPSS 12.0 for Windows (SPSS, Chicago, IL) and Stata 8.0 (Stata Corp, College Station, TX). The study was approved by the Royal Children's Hospital Ethics in Human Research Committee (reference number 22056A), the Victorian Government Department of Education, and the Catholic Directorate of Education.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
Correlation coefficients for the correlation between severity of hearing loss and outcomes within the SNHL group were not significantly different from 0, with the exception of a positive correlation between children's self-reported levels of hearing-related problems on the emotional scale and their PTA values (Pearson r = 0.53; P = .003). Table 5 shows that tests for interactions did not provide strong evidence that outcome differences between the slight/mild bilateral SNHL and normal-hearing groups in the main summary measures of language, reading, and phonologic short-term memory were greater or smaller in grade 5 than grade 1.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Because of the study's large sample size, random population sampling, very high response rate, and careful ascertainment of slight/mild bilateral SNHL, we have confidence in the prevalence estimates. Significant bias in outcome assessment is also unlikely. Parents reported on child behavior and HRQoL before the audiometric assessments, when most parents were unaware of their child's hearing loss and therefore were not biased by the knowledge, and direct assessments were undertaken by researchers who almost always were blinded to the child's hearing status. The high response rate and low potential for bias may explain much of the difference in conclusions between our study and that of Bess et al.1 The fact that children underwent their hearing tests before reporting on their own HRQoL did not seem to affect their scores, although a nearly significant trend for higher Hearing Handicap Inventory scores was noted for grade 5 children.
The study has some limitations. Firstly, although the lower-than-anticipated number of cases of slight/mild bilateral SNHL increased the precision of the prevalence estimate, it also reduced the size of the "exposed" sample and therefore the available power with which to draw conclusions about the impact of the condition. Point estimates for all measures (except phonologic skills) indicated similar functioning but, on the basis of the 95% CIs provided, we cannot rule out real but clinically small effects (in either direction) of slight/mild bilateral SNHL.
Secondly, because this was a cross-sectional study, it was not possible to separate congenital and acquired losses or to determine whether any cases of slight/mild bilateral SNHL had originated or progressed very recently. A prospective study would be needed to determine this, but such a study would be challenging (if not impossible) because of the technical difficulty of ascertainment of very mild hearing loss in infants, toddlers, and preschool-aged children and the extremely large number of children who would need repeated hearing tests for ascertainment of those who acquire losses. More cases of hearing loss were detected in the grade 5 children, which might suggest that a proportion of the older children had acquired or progressive losses, which might not have the same impact as congenital losses. Data analyses did not indicate differential effects according to age group, but the numbers were small.
Thirdly, we excluded children with unilateral hearing loss and so could not study those effects. Ninety-five children (1.52%) had a LPTA and/or HPTA of >15 dB HL in one ear and normal hearing in the other ear. These were almost invariably minimal losses, with only 10 children (prevalence: 0.16%; 95% CI: 0.08%–0.29%) meeting the definition by Bess et al1 of unilateral losses of 20 dB HL in the worse ear and normal hearing in the other ear, compared with 3.0% in the study by Bess et al.1 The low prevalence precluded meaningful comparison of possible language, behavior, and achievement effects between those with unilateral losses and normal hearing, but it also suggests a small population burden from this condition, even if some effects were present for the individuals affected.
Finally, few of the children approached the upper end of the mild range (up to 40 dB HL), where adverse effects would be most likely. Therefore, this study cannot determine whether there is a specific level of hearing impairment in the upper mild/lower moderate range beyond which reductions in language skills and academic achievement and other problems increase rapidly in prevalence or magnitude. We note that our exploratory analyses did not suggest a relationship between poorer outcomes and more severe hearing loss within the slight/mild SNHL range, which probably reflects the population ascertainment of the sample. Our previous study of children with congenital bilateral hearing loss with hearing amplification showed that children with mild losses had poorer spoken language skills and academic achievement than the normative samples.5 Similarly, Yoshinaga-Itano et al26 found that children receiving early intervention services for early-detected, as opposed to later-detected, mild, congenital, bilateral hearing loss had better language outcomes during the preschool years. There are several important differences between those 2 studies and the study reported here, including (1) the main research questions, (2) the later age at which hearing loss and outcomes were ascertained, and (3) the population-based, rather than clinical, sampling method.
At 0.88% (95% CI: 0.66%–1.15%), the prevalence of slight/mild bilateral SNHL was lower than reported previously.1–3 This is plausible, because it likely reflects the rigorous exclusion of conductive hearing loss (through air-conduction and bone-conduction threshold tests) in all potential cases, which was not possible in the NHANES and HHANES. The difference in prevalence rates between NHANES/HHANES (
3%) and our study (slightly less than 1%) could be fully explained by an underlying 2% prevalence of conductive hearing loss, which would be entirely in keeping with the known epidemiologic features of conductive hearing loss. It seems unlikely that Australian children would suffer lower rates of hearing loss than US and European children. Hearing risk factors and the prevalence of moderate or greater bilateral congenital hearing loss in Australia are similar to those of other developed countries, and the general health levels (measured with the Pediatric Quality of Life Inventory in population studies) of US children27 and Australian children28 are also similar. Higher systematic early detection rates cannot explain these findings, because (1) Victoria was not offering UNHS when these children were born, (2) universal school-entry hearing screening ceased in Victoria in 1997 (ie, the year before the oldest children in this sample entered the school system), and (3) very few of the cases detected had been diagnosed previously.
Two other points regarding prevalence merit additional discussion. Firstly, the prevalence of bilateral slight/mild SNHL seemed higher in children of Asian origin (Table 2). Subsequent genetic analysis revealed that the excess prevalence over the base rate could be attributed to the relatively large proportion of Asian children who are homozygous for the V37I change in the GJB2 gene, coding for connexin 26.29 We estimated that the carrier frequency of the V37I change in Asian children in our cohort was 16%, whereas the frequency in children of non-Asian background was <2%. We concluded from the genetic studies that the V37I change is a mutation associated with hearing loss and that it is a common cause of slight/mild SNHL in children of Asian ethnicity.29 Secondly, children in grade 5 had a higher prevalence of slight/mild SNHL than did those in grade 1 (1.22% vs 0.56%; difference: 0.66%; 95% CI: 0.19%–1.12%; P = .006). This could represent a true increase in slight/mild bilateral SNHL, with new (incident) cases developing after grade 1.
Slight/mild SNHL led to a reduction in phonologic processing abilities, but this did not translate into poorer functioning in a range of child developmental, behavioral, and academic domains, including reading. Again, these results are plausible and are consistent with the report by Briscoe et al10 that children with mild/moderate SNHL were not impaired in tests of language, literacy, comprehension, and sentence and digit recall but were as impaired as children with specific language impairment in nonword repetition, phonologic awareness, and phonologic discrimination. The phonologic deficit is likely caused by the peripheral nature of the hearing loss, and it is unlikely that performance would improve over time. Longitudinal studies would be needed to determine whether these deficits persist or progress. These findings are also consistent with outcomes found among normally developing children with a history of otitis media with effusion who experienced similar levels of hearing loss.30
This study should provide reassurance to professionals counseling parents of children with hearing loss in this range who are otherwise healthy and of normal intelligence. The study does not provide data to support screening in the early school years to detect children with slight/mild hearing loss. However, additional research may still be required to determine whether children with bilateral hearing loss toward the upper end of the mild range experience adverse outcomes at a population level and, if so, whether they gain more benefit than harm from systematic intervention.
|
ACKNOWLEDGMENTS |
|---|
We acknowledge sincerely the contribution of the field workers responsible for data collection, and we thank all of the children, families, and schools for participating in the study.
|
FOOTNOTES |
|---|
Address correspondence to Melissa Wake, MBChB, MD, Centre for Community Child Health, Royal Children's Hospital, Flemington Rd, Parkville 3052, Australia. E-mail: melissa.wake{at}rch.org.au
Dr Wake had full access to all of the data in the study, takes responsibility for the integrity of the data and the accuracy of the data analysis, and is the guarantor.
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Brunnberg, M. L. Bostrom, and M. Berglund Self-Rated Mental Health, School Adjustment, and Substance Use in Hard-of-Hearing Adolescents J. Deaf Stud. Deaf Educ., July 1, 2008; 13(3): 324 - 335. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Tharpe Unilateral and Mild Bilateral Hearing Loss in Children: Past and Current Perspectives Trends in Amplification, March 1, 2008; 12(1): 7 - 15. [Abstract] [PDF]
|
|
|
|
|
|
A. M. Tharpe and D. P. Sladen Causation of Permanent Unilateral and Mild Bilateral Hearing Loss in Children Trends in Amplification, March 1, 2008; 12(1): 17 - 25. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2006 American Academy of Pediatrics. All rights reserved. |
||
|
||
