Universal neonatal hearing screening

Universal neonatal hearing screening
Universal Neonatal
Hearing Screening
November 2007
MSAC reference 17
Assessment report
© Commonwealth of Australia [YEAR]
ISBN <number>
ISSN (Print)
1443-7120
ISSN (Online) 1443-7139
First printed <add month/year>
Paper-based publications
© Commonwealth of Australia [add year]
This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be
reproduced by any process without prior written permission from the Commonwealth. Requests and
inquiries concerning reproduction and rights should be addressed to the Commonwealth Copyright
Administration, Attorney General’s Department, Robert Garran Offices, National Circuit, Barton ACT
2600 or posted at http://www.ag.gov.au/cca
Internet sites
© Commonwealth of Australia [add year of release of home page or electronic document]
This work is copyright. You may download, display, print and reproduce this material in unaltered form
only (retaining this notice) for your personal, non-commercial use or use within your organisation. Apart
from any use as permitted under the Copyright Act 1968, all other rights are reserved. Requests and
inquiries concerning reproduction and rights should be addressed to Commonwealth Copyright
Administration, Attorney General’s Department, Robert Garran Offices, National Circuit, Barton ACT
2600 or posted at http://www.ag.gov.au/cca
Electronic copies of the report can be obtained from the Medical Service Advisory Committee’s Internet site
at http://www.msac.gov.au/
Printed copies of the report can be obtained from:
The Secretary
Medical Services Advisory Committee
Department of Health and Ageing
Mail Drop 106
GPO Box 9848
Canberra ACT 2601
Enquiries about the content of the report should be directed to the above address.
The Medical Services Advisory Committee (MSAC) is an independent committee which has been
established to provide advice to the Minister for Health and Ageing on the strength of evidence available
on new and existing medical technologies and procedures in terms of their safety, effectiveness and costeffectiveness. This advice will help to inform government decisions about which medical services should
attract funding under Medicare.
MSAC recommendations do not necessarily reflect the views of all individuals who participated in the
MSAC evaluation.
This report was prepared on behalf of the Medical Services Advisory Committee by Tracy Merlin, Hedyeh
Hedayati, Thomas Sullivan, Liz Buckley, Skye Newton, Brent Hodgkinson, Petra Bywood, Fiona Jenner,
John Moss and Janet Hiller from Adelaide Health Technology Assessment (AHTA), Discipline of Public
Health, University of Adelaide, Adelaide, South Australia. The report was edited by Jo Mason from
MasonEdit, South Australia. This recommendation was endorsed by the Minister for Health and Ageing
on <date>.
Publication approval number: <number>
Contents
Executive summary................................................................................................. ix
Introduction ............................................................................................................14
Rationale for assessment............................................................................................... 14
Rationale for universal neonatal hearing screening .................................................. 14
Background.............................................................................................................16
Neonatal hearing testing in Australia.......................................................................... 16
What is permanent childhood hearing impairment? ................................................ 17
Testing for hearing impairment in neonates.............................................................. 20
Marketing status of the screening tests....................................................................... 23
Current reimbursement arrangement ......................................................................... 23
Approach to assessment ........................................................................................ 24
Objective......................................................................................................................... 24
Research questions ........................................................................................................ 24
Assessment of screening programs............................................................................. 24
Review of literature ....................................................................................................... 28
Expert advice.................................................................................................................. 35
Results of assessment ............................................................................................ 36
How prevalent is permanent childhood hearing impairment?................................ 36
Safety of universal neonatal hearing screening.......................................................... 42
How accurate are the screening tests? ........................................................................ 51
Is it effective to screen all neonates for hearing impairment?................................. 59
What are the economic considerations?................................................................... 100
Economic model ................................................................................................... 106
Possible hearing testing programs for neonates...................................................... 107
Short-term costs........................................................................................................... 109
Cost-effectiveness of screening protocols ............................................................... 113
Incremental cost-effectiveness ratios........................................................................ 124
Long-term costs ........................................................................................................... 125
Government expenditure ........................................................................................... 134
Discussion............................................................................................................. 142
Prevalence of permanent childhood hearing impairment...................................... 142
Safety of universal neonatal hearing screening........................................................ 142
Effectiveness ................................................................................................................ 144
Diagnostic accuracy of the screening tests............................................................... 144
Effectiveness of universal neonatal hearing screening........................................... 146
Cost-effectiveness of universal neonatal hearing screening .................................. 148
Conclusions........................................................................................................... 153
Clinical need ................................................................................................................. 153
Universal neonatal hearing screening
iii
Safety ............................................................................................................................. 153
Effectiveness ................................................................................................................ 153
Economic considerations ........................................................................................... 154
Research implications.................................................................................................. 154
Implementation issues ................................................................................................ 155
Recommendation.................................................................................................. 157
Appendix A
MSAC terms of reference and membership.................................. 158
Appendix B
Advisory Panel, Evaluator and Project Manager.......................... 160
Appendix C
Search strategies ............................................................................ 162
Appendix D Internet sites searched................................................................... 165
Appendix E
Critical appraisal checklists........................................................... 170
Appendix F Studies included in the review ....................................................... 176
Included studies on prevalence of permanent childhood hearing
impairment.................................................................................................................... 176
Included controlled and descriptive studies on screening safety.......................... 181
Included studies on diagnostic accuracy .................................................................. 185
Included controlled studies on effectiveness of screening .................................... 188
Included descriptive studies on effectiveness of screening ................................... 192
Included economic studies......................................................................................... 214
Appendix G
Excluded studies ........................................................................... 219
Appendix H Guidelines for using screening devices .......................................224
Use of the Natus Algo Portable AABR Screener ................................................... 224
Use of the ECHOCHECK hand-held ILO OAE Screener ................................. 228
Appendix I
Analysis of published economic evaluations (up to 2003)..............232
Glossary .................................................................................................................254
Abbreviations ........................................................................................................256
References .............................................................................................................258
iv
Universal neonatal hearing screening
Tables
Table 1
Neonatal hearing screening in Australian States and Territories as at
9th October 2007 ..................................................................................................... 16
Table 2
Neonatal hearing screening equipment listed on the ARTG ........................... 23
Table 3
Evidence dimensions.............................................................................................. 32
Table 4
Grading system used to rank included diagnostic studies ................................ 33
Table 5
Designations of levels of evidence* according to type of research
question (NHMRC 2005) ...................................................................................... 33
Table 6
Quality checklists .................................................................................................... 34
Table 7
Body of evidence assessment matrix ................................................................... 35
Table 8
Prevalence of permanent childhood hearing impairment (PCHI) .................. 37
Table 9
Safety of universal neonatal hearing screening (comparative studies) ............ 46
Table 10
Safety of universal neonatal hearing screening (noncomparative
studies)...................................................................................................................... 50
Table 11
Diagnostic accuracy of hearing screening tests .................................................. 56
Table 12
Effectiveness of universal neonatal hearing screening for secondary
outcomes (controlled studies) ............................................................................... 63
Table 13
Descriptive (uncontrolled) studies of 1-stage universal neonatal
hearing screening .................................................................................................... 71
Table 14
Descriptive (uncontrolled) studies of 2-stage universal neonatal
hearing screening .................................................................................................... 73
Table 15
Descriptive (uncontrolled) studies of 3-stage (or more) universal
neonatal hearing screening .................................................................................... 88
Table 16
Relevant resource items for a neonatal hearing screening program.............. 103
Table 17
Region-dependent design of UNHS program.................................................. 109
Table 18
Identification of 2003 unit costs for neonatal hearing screening................... 110
Table 19
2003 cost of targeted screening by method of AABR delivery for a
cohort of 4,000 infants per year.......................................................................... 115
Table 20
Transitional probabilities for an Australian targeted screening
program .................................................................................................................. 116
Table 21
Transitional probabilities for an Australian UNHS program (protocol
A)............................................................................................................................. 119
Table 22
Transitional probabilities for an Australian UNHS program (protocol
B) ............................................................................................................................. 121
Table 23
Incremental cost of UNHS by choice of 2-stage screening method............. 123
Table 24
Summary cost-effectiveness of three screening options for PCHI in a
birth cohort of 4,000 infants per year ................................................................ 124
Table 25
ICER of three screening options for PCHI in a birth cohort of 4,000
infants per year ...................................................................................................... 124
Universal neonatal hearing screening
v
Table 26
Resources used (in 2003 $AUD) in therapy, rehabilitation and
education of infants with PCHI to 18 years of age (or school Year 12) ...... 128
Table 27
2003 cost (discounted @ 5% p.a.) of extra rehabilitation and
education per child with bilateral PCHI dependent on age of
identification and whether unilateral or bilateral: best case scenario ............ 130
Table 28
Summary: Incremental yield and discounted 2003 costs ($million,
discounted @ 5% p.a.) for a UNHSa program over the lifetime of an
Australian annual birth cohort of 250,000 infants ........................................... 131
Table 29
One-way threshold analysis: Incremental yield and discounted 2003
costs ($million, discounted @ 5% p.a.) for a UNHSa program over
the lifetime of an Australian annual birth cohort of 250,000 infants............ 132
Table 30
Estimated additional government expenditure (all jurisdictions
combined) over the first 8 years of a national program of UNHS
(2003 costs in $’000)............................................................................................. 136
Table 31
Savings on transfer payments per child with unilateral or bilateral
hearing impairment............................................................................................... 139
vi
Universal neonatal hearing screening
Figures
Figure 1
Anatomy of the ear................................................................................................. 18
Figure 2
Clinical pathway for universal neonatal hearing screening ............................... 20
Figure 3
Summary of the process used to identify and select studies for the
assessment of universal neonatal hearing screening .......................................... 30
Figure 4
Age at first hearing aid fitting for children born between 1986 and
2002........................................................................................................................... 41
Figure 5
2003 timeline of identification of PCHI in Australia in the absence of
a UNHS program.................................................................................................. 106
Figure 6
Options for neonatal hearing screening programs .......................................... 108
Figure 7
Decision model for targeted hearing screening................................................ 113
Figure 8
Decision model for universal neonatal hearing screening .............................. 118
Figure 9
Two-way threshold analysis: combinations of proportions of hearingimpaired persons who are unemployed and PCHI infants who attain
normal language skills, where a UNHS program for an Australian
annual birth cohort will be less costly over the lifetime than no
organised screening program .............................................................................. 133
Figure 11 Decision tree for universal and targeted neonatal hearing screening
model (Keren et al 2002) ..................................................................................... 233
Figure 12 Decision tree for universal and targeted neonatal hearing screening
model (Kemper & Downs 2000)........................................................................ 236
Figure 13 Decision tree for modelled universal neonatal hearing screening
protocols (Kezirian et al 2001)............................................................................ 238
Figure 14 Decision tree for modelled universal neonatal hearing screening
protocols (Gorga et al 2001) ............................................................................... 240
Figure 15 Decision tree for modelled universal neonatal hearing screening
protocols (Boshuizen et al 2001) ........................................................................ 241
Figure 16 Screening protocols for existing 1- and 2-stage universal neonatal
hearing screening programs (Vohr et al 2001).................................................. 243
Figure 17 Protocol for universal screening of well and neonatal intensive care
unit babies (Gorga et al 2001)............................................................................. 245
Figure 18 Protocol for 2-stage TEOAE or AABR universal neonatal hearing
screening (Lemons et al 2002) ............................................................................ 246
Figure 19 Protocol for 2-stage TEOAE universal neonatal hearing screening
(Weirather et al 1997) ........................................................................................... 247
Figure 20 Protocol for 2-stage TEOAE universal neonatal hearing screening
(Maxon et al 1995) ................................................................................................ 248
Universal neonatal hearing screening
vii
Executive summary
Rationale for assessment
At the July 2002 Australian Health Ministers’ Conference (AHMC) in Darwin, the health
ministers considered a proposal by the Queensland Health Minister to establish a
universal neonatal hearing screening (UNHS) program for Australia. It was noted that
the type of screening protocol used would have a significant impact on both the direct
screening and the ongoing diagnostic costs, and that current Australian data on the cost
of establishing a UNHS program was not available. In order to inform decision-making
regarding this proposal, the AHMC requested the MSAC to commission a full health
technology assessment on the safety, effectiveness and cost-effectiveness of UNHS. This
would allow an estimation of the potential costs and cost-savings of such a program, and
assist with the development of (1) a national evidence-based screening protocol, (2)
guidelines/standards of practice, (3) a national minimum data set, and (4) mechanisms to
ensure program sustainability.
The impetus for universal screening has been the belief that targeted ‘risk factor’
screening fails to identify more than 50 to 60 per cent of all neonates who eventually
display some form of permanent hearing loss. Delays in the identification of hearing
impairment have been suggested to profoundly affect the quality of life of children in
terms of their communication skills and subsequent education and employment
prospects. With the development of transient evoked otoacoustic emissions (TEOAE)
tests and automated auditory brainstem response (AABR) tests, screening for hearing
impairment can take place when infants are only hours or days old. Potentially, this can
have an impact on tertiary prevention by improving the times to referral, diagnosis and
early management for hearing-impaired infants.
The procedure
Universal neonatal hearing screening allows for early identification and management of
permanent hearing impaired children (PCHI). This means obtaining hearing screening on
every infant born in the hospital as early as possible. While, to date, primary prevention
of hearing loss in neonates is not apparent, universal hearing screening may reduce the
disabilities that appear to be associated with PCHI, such as impairment in language
acquisition, learning and speech development. Initial efforts of screening children
concentrated on high-risk populations. This included children with a family or perinatal
history, or physical examination findings amongst other risk factors. Although, as
previously mentioned, this constitutes a large proportion of children, it fails to identify
up to 60 per cent of all neonates whom eventually display some form of permanent
hearing loss (Thompson et al 2001). Screening tools utilised for universal neonatal
hearing screening are the otoacoustic emissions (OAE) test and the automated auditory
brainstem response (AABR) test. These are objective measures that respond to different
stimuli; sound stimuli evoke otoacoustic emissions from outer hair cells, while AABR, a
neurological test of auditory brainstem function, responds to click or tone stimuli.
Currently, conventional auditory brainstem response testing is the gold-standard for the
diagnosis of hearing impairment in infants.
Universal neonatal hearing screening
ix
Medical Services Advisory Committee – role and approach
The Medical Services Advisory Committee (MSAC) was established by the Australian
Government to strengthen the role of evidence in health financing decisions in Australia.
The MSAC advises the Minister for Health and Ageing on the evidence relating to the
safety, effectiveness and cost-effectiveness of new and existing medical technologies and
procedures, and under what circumstances public funding should be supported.
A rigorous assessment of evidence is thus the basis of decision making when funding is
sought under Medicare. In 2003 a team from Adelaide Health Technology Assessment,
Discipline of Public Health at the University of Adelaide was engaged to conduct a
systematic review of literature on universal neonatal hearing screening. An advisory panel
with expertise in this area then evaluated the evidence and provided advice to the MSAC.
In 2007 Adelaide Health Technology Assessment was commissioned to update the
systematic literature review.
MSAC’s assessment of universal neonatal hearing screening
Clinical need
According to the international literature, the median prevalence of moderate to profound
(>35 dB) bilateral permanent childhood hearing impairment (PCHI) is 1.3 per 1,000
infants. The median prevalence of unilateral PCHI of similar severity is 0.6 per 1,000
infants.
There are no population-based data on the prevalence of congenital PCHI in Australia.
Using the median international prevalence estimate and multiplying this by the number
of yearly birth registrations in Australia (259,800) suggests that 325 Australian children
are born annually with moderate to profound bilateral PCHI. Unilateral PCHI of similar
severity is estimated to occur in an additional 156 children each year. Overall, it is
estimated that 481 Australian children are born each year with either unilateral or bilateral
moderate to profound PCHI.
Safety
None of the available studies reported any physical harm resulting from universal
neonatal hearing screening (UNHS).
The data available on the psychosocial harms from UNHS are of poor to average quality.
The most commonly reported psychosocial outcome of UNHS was maternal anxiety
regarding (1) the screen, (2) a false positive result and (3) a screen positive result. Overall
anxiety levels were within the normal range, so although anxiety levels were frequently
reported as higher when infants screened positive rather than negative, no clinically
important differences in anxiety level were found (level III-2 interventional evidence).
Likewise, no differences in anxiety were found between parents of unscreened babies, or
screened babies, regardless of whether the screening outcome was positive or negative
(level III-2 interventional evidence).
Maternal concern about infant’s hearing increased as the number of required tests
increased. Levels of depression increased and the quality of interactions with the infant
were reported to be statistically significantly lower when babies were screened as positive
(level III-2 interventional evidence). It has been suggested that screen status or anxiety
x
Universal neonatal hearing screening
may have an impact on the parental relationship with the child, but in the one controlled
study (level III-2 interventional evidence) that reported on parental attitudes to the child,
as distinct from quality of interactions, no differences were ascertained. No studies
reported on the psychosocial effects of false reassurance or of a true-positive diagnosis.
Diagnostic accuracy of the screening tests
Transient evoked otoacoustic emissions testing versus auditory brainstem
response testing
Limited but good quality evidence indicates that the accuracy of transient evoked
otoacoustic emissions (TEOAE) is influenced by the level of local ambient noise and the
condition of infant ears at testing. Under quiet conditions, the test has been found to
possess excellent sensitivity (up to 100%) and good specificity (92%). The ability of an
initial TEOAE test to correctly diagnose permanent childhood hearing impairment
(PCHI) is very low however, with a positive predictive value of 1.5 per cent. This is likely
a consequence of both the frequency of transient hearing losses in newborns due to ear
occlusion and the low prevalence of PCHI.
Automated auditory brainstem response testing versus auditory brainstem
response testing
One study on an early model of the automated auditory brainstem response (AABR) test
demonstrates that it possesses excellent specificity (96%) and good sensitivity (80%) in
diagnosing PCHI. The positive predictive value of AABR is very low (2.2%), although
marginally better than TEOAE conducted under quiet conditions. Expert opinion
indicates that later models of the AABR may have better diagnostic accuracy, although
this has yet to be confirmed empirically.
False positives associated with either test could be reduced with the introduction of a
second-stage or third-stage screen of initial failures, prior to diagnostic testing. This may,
however, result in unnecessary caregiver anxiety and added costs and delays in
rehabilitation.
Effectiveness
Change in clinical management
The best evidence (III-1 screening evidence) available indicates that universal neonatal
hearing screening has an impact on the clinical management of PCHI. Referral for
diagnostic testing, actual PCHI diagnosis, and management of PCHI commonly occurs
earlier and more frequently with universal neonatal hearing screening than without
universal neonatal hearing screening. Level III-1 screening evidence indicates that the
probability of referring an infant for diagnostic testing before the age of six months is
nearly three times more likely [RR=2.9, 95%CI 1.4, 6.3] (19 times when controlling for
the severity of hearing impairment) with universal neonatal hearing screening, than when
universal screening is not available. Infants born during periods of universal neonatal
hearing screening are twice as likely to receive a diagnosis of PCHI, than infants born in
periods without universal hearing screening [RR=2.3, 95%CI 1.1, 4.7]. The absolute
increase in benefit is small, however - an extra five children identified per 10,000 because of the low prevalence of the condition. There is also an indication that screening
may increase the likelihood of PCHI management before the age of ten months by nearly
Universal neonatal hearing screening
xi
two and a half times [RR=2.4, 95%CI 1.0, 5.8] (eight times when controlling for the
severity of PCHI).
Change in health outcomes
There is limited information available on the effect of universal neonatal hearing
screening on primary or patient-relevant outcomes. The results, from two cohort studies
(Level III-2 screening evidence) indicate that children with bilateral PCHI born in
hospitals with universal neonatal hearing screening have better receptive language, but
not expressive language abilities and communicative abilities than children born in
hospitals without screening. Information on the impact of universal neonatal hearing
screening programs on the longer term outcomes (i.e. educational and employment
status) has yet to be reported.
Descriptive data indicate that the majority of universal neonatal hearing screening
programs manage to screen over 90 per cent of infants in their catchment area. These
programs are largely hospital-based with initial screening occurring prior to discharge.
Community-based studies also obtain very good coverage when screening is “piggybacked” onto other health or immunisation checks at the health clinic or when it occurs
in the home. Losses to follow-up commonly occur when there is a long delay prior to rescreening or diagnostic testing of the infant, or when infants and mothers are discharged
early from the hospital.
Economic considerations
The economic questions are whether the value to Australian society of implementing a
universal neonatal hearing screening (UNHS) program is likely to be greater than that of
the current situation, and how widespread the screening coverage should be. The existing
situation is varied, and the design of a comprehensive screening system that will cover all
Australian infants remains to be completed.
Information published up until 2003 on the cost-effectiveness of UNHS was limited and
at times contradictory. Furthermore, no Australian UNHS program has yet to be
reported in detail in the literature. As the majority of identified published research
examining the cost-effectiveness of existing or modelled UNHS programs is from an
American perspective, the results obtained can only suggest what might occur under
Australian conditions.
From the available literature it can be concluded that, in the short term, the costs for the
additional cases identified and diagnosed by UNHS are greater per unit than those of
targeted screening. However, taking a societal perspective over the long term suggests
that identifying a larger proportion of hearing-impaired infants at an early stage (ie ≤6
months of age) would result in a cost saving overall. The validity of these estimates of
long-term cost savings should be regarded with caution as they are based primarily on
observational data and expert opinion.
When an experiment is either not ethical or not feasible, or has simply not yet been
carried out, decision analytic modelling can provide insight into the relationship between
the costs and outcomes of the intervention. The base case of the model is simulated with
the available data thought best to approximate the true situation.
In the short term the model presented in this report predicts that implementing a twostage automated auditory brainstem response (AABR) universal neonatal hearing
xii
Universal neonatal hearing screening
screening (UNHS) program for a cohort of 250,000 newborns would identify an extra
607 infants with unilateral or bilateral hearing impairment by the age of 6 months
compared to no formal screening program, at an incremental cost of $6–$11 million.
Where a targeted screening program is already in place, expanding to a universal
screening program would identify 319 more infants, at an incremental cost of $4–$8
million. These figures were reported in 2003 Australian dollars as only the literature
review for UNHS has been updated until 2007.
The long-term direct cost savings from the reduced need for special education and
rehabilitation and the possibility of indirect savings from enhanced productivity in
adulthood outweigh by an order of magnitude the costs of the actual screening and
diagnosis. Unfortunately, these potential savings are less well researched than the shortterm costs and so the estimates are more uncertain. Nevertheless, the long-term savings
from implementing a UNHS program derived from this model are in general agreement
with previous reports.
Given the salience of the estimate of indirect cost savings in the eventual decision
whether or not to implement and to continue to support a national UNHS program, it is
important that more valid and accurate estimates of the indirect costs of hearing and
language skill impairment be obtained for Australian conditions.
The results of the model presented in this report are influenced by the proportion of
infants diagnosed and treated before 6 months of age who actually do achieve normal
language skills. Systematic follow-up of these infants is therefore a high research priority.
The detection and long-term management of permanent childhood hearing impairment
involves public expenditures from both Federal and State/Territory levels of
government, and from both health and non-health departments. Over the long term, the
States/Territories stand to save on special education and rehabilitation, and the Federal
Government to save on disability support pensions.
Recommendation
MSAC recommended that on the strength of evidence pertaining to <application name>
public funding <should/should not> be supported for this procedure.
- The Minister for Health and Ageing endorsed/did not endorse this recommendation on
<date>… OR
Since there is currently insufficient evidence pertaining to <application name>, MSAC
recommended that public funding should not be supported at this time for this
procedure.
- The Minister for Health and Ageing endorsed/did not endorse this recommendation on
<date> -
Universal neonatal hearing screening
xiii
Introduction
The Medical Services Advisory Committee (MSAC) has reviewed the use of universal
neonatal hearing screening. The MSAC evaluates new and existing health technologies
and procedures for which funding is sought under the Medicare Benefits Scheme in
terms of their safety, effectiveness and cost-effectiveness, while taking into account other
issues such as access and equity. The MSAC adopts an evidence-based approach to its
assessments, based on reviews of the scientific literature and other information sources,
including clinical expertise.
The MSAC’s terms of reference and membership are at Appendix A. The MSAC is a
multidisciplinary expert body, comprising members drawn from such disciplines as
diagnostic imaging, pathology, surgery, internal medicine and general practice, clinical
epidemiology, health economics, consumer health and health administration.
A team from Adelaide Health Technology Assessment (AHTA), in the Discipline of
Public Health, School of Population Health and Clinical Practice, University of Adelaide
was engaged to conduct a systematic review of the literature on universal neonatal
hearing screening. An advisory panel with expertise in this area then evaluated the
evidence and provided advice to the MSAC. The advisory panel members are listed at
Appendix B.
This report summarises the assessment of current evidence for universal neonatal hearing
screening.
Rationale for assessment
At the July 2002 Australian Health Ministers’ Conference (AHMC) in Darwin, the health
ministers considered a proposal by the Queensland Health Minister to establish a
universal neonatal hearing screening (UNHS) program for Australia. It was noted that the
type of screening protocol used would have a significant impact on both the direct
screening and ongoing diagnostic costs, and that current Australian data on the cost of
establishing a UNHS program was not available. In order to inform decision-making
regarding this proposal, the AHMC requested the MSAC to commission a full health
technology assessment on the safety, effectiveness and cost-effectiveness of universal
neonatal hearing screening. This would allow an estimation of the potential costs and
cost savings of such a program, and assist with the development of (1) a national
evidence-based screening protocol, (2) guidelines/standards of practice, (3) a national
minimum data set and (4) mechanisms to ensure program sustainability.
Rationale for universal neonatal hearing screening
In general, it is believed that ‘children with hearing loss have delayed development in
vocabulary, grammar, conversation and reading’ (Helfand et al 2001). In an annual
nationwide survey of over 38,000 students with hearing impairment in the USA, 33 to 50
per cent had at least one limitation in thinking/reasoning, maintaining attention or
communication (Gallaudet Research Institute 2001). At present, little can be done in the
way of primary prevention of hearing loss in neonates. However, early identification and
intervention may reduce the disabilities that appear to be associated with permanent
childhood hearing impairment (PCHI), such as language and communication deficit.
14
Universal neonatal hearing screening
The impetus for universal screening has been the belief that targeted ‘risk factor’
screening fails to identify more than 50 to 60 per cent of all neonates who eventually
display some form of permanent hearing loss (Thompson et al 2001). Hearing-impaired
infants may not be identified until months or even years later when undergoing
behavioural response testing (at approximately 8 months of age) or school entry
screening (at approximately 5 years of age). Between 2003 and 2006 the average reported
age of initial fitting of hearing aids in Australian children was 15 months (Australian
Hearing 2007).With universal neonatal hearing screening (UNHS) the fitting of hearing
aids could occur as early as 3 months of age (Australian Hearing 2007).
Delays in the identification of hearing impairment have been suggested to profoundly
affect the quality of life of children in terms of their communication skills and subsequent
education and employment prospects (White 1997; Yoshinaga-Itano et al 2001). A
population-based cohort study of Australian hearing-impaired children, aged 7–8 years
with normal cognition, found substantial delays in their receptive and expressive language
abilities, social-emotional development and reading age, compared to a normative
population (Wake et al 2005). This is despite receiving amplification and ongoing
rehabilitation once diagnosed (Wake et al 2003).
With the development of OAE and AABR methods, screening for hearing impairment
can take place when infants are only hours or days old. This has the potential to impact
on tertiary prevention by improving the times to referral, diagnosis and early
management for hearing-impaired infants. As both of these tests are currently available
for use in Australia for infants designated ‘at risk’ (ie targeted screening), it is suggested
that the benefits and harms of UNHS be systematically evaluated.
Universal neonatal hearing screening
15
Background
Neonatal hearing testing in Australia
The procedure
In Australia there are two screening tools currently being used to identify infants with
possible permanent childhood hearing impairment (PCHI) that may require further
diagnostic assessment. These are the otoacoustic emissions (OAE) test and the
automated auditory brainstem response (AABR) test.
Universal screening is currently implemented broadly throughout the Australian Capital
Territory, New South Wales, Queensland and South Australia (Table 1). At present the
Australian States/Territories differ in their practice of screening neonates for hearing
impairment depending on the choice or availability of technology (a two- or three-stage
protocol of either OAE testing or AABR testing or both).
Table 1
Neonatal hearing screening in Australian States and Territories as at 9th October 2007
State/Territory
Universal neonatal hearing screeninga
State/Territory coverage
Type of screening protocolb
Territory-wide
Tertiary setting
3-stage: AABR
New South Wales
Statewide
Tertiary setting
2-stage: AABR - AABR
Northern Territory
February-March 2008
Tertiary setting
2-stage: AABR - AABR
Queensland
Statewide
Tertiary setting
2-stage: AABR - AABR
South Australia
Statewide
Tertiary setting
1-stage: OAE
Tasmania
Partial
Tertiary setting
2-stage: AABR - AABR
Victoria
Partial
Tertiary setting
1-stage: OAE
Western Australia
Partial
Tertiary setting
2-stage: OAE - AABR
Australian Capital Territory
a All babies are screened, irrespective of risk status; b All infants failing the screening stages are referred for diagnostic assessment; AABR =
automated auditory brainstem response test; OAE = otoacoustic emissions test.
UNHS involves the testing of all newborns, regardless of their risk factor status. This
usually involves testing just prior to discharge from hospital or within a few days of
delivery. Community-based initiatives have only been piloted in one state, South
Australia. In this program initial screening was conducted in a tertiary setting but with
comprehensive community-based follow-up (Child and Youth Health 2001).
Approximately 50–70 per cent of children with permanent congenital hearing loss have a
‘risk factor for deafness’ identified before or after ascertainment of the hearing loss
(Kennedy et al 1998; Mehl & Thomson 2002). These risk factors are listed in Box 1.
Based on the highest level of evidence of UNHS, 8.1 per cent of screened infants were
16
Universal neonatal hearing screening
identified as ‘at risk’ prior to testing (Kennedy et al 1998). Of the children identified with
permanent childhood hearing impairment (PCHI), 41 per cent had been admitted to a
neonatal intensive care unit or special care baby unit, while a further 33 per cent were
well babies who had risk factors for PCHI (eg a family history or a mild craniofacial
malformation) (Kennedy et al 1998). Risk factors for PCHI are listed in Box 1.
Box 1
Established risk factors for targeted neonatal hearing screening
Residence in neonatal intensive care unit / special care baby unit for ≥48 hours
Prolonged usage (>7 days) of aminoglycosides
Family history of permanent childhood deafness
Craniofacial abnormality noticeable at birth
Perinatal infection (either suspected or confirmed), eg toxoplasmosis, rubella, cytomegalovirus,
herpes or acquired meningitis
Birthweight <1.5 kilograms
Birth asphyxia
Chromosomal abnormality, including Down syndrome (Trisomy 21)
Exchange transfusion or intrauterine transfusion, eg hyperbilirubinaemia
Intracranial haemorrhage
What is permanent childhood hearing impairment?
Hearing impairment occurs when there is a reduction in the ability to perceive sound,
resulting from an abnormality anywhere in the auditory system (Columbia University
College of Physicians and Surgeons 2002; Pugh 2000).
Hearing impairment can be categorised as either congenital or acquired. Congenital
hearing impairment is present at birth or arises shortly thereafter as a consequence of
progressive loss, whereas acquired hearing impairment occurs later in the lifespan
(Australian Hearing 2003). Acquired hearing loss can be described as perinatal or
postnatal. Causes of acquired hearing loss include severe hypoxia, neonatal
sepsis/meningitis, viral infections such as mumps, ototoxicity from some medications,
hyperbilirubinaemia, prematurity and trauma (especially secondary to head injury).
Hearing impairment may be unilateral or bilateral. In unilateral hearing impairment one
ear has normal hearing and the other is hearing impaired. Bilateral hearing impairment
indicates that there is hearing loss in both ears.
Hearing impairment can result from disorders of the auricle, external auditory canal,
middle ear, inner ear, auditory nerve, central auditory pathways and auditory cortex
(Figure 1) (Braunwald et al 2001).
Universal neonatal hearing screening
17
Figure 1
Anatomy of the ear
F Jenner. 2003
Hearing losses are usually categorised as conductive, sensorineural or mixed. Conductive
hearing impairment occurs when there is interference with the acoustic transmission of
sound to the cochlea (Department of Medical Oncology 1998). Causes of conductive
hearing impairment in the outer ear include meatal atresia, wax obstruction, infections or
tumours of the external ear canal, tympanic membrane perforation or severe scarring
(tympanosclerosis). Causes of conductive hearing impairment in the middle ear include
middle ear effusions, cholesteatoma, ossicular fixation or discontinuity, or middle ear
tumours. Conductive hearing loss may be transient, and is often reversible once the cause
has been identified and treated (Columbia University College of Physicians and Surgeons
2002). Speech, language and educational outcomes may be affected if the blockage is
chronic, repetitive or not amenable to treatment (Nussbaum 1999).
Sensorineural hearing loss (SNHL) occurs when there is damage to the cochlear hair cells
or auditory nerve. This may occur along the auditory pathway from cochlea to the
brainstem (Columbia University College of Physicians and Surgeons 2002). SNHL may
be genetic or non-genetic and each of these groups may have congenital, acquired or
progressive SNHL. Infection, inflammatory causes, ototoxic medications, trauma, loud
noise exposure and presbyacusis are potential causes of acquired SNHL (Braunwald et al
2001). SNHL can be further categorised as sensory or neural. Sensory losses are caused
by insult to the cochlea by, for example, acoustic trauma or certain viruses such as
mumps. Neural loss occurs with auditory nerve problems such as tumours or neurologic
disorders. Generally, SNHL cannot be reversed and thus constitutes the majority of
permanent childhood hearing impairment (Columbia University College of Physicians
and Surgeons 2002). Early identification and intervention can be considered necessary for
appropriate language development (Nussbaum 1999).
Mixed hearing loss is the result of a problem in both the conductive pathway (outer or
middle ear) and the nerve pathway (inner ear) (Australian Hearing 2003).
18
Universal neonatal hearing screening
Hearing impairment may be defined as slight or mild, moderate, severe or profound. The
grades of hearing impairment differ across organisations and countries. The World
Health Organization has defined hearing loss (in the better ear) in adults:
•
at 26–40 dB as slight or mild hearing impairment. With this hearing loss an
individual should be able to hear and repeat words spoken in a normal voice at a
distance of one metre.
•
at 41–60 dB as moderate impairment. With this impairment an individual can
hear and repeat words spoken in a raised voice at a distance of one metre.
•
at 61–80 dB as severe hearing impairment. At this level an individual is able to
hear some words when shouted into the better ear.
•
at 81 dB or greater as profound hearing impairment, including deafness.
Individuals with this type of impairment are unable to hear and understand even
a shouted voice (Informal Working Group on Prevention of Deafness and
Hearing Impairment Programme Planning 1997).
Universal neonatal hearing screening
19
Testing for hearing impairment in neonates
A flowchart outlining the process of clinical decision-making associated with universal
neonatal hearing screening, and outcomes of interest, is presented in Figure 2.
Figure 2
Clinical pathway for universal neonatal hearing screening
Universal screening
Not universal screening
Initial screen in nursery/community
(OAE and/or AABR)
One-stage
Passa (well)
Testing by indication (risk
factors or behavioural). Includes
targeted and opportunistic
testing using OAE, AABR,
HVDT or school age testing.
Failb
Rescreen (OAE and/or AABR)
within 3 weeks
Two-stage
Passa (well)
Passa
Failb
Diagnostic assessment: diagnostic ABR, OAE,
multi-frequency tympanometry, SSEP
Passa
Failb
Referral/management of PCHI
OUTCOMES: Primary – screening yield, rate and quality of language acquisition, behaviour, family
functioning, communication ability / social functioning, educational achievement,
employment status, socioeconomic status, quality of life
Secondary – age of referral for diagnostic testing, age of PCHI diagnosis, age receiving
therapeutic intervention
OAE = otoacoustic emissions test; AABR = automated auditory brainstem response test; HVDT = health visitor distraction test; ABR = conventional
auditory brainstem response test; SSEP = steady state evoked potentials test – often not readily available and may not always be necessary; PCHI =
permanent childhood hearing impairment. a Some children who pass but have risk factors may be brought in for diagnostic testing at a later stage; b fail in
one or both ears.
20
Universal neonatal hearing screening
The two tools used for universal neonatal hearing screening are the otoacoustic emissions
(OAE) test and the automated auditory brainstem response (AABR) test. These tests are
used either singly, followed by diagnostic testing; or as part of a two- or even three-stage
screening protocol that includes repeat testing on the same instrument or repeat testing
using the other screening tool, or both.
Otoacoustic emissions testing
The cochlea produces what are now termed otoacoustic emissions (OAEs) in response to
a sound stimulus. When sound enters the ear in the form of a pressure wave, the ossicles
of the middle ear turn this wave into a mechanical force that causes movement of
cochlear fluid, resulting in the generation of a wave within the basal membrane (Kemp
2002). This wave travels towards the apex of the cochlea, peaking and then stopping at a
specific region of the cochlea unique to the input frequency of the sound. The wave
stimulates hair cells located in the organ of Corti, which consists of one row of inner and
three rows of outer hair cells. The inner hair cells are necessary for signal transduction to
the hearing centres of the brain (Kemp 2003), and the outer hair cells are responsible for
sound quality and for amplifying the energy of the wave travelling along the basement
membrane. This amplification is necessary as the initial wave rapidly loses energy due to
the viscosity of the cochlear fluids and the energy required to stimulate the inner hair
cells. While most of the wave energy proceeds in a forward direction toward the apex of
the cochlea, some is lost in a backward direction toward the tympanum (eardrum). This
backward energy causes the tympanum to vibrate, producing OAEs. Outer hair cells,
rather than inner hair cells, are responsible for the production of OAEs. From sound
input to otoacoustic emission, the whole process takes anywhere from 3 to 15
milliseconds.
Transient evoked and distortion product otoacoustic emissions testing
Otoacoustic emissions can occur spontaneously in the absence of any external stimuli.
They can also be elicited by applying either a single click or tone (via a transient evoked
otoacoustic emissions test – TEOAE) or two simultaneous tones (via a distortion
product otoacoustic emissions test – DPOAE) to the ear and recording the response
from the cochlea (Probst et al 1991).
The term ‘transient’ refers to the presence of a latency period between the intermittent
click stimuli and the returning OAEs. Wide band stimuli such as clicks can excite and
elicit responses from a wide region of the cochlea. Higher frequency vibrations stimulate
areas of the cochlea at its base, while lower frequency vibrations travel farther down
towards the apex (Robles & Ruggero 2001). Therefore, OAEs produced by different click
frequencies reach the tympanum at different times, depending upon their origin in the
cochlea (ie from nearer the base or the apex) (Kemp 2002). By measuring OAE
responses at different frequencies, the functionality of different regions of the cochlea
can be assessed.
DPOAE responses are produced by introducing two simultaneous tones at different
frequencies (F1 and F2) that, in stimulating their respective sites on the basal membrane,
generate a distortion product wave of a different frequency that stimulates another set of
outer hair cells (Kemp 2002). The resulting distortion product is located somewhere
between F1 and F2 and is defined by the function 2F-F2.
These two methods have been described as being complementary (Kemp 2002). TEOAE
testing is described as sensitive, having good frequency resolution and efficient – it can
Universal neonatal hearing screening
21
cover a wide range of frequencies in one measurement. However, most ears do not
produce TEOAEs over 4 kHz, so background noise must be minimal when conducting
the test. Distortion product technology is able to elicit OAEs in normal ears at
frequencies over 10 kHz.
Auditory brainstem response testing (reference standard)
Conventional auditory brainstem response (ABR) audiometry consists of a neurological
test of auditory brainstem function in response to click or tone stimuli (Scott &
Bhattacharyya 2002). These audible stimuli result in the production of action potentials,
generated by neurons along the auditory pathway, which are detected by surface
electrodes placed on the scalp and earlobe of the individual. Repeated stimulus allows for
the collection of multiple responses that are then averaged in an attempt to distinguish
real responses from noise (real responses will provide a consistent signal and are more
likely to produce a higher average response, while noise is non-synchronous). This
technique allows properly trained professionals to identify the presence or absence of
auditory brainstem responses from the surrounding noise and determine the hearing level
threshold. The advantage of ABR testing is that it is sensitive to both cochlear and retrocochlear pathology. Currently, conventional ABR testing is the gold-standard for the
diagnosis of hearing impairment in infants.
Automated auditory brainstem response testing
With the introduction in many countries of neonatal hearing screening programs, there
developed a need for more rapid testing using less technically trained staff. The
automated auditory brainstem response (AABR) system consists of a small handheld
device that presents fully automated ‘pass/refer’ screening results that require no
interpretation by the user (see figure below). The stimulus is provided through an ear-cup
fitted over the infant’s ear, and responses are measured through probes attached to the
scalp, earlobe, neck and shoulder.
In the automated form of the test a detection algorithm assumes that the infant is hearing
impaired and only provides a ‘pass’ when enough responses have been detected to
assume with very high probability that the infant has ‘normal’ hearing (a statistical
likelihood ratio model). This assumption is based on cross-matching the response with a
template from an infant with confirmed normal hearing. Hearing impairment is measured
by the level of loss as the ‘decibels (dB) hearing level (HL)’. The AABR test is ideally
performed with a click at 35 dB HL that sweeps across a range of frequencies from 500
Hz to 5 kHz. This allows for hearing loss specifically at the intensity level and frequency
range of normal speech (≥30 dB HL, 500 Hz to 4 kHz). The automated test is, however,
by no means diagnostic, and it is critical to ensure that the equipment is independently
calibrated to accurately capture hearing losses greater than 35 dB.
22
Universal neonatal hearing screening
Used with permission, Natus Algo.
Marketing status of the screening tests
The two major categories of tests used for screening for neonatal hearing impairment
(otoacoustic emissions testing and automated auditory brainstem response testing) are
included on the Australian Register of Therapeutic Goods (ARTG) (see Table 2).
Table 2
Neonatal hearing screening equipment listed on the ARTG
Product name
ARTG #
Product #
Sponsor
Oto-acoustic emission instrument
129297
213915
Medtel Pty Ltd
Oto-acoustic emission instrument
145411
233456
Central Neurophysiology Supplies Pty Ltd
Oto-acoustic emission instrument
94690
165324
GN Resound Pty Ltd
Natus ALGO Portable
141199
228603
Scanmedics Pty Ltd
Natus ALGO 3
97470
168390
Scanmedics Pty Ltd
Current reimbursement arrangement
Otoacoustic emissions (OAE) testing (MBS Item 11332) and conventional auditory
brainstem response testing (ie for diagnostic assessment, MBS Item 11300) are currently
listed on the Medicare Benefits Schedule (MBS). Medicare benefits are, however, not
payable for OAE audiometry that is undertaken for the purpose of routine screening of
infants. According to MBS requirements, the equipment used must be able to display a
recorded emission rather than a pass/fail indicator. Automated auditory brainstem
response testing is not listed on the Schedule.
Universal neonatal hearing screening
23
Approach to assessment
Objective
The objective of this MSAC assessment was to determine whether there is sufficient
evidence to establish a program of universal neonatal hearing screening (UNHS) in
Australia.
Research questions
The collated literature was assessed as to its suitability to answer specific, relevant
research questions:
1. What is the prevalence of permanent hearing impairment in neonates and infants in
Australia?
2. What is the diagnostic accuracy of the tests for permanent childhood hearing
impairment when conducted on the neonate or infant?
3. Does universal neonatal hearing screening, and the finding of a positive and/or
negative test, affect the clinical management or treatment options available to
permanently hearing-impaired infants?
4. Does universal neonatal hearing screening, and therefore possible alterations in
clinical management, have an impact on the adverse outcomes associated with
permanent childhood hearing impairment?
Assessment of screening programs
Screening is ‘a public health service in which members of a defined population, who do
not necessarily perceive they are at risk of, or are already affected by, a disease or its
complications, are asked a question or offered a test to identify those individuals who are
more likely to be helped than harmed by further tests or treatment to reduce the risk of
disease or its complications’ (UK National Screening Committee 2000).
Screening programs can be:
•
systematic/universal – mass screening of an entire segment of the population;
•
selective – targeting high risk groups (see Box 1) in the population. This includes
cascade screening or case finding; and
•
opportunistic – screening as a result of a patient’s consultation with a clinician.
(Davies et al 2000; Muir 2001; Murray et al 1999; Schersten et al 1999)
Criteria for appraising the viability, effectiveness and appropriateness of a screening
program have been developed by the UK National Screening Committee (UK National
Screening Committee 2000) and are presented in Box 2.
24
Universal neonatal hearing screening
It was the aim of the evaluation team to summarise the analysis of universal neonatal
hearing screening in terms of these criteria. The evaluation team recognised, however,
that information or evidence addressing all of these criteria was unlikely to be available.
Universal neonatal hearing screening
25
Box 2
Criteria for appraising the viability, effectiveness and appropriateness of a screening
program (UK National Screening Committee 2000)
The condition
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.1 The condition should be an important health problem.
1.2 The epidemiology and natural history of the condition, including
development from latent to declared disease, should be adequately
understood and there should be a detectable risk factor, or disease
marker and a latent period or early symptomatic stage.
1.3 All the cost-effective primary prevention interventions should have been
implemented as far as practicable.
The test
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.4 There should be a simple, safe, precise and validated screening test.
1.5 The distribution of test values in the target population should be known
and a suitable cut-off level defined and agreed.
1.6 The test should be acceptable to the population.
1.7 There should be an agreed policy on the further diagnostic
investigation of individuals with a positive test result and on the choices
available to those individuals.
The treatment
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.8 There should be an effective treatment or intervention for patients
identified through early detection, with evidence of early treatment
leading to better outcomes than late treatment.
1.9 There should be agreed evidence based policies covering which
individuals should be offered treatment and the appropriate treatment
to be offered.
1.10 Clinical management of the condition and patient outcomes should be
optimised by all health care providers prior to participation in a
screening program.
The screening program
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.11 There must be evidence from high quality Randomised Controlled Trials
that the screening program is effective in reducing mortality or
morbidity.
1.12 Where screening is aimed solely at providing information to allow the
person being screened to make an “informed choice” (e.g. Down
syndrome, cystic fibrosis carrier screening), there must be evidence
from high quality trials that the test accurately measures risk. The
information that is provided about the test and its outcome must be of
value and readily understood by the individual being screened.
1.13 There should be evidence that the complete screening program (test,
diagnostic procedures, treatment/intervention) is clinically, socially and
ethically acceptable to health professionals and the public.
1.14 The benefit from the screening program should outweigh the physical
and psychological harm (caused by the test, diagnostic procedures and
treatment).
26
Universal neonatal hearing screening
Box 2 (cont.) Criteria for appraising the viability, effectiveness and appropriateness of a screening
program (UK National Screening Committee 2000)
The screening program (cont.)
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.15 The opportunity cost of the screening program (including testing,
diagnosis, treatment, administration, training and quality assurance)
should be economically balanced in relation to expenditure on medical
care as a whole (i.e. value for money).
1.15 There must be a plan for managing and monitoring the screening
program and an agreed set of quality assurance standards.
1.16 Adequate staffing and facilities for testing, diagnosis, treatment and
program management should be made available prior to the
commencement of the screening program.
1.17 All other options for managing the condition should have been
considered (e.g. improving treatment, providing other services), to
ensure that no more cost effective intervention could be introduced or
current interventions increased within the resources available.
1.18 Evidence-based information, explaining the consequence of testing,
investigation and treatment, should be made available to potential
participants to assist them in making an informed choice.
1.19 Public pressure for widening the eligibility criteria, for reducing the
screening interval, and for increasing the sensitivity of the testing
process, should be anticipated. Decisions about these parameters
should be scientifically justifiable to the public.
This assessment of universal neonatal hearing screening is based on the framework
outlined in the MSAC “Guidelines for the assessment of diagnostic technologies” handbook(MSAC
2005).
Assessing the safety and effectiveness of universal neonatal hearing screening (UNHS)
was approached using two complementary techniques:
1. assessing studies of screening programs as a whole (direct evidence of the impact of
screening on patient relevant outcomes)
2. assessing the diagnostic test performance (diagnostic accuracy) – sensitivity,
specificity and accuracy.
Diagnostic accuracy was assessed as it was expected that only limited data would be
available from the trials or studies of UNHS programs.
The cost-effectiveness of UNHS was determined on the basis of previously published
reports, and was also calculated independently using Australian cost estimates and
effectiveness estimates derived from this systematic review of the literature. Information
on the methods and results of the economic analysis are presented elsewhere in this
assessment report.
Universal neonatal hearing screening
27
Review of literature
The medical literature was searched to identify relevant studies concerning universal
neonatal hearing screening (UNHS) for the period between 1966 and August 2007.
Appendix C describes the electronic databases that were used for this search and the
other sources of evidence – particularly grey literature – that were investigated.
The search terms, presented in Appendix C, were used to identify literature in electronic
bibliographic databases on the prevalence of permanent childhood hearing impairment
(PCHI); the diagnostic accuracy of the screening tests; the safety, effectiveness and costeffectiveness of UNHS.
Inclusion/Exclusion criteria
The criteria for including articles varied depending on the type of research question being
addressed. Often a study was assessed more than once because it addressed more than
one research question. Two researchers separately applied the inclusion criteria to the
collated literature to ensure that all potentially relevant studies were captured. In general,
articles were excluded if they did not:
•
address the research question
•
provide information on the pre-specified target population
•
include one of the pre-specified interventions
•
compare results to a pre-specified comparator
•
address one of the pre-specified outcomes and/or provided inadequate numerator
and/or denominator data
•
have the appropriate study design.
The inclusion criteria relevant to each of the research questions posed in this assessment
are provided in Box 3, Box 4, Box 5, Box 6 in the results section of this report.
Grey literature 1 was included in the search strategy. Unpublished literature, however, was
not canvassed as it is difficult to search for this literature exhaustively and systematically,
and trials that are difficult to locate are often smaller and of lower methodological quality
(Egger et al 2003). It is, however, possible that these unpublished data (particularly from
local screening programs) could have an impact on the results of this review.
In terms of the ‘pre-specified target population’, the definition of a neonate is a live birth
who is less than 28 days old (Australian Institute of Health and Welfare 2002). As this is a
review of neonatal screening, it would be expected that the target population would only
fit into this age category. Neonates and infants up to 6 months of age were, however,
included in this assessment of hearing screening for several reasons:
1
28
Literature that is difficult to find including published government reports, theses, technical reports, nonpeer reviewed literature etc.
Universal neonatal hearing screening
•
to include the critical period for language acquisition in children. Given that the first
babbling stage is around 4–6 months of age , it makes sense that for a screening
program to have its maximum effect, recognition of PCHI and initial management
should occur prior to this stage;
•
to ensure that the age period includes both the chronological and corrected (due to
pre-term birth) age of the target population;
•
to ensure that programs concerning infants not born in hospitals and not immediately
screened due to access issues (ie in rural and remote areas) can still be assessed; and
•
to exclude the effect of acquired or transient hearing loss (through infection) on the
results. Older children are more exposed to situations where they can acquire
infections or experience traumas.
Search results
The process of study selection for this report went through six phases:
1. All reference citations from all literature sources were collated into an Endnote
8.0 database;
2. Duplicate references were removed;
3. Studies were excluded, on the basis of the citation information, if it was obvious
that they did not meet the pre-specified inclusion criteria. Citations were assessed
independently by two reviewers. Studies marked as requiring further evaluation
by either reviewer were retrieved for full-text assessment (after discussion);
4. Studies were included to address the research questions if they met the prespecified criteria again independently applied by two reviewers to the full-text
articles. Those articles meeting the criteria formed part of the evidence-base. The
remainder provided background information;
5. The reference lists of the included articles were pearled for additional relevant
studies. These were retrieved and assessed according to phase 4; and
6. The evidence-base consisted of articles from phases 4 and 5 that met the
inclusion criteria.
Any doubt concerning inclusions at phase 4 was resolved by consensus between the two
reviewers. A third reviewer was included to arbitrate where necessary. The results of the
process of study selection are provided in Figure 3.
Universal neonatal hearing screening
29
Figure 3
Summary of the process used to identify and select studies for the assessment of universal
neonatal hearing screening
Potentially relevant studies identified
in the literature searches and
screened for retrieval:
prevalence (n=3810)
accuracy in diagnosis (n=4501)
screening (n=2703)
cost effectiveness (n=2000)
Studies retrieved for more
detailed evaluation:
prevalence (n=148)
accuracy in diagnosis (n=168)
screening (n=276)
cost effectiveness (n=145)
Potentially appropriate studies to be
included in the systematic review:
prevalence (n=38)
accuracy in diagnosis (n=11)
screening (n=105)
cost effectiveness (n=12)
Studies excluded because did not meet the inclusion
criteria:
prevalence (n=3662)
accuracy in diagnosis (n=4333)
screening (n=2427)
cost effectiveness (n=1855)
Studies excluded because did not meet inclusion
criteria:
prevalence (n=110)
accuracy in diagnosis (n=157)
screening (n=171)
cost effectiveness (n=133)
Studies excluded, with reasons:
Unable to extract data (n= 34)
Data included in another paper (n= 15)
Data not available (n=10)
Studies included in the systematic
review (n=103) and by outcome:
safety (n=10)
prevalence (n=19)
accuracy in diagnosis (n=5)
screening (n=64)
cost effectiveness (n=12)
Data extraction and analysis
A study profile was developed for each included study (see Appendix F) – outlining the
level of evidence, study quality, authors, publication year, location, study design, study
population characteristics, type of intervention, testing or screening protocol, comparator
or reference standard, and outcomes assessed. Studies that were unable to be retrieved, or
that met the inclusion criteria but contained insufficient or inadequate data, are provided
in Appendix G. Definitions of all technical terms and abbreviations are provided in the
appended Glossary and Abbreviations.
Descriptive statistics were extracted or calculated for all safety and effectiveness
outcomes (defined in the assessment protocol) in the individual studies, including
30
Universal neonatal hearing screening
numerator and denominator information, means and standard deviations. The power of
individual controlled studies to detect a clinically important effect was calculated,
assuming that α = 0.05.
Relative risk/rate ratios (RR), absolute risk differences, number needed to screen or
diagnose to benefit or harm, and associated 95 per cent confidence intervals, were
calculated from individual comparative studies containing count data. The calculated
number needed to diagnose was always rounded up to the next whole value, as the value
represents the number of infants required for one extra case to be diagnosed, and this
cannot be reported as a fraction. By rounding up the value, it will only overestimate the
number of infants who need to undergo screening to diagnose one case, not
underestimate. Mean differences and 95 per cent confidence intervals were calculated for
normally distributed continuous outcomes in individual studies using the independent ttest. Where authors had reported statistical analyses, particularly adjusted analyses (eg
analysis of covariance, regression), their results were reported as the primary data were
not available to replicate or confirm the result.
Assessing diagnostic accuracy
In the analysis of diagnostic accuracy, calculations of sensitivity, specificity and positive
predictive values of tests, with 95 per cent confidence intervals, were undertaken where
possible. False positive rates (the complement of test specificity) and false alarm rates
(the complement of the positive predictive value of a test) were also calculated in the
analyses of diagnostic accuracy and screening effectiveness, respectively. As an analysis of
screening effectiveness, as opposed to efficacy (Walter 2003), was undertaken, all data
were presented according to intention-to-screen principles. That is, calculations were
undertaken using as the denominator those infants that were intended for screening, as
opposed to those who were actually screened. For example, if 100 infants were referred
for re-screening (eg 20 were lost to follow-up, 40 failed the screen and 40 passed the
screen), the failure rate would be calculated as 40/100, rather than 40/80. Data on loss to
follow-up, or compliance, were assessed separately.
Meta-analysis was not undertaken as the evidence-base was heterogeneous and there were
very few controlled trials of screening. A narrative synthesis of the data was therefore
undertaken.
All statistical calculations and testing were undertaken using the biostatistical computer
package, Stata version 7.0 (Stata Corporation 2001).
Appraisal of the evidence
The evidence presented in the selected studies was assessed and classified using the
dimensions of evidence defined by the National Health and Medical Research Council
(NHMRC 2000a).
These dimensions (Table 3) consider important aspects of the evidence supporting a
particular intervention and include three main domains: strength of the evidence, size of
the effect and relevance of the evidence. The first domain is derived directly from the
literature identified as informing a particular intervention. The last two require expert
clinical input as part of its determination.
Universal neonatal hearing screening
31
Table 3
Evidence dimensions
Type of evidence
Strength of the evidence
Level
Quality
Statistical precision
Definition
The study design used, as an indicator of the degree to which bias has been eliminated by
design.*
The methods used by investigators to minimise bias within a study design.
The p-value or, alternatively, the precision of the estimate of the effect. It reflects the
degree of certainty about the existence of a true effect.
Size of effect
The distance of the study estimate from the “null” value and the inclusion of only clinically
important effects in the confidence interval.
Relevance of evidence
The usefulness of the evidence in clinical practice, particularly the appropriateness of the
outcome measures used.
*See Table 5
The three sub-domains (level, quality and statistical precision) are collectively a measure
of the strength of the evidence. With respect, specifically, to diagnostic evidence the
individual studies assessing diagnostic effectiveness were graded according to the prespecified quality and applicability criteria (MSAC 2005) as shown in Appendix E. The
designations of the levels of evidence are shown in Table 5. Study quality was assessed
using the critical appraisal checklists provided in * A systematic review will only be
assigned a level of evidence as high as the studies it contains, excepting where those
studies are of level II evidence.
** The dimensions of evidence apply only to studies of diagnostic accuracy. To assess the effectiveness of a diagnostic test there also needs
to be a consideration of the impact of the test on patient management and health outcomes. See MSAC (2004) Guidelines for the assessment
of diagnostic technologies. Available at: 52Hwww.msac.gov.au .
§§ The validity of the reference standard should be determined in the context of the disease under review. Criteria for determining the validity
of the reference standard should be pre-specified. This can include the choice of the reference standard(s) and its timing in relation to the index
test. The validity of the reference standard can be determined through quality appraisal of the study. See Whiting P, Rutjes AWS, Reitsma JB,
Bossuyt PMM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in
systematic reviews. BMC Medical Research Methodology 2003, 3: 25.
†† Well-designed population based case-control studies (eg. population based screening studies where test accuracy is assessed on all cases,
with a random sample of controls) do capture a population with a representative spectrum of disease and thus fulfil the requirements for a valid
assembly of patients. However, in some cases the population assembled is not representative of the use of the test in practice. In diagnostic
case-control studies a selected sample of patients already known to have the disease are compared with a separate group of normal/healthy
people known to be free of the disease. In this situation patients with borderline or mild expressions of the disease, and conditions mimicking
the disease are excluded, which can lead to exaggeration of both sensitivity and specificity. This is called spectrum bias because the spectrum
of study participants will not be representative of patients seen in practice.
‡‡ Studies of diagnostic yield provide the yield of diagnosed patients, as determined by an index test, without confirmation of the accuracy of
this diagnosis by a reference standard. These may be the only alternative when there is no reliable reference standard.
Note 1: Assessment of comparative harms/safety should occur according to the hierarchy presented for each of the research questions, with
the proviso that this assessment occurs within the context of the topic being assessed. Some harms are rare and cannot feasibly be captured
within randomised controlled trials; physical harms and psychological harms may need to be addressed by different study designs; harms from
diagnostic testing include the likelihood of false positive and false negative results; harms from screening include the likelihood of false alarm
and false reassurance results.
Note 2: When a level of evidence is attributed in the text of a document, it should also be framed according to its corresponding research
question eg. level II intervention evidence; level IV diagnostic evidence.
.
32
Universal neonatal hearing screening
Table 4
Grading system used to rank included diagnostic studies
Validity criteria
Description
Grading System
Appropriate
Did the study evaluate a direct comparison of the
C1 direct comparison
comparison
index test strategy versus the comparator test
CX other comparison
strategy?
Applicable population
Did the study evaluate the index test in a population
P1 applicable
that is representative of the subject characteristics
P2 limited
(age and sex) and clinical setting (disease
P3 different population
prevalence, disease severity, referral filter and
sequence of tests) for the clinical indication of
interest?
Quality of study
Was the study designed to avoid bias?
Study design: NHMRC level of evidence
High quality = no potential for bias based on predefined key quality criteria
Study quality (QUADAS checklist):
Q1 high quality (≥12/14)
Medium quality = some potential for bias in areas
other than those pre-specified as key criteria
Poor quality = poor reference standard and/or
potential for bias based on key pre-specified criteria
Table 5
Q2 medium (10-11/14)
Q3 poor reference standard
poor quality (<10/14)
or insufficient information
Designations of levels of evidence* according to type of research question (NHMRC 2005)
Level
Diagnostic accuracy **
Screening
I
A systematic review of level II studies
A systematic review of level II studies
II
A study of test accuracy with: an independent,
blinded comparison with a valid reference standard,
§§ among consecutive patients with a defined clinical
presentation ††
A randomised controlled trial
III-1
A study of test accuracy with: an independent,
blinded comparison with a valid reference standard,
§§ among non-consecutive patients with a defined
clinical presentation††
A pseudorandomised controlled trial
(i.e. alternate allocation or some other method)
III-2
A comparison with reference standard that does not
meet the criteria required for Level II and III-1
evidence
A comparative study with concurrent controls:
Non-randomised, experimental trial
Cohort study
Case-control study
III-3
Diagnostic case-control study ††
A comparative study without concurrent controls:
Historical control study
Two or more single arm study
IV
Study of diagnostic yield (no reference standard) ‡‡
Case series
*
* A systematic review will only be assigned a level of evidence as high as the studies it contains, excepting where those studies are of level II
evidence.
** The dimensions of evidence apply only to studies of diagnostic accuracy. To assess the effectiveness of a diagnostic test there also needs
to be a consideration of the impact of the test on patient management and health outcomes. See MSAC (2004) Guidelines for the assessment
of diagnostic technologies. Available at: www.msac.gov.au .
§§ The validity of the reference standard should be determined in the context of the disease under review. Criteria for determining the validity
of the reference standard should be pre-specified. This can include the choice of the reference standard(s) and its timing in relation to the index
test. The validity of the reference standard can be determined through quality appraisal of the study. See Whiting P, Rutjes AWS, Reitsma JB,
Bossuyt PMM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in
systematic reviews. BMC Medical Research Methodology 2003, 3: 25.
Universal neonatal hearing screening
33
†† Well-designed population based case-control studies (eg. population based screening studies where test accuracy is assessed on all cases,
with a random sample of controls) do capture a population with a representative spectrum of disease and thus fulfil the requirements for a valid
assembly of patients. However, in some cases the population assembled is not representative of the use of the test in practice. In diagnostic
case-control studies a selected sample of patients already known to have the disease are compared with a separate group of normal/healthy
people known to be free of the disease. In this situation patients with borderline or mild expressions of the disease, and conditions mimicking
the disease are excluded, which can lead to exaggeration of both sensitivity and specificity. This is called spectrum bias because the spectrum
of study participants will not be representative of patients seen in practice.
‡‡ Studies of diagnostic yield provide the yield of diagnosed patients, as determined by an index test, without confirmation of the accuracy of
this diagnosis by a reference standard. These may be the only alternative when there is no reliable reference standard.
Note 1: Assessment of comparative harms/safety should occur according to the hierarchy presented for each of the research questions, with
the proviso that this assessment occurs within the context of the topic being assessed. Some harms are rare and cannot feasibly be captured
within randomised controlled trials; physical harms and psychological harms may need to be addressed by different study designs; harms from
diagnostic testing include the likelihood of false positive and false negative results; harms from screening include the likelihood of false alarm
and false reassurance results.
Note 2: When a level of evidence is attributed in the text of a document, it should also be framed according to its corresponding research
question eg. level II intervention evidence; level IV diagnostic evidence.
Table 6
Quality checklists
Study type
Checklist/s
Systematic reviews/HTA reports
NHMRC Checklist Table 1.4 (NHMRC 2000b)
Randomised controlled trials
NHMRC Checklist Box 6.1 (NHMRC 2000a)
Cohort study
NHMRC Checklist Table 1.4 (NHMRC 2000b)
Case-control
NHMRC Checklist Table 1.4 (NHMRC 2000b)
Diagnostic test cross sectional
study
QUADAS quality assessment tool (Whiting P 2003)
Intervention case series
NHS CRD Quality Assessment Scale (Box 5.9) (Khan et al 2001)
Statistical precision
Statistical precision was determined using standard statistical principles. Small confidence
intervals and p-values give an indication as to the probability that the reported effect is
real (NHMRC 2000b).
Size of effect in individual studies
It is important to establish whether statistically significant differences are also clinically
important. The size of the effect needs to be determined, as well as whether the 95
per cent confidence interval includes only clinically important effects (NHMRC 2000b).
Relevance of evidence in individual studies
Similarly, the outcome being measured in the studies should be appropriate and clinically
relevant. Inadequately validated (predictive) surrogate measures of a clinically relevant
outcome should be avoided (NHMRC 2000b).
Assessment of the body of evidence
Once the results of the studies were synthesised, the overall conclusion as derived from
the body of evidence (Table 7) was presented to answer each clinical question – see
Discussion section.
34
Universal neonatal hearing screening
Table 7
Body of evidence assessment matrix
Component
A
B
C
D
Excellent
Good
Satisfactory
Poor
several level I or II
studies with low risk of
bias
one or two level II
studies with low
risk of bias or a
SR/multiple level
III studies with low
risk of bias
level III studies with
low risk of bias, or
level I or II studies
with moderate risk
of bias
level IV studies, or
level I to III studies
with high risk of
bias
all studies consistent
most studies
consistent and
inconsistency
may be explained
some
inconsistency
reflecting genuine
uncertainty around
clinical question
evidence is
inconsistent
Clinical impact
very large
substantial
moderate
slight or restricted
Generalisability
population/s studied in
body of evidence are
the same as the target
population
population/s
studied in the
body of evidence
are similar to the
target population
population/s
studied in body of
evidence different
to target population
but it is clinically
sensible to apply
this evidence to
target population
population/s
studied in body of
evidence different
to target population
and hard to judge
whether it is
sensible to
generalise to target
population
directly applicable to
Australian healthcare
context
applicable to
Australian
healthcare
context with few
caveats
probably applicable
to Australian
healthcare context
with some caveats
not applicable to
Australian
healthcare context
Evidence base
Consistency
Applicability
Expert advice
An advisory panel with expertise in paediatrics, otorhinolaryngology, audiology, deaf
education, epidemiology, consumer issues and neonatal hearing screening was established
to evaluate the evidence and provide advice to the MSAC from a clinical and client
perspective. In selecting members for advisory panels, the MSAC’s practice is to
approach the appropriate medical colleges, specialist societies and associations and
consumer bodies for nominees. Membership of the advisory panel is provided at
Appendix B.
Universal neonatal hearing screening
35
Results of assessment
How prevalent is permanent childhood hearing impairment?
Nineteen studies met the inclusion criteria delineated in the assessment protocol (Box 3)
and provided estimates of the prevalence of permanent childhood hearing impairment
(PCHI). Profiles of these studies, including the raw prevalence data, are provided in
Appendix F. Eight studies were conducted in the UK, three in the USA, two in Australia,
two in Italy, and one each in France, Cyprus, Italy and Austria.
Box 3
Study selection criteria for prevalence
Research question
What is the prevalence of permanent hearing impairment in neonates and infants in Australia?
Selection criteria
Population
Outcome
Study design
Search period
Language
a
Inclusion criteria
Neonates and infants ≤6 months of age born in (1) Australia or, if this information was
unavailable, in (2) Western countries of similar demographic composition. If information on the
prevalence in this age group was unavailable, the criterion was widened to ≤6 years of age.
Prevalence – proportion of infants with permanent childhood hearing impairment (defined as ≥35
dB).a The source of prevalence estimates was clearly defined – neonatal screening programs,
school-entry screening etc.
Cross-sectional surveys (with random sampling), case series of consecutive children, or cohort
studies.
In order to obtain relatively recent prevalence estimates, studies published before 1980 were not
included.
Studies relevant to Australia’s demographic composition are most likely to be published in
English. Therefore, studies in languages other than English were not included.
interested in a total population prevalence or prevalence estimate, rather than the prevalence in ‘high-risk’ groups alone
There were three sources from which prevalence data were ascertained: neonatal hearing
screening of well babies only, universal neonatal hearing screening of well and ‘at-risk’
babies and examination of school and/or health records or databases. Table 8
summarises the available prevalence data stratified by level of hearing impairment and
source of data ascertainment.
Three studies assessed the prevalence of PCHI greater than 35 dB HL, as ascertained
through universal neonatal screening of well and ‘at-risk’ babies. One study conducted in
the USA found that sensorineural bilateral hearing loss occurred in 1.8 per 1,000
neonates born between December 1996 and December 1997. Unilateral hearing loss
occurred in 0.9 per 1,000 neonates (Stewart et al 2000). A further study conducted in the
USA reported a prevalence of sensorineural or conductive hearing loss of 4.4 per 1,000
neonates born between January 1997 and January 2002 (Connolly et al 2005). The higher
prevalence estimate reported in this study is likely a result of the relatively high
proportion of ‘at-risk’ neonates screened (18.1 per cent of the total sample). An
Australian study found lower rates of PCHI in infants born in Western Australia between
February 2000 and June 2001. Bilateral hearing loss occurred in 0.7 per 1,000 neonates
and unilateral hearing loss in 0.2 per 1,000 neonates (Bailey et al 2002).
36
Universal neonatal hearing screening
Table 8
PCHI level
Prevalence of permanent childhood hearing impairment (PCHI)
Country
Source of data
Prevalence (per 1,000)a
Australia
UNHS – well and at-risk babies
Bilateral >35 dB: 0.7/1,000
Unilateral >35 dB: 0.2/1,000
>35 dB HL
USA
UNHS – well and at-risk babies
Sensorineural/conductive: 4.43/1,000b
USA
UNHS – well and at-risk babies
Sensorineural: 2.7/1,000
Bilateral: 1.8/1,000
Unilateral: 0.9/1,000
France
UNHS – well babies only
Sensorineural (bilateral): 1.4/1,000
UK
Quasi-randomised trial. Children
born during periods with/without
UNHS - well and at-risk babies.
Health records also reviewed
Group with UNHS:
Bilateral >40 dB: 1.2/1,000
Group without UNHS:
Bilateral >40 dB: 1.2/1,000
Overall:
Bilateral >40 dB: 1.2/1,000
>40 dB HL
Germany
UNHS – well and at-risk babies
Congenital >40 dB: 2.1/1,000b
Italy
UNHS – well and at-risk babies
Congenital >40 dB: 3.2/1,000
Bilateral: 1.7/1,000
Unilateral: 1.5/1,000
UK
UNHS – well and at-risk babies
Bilateral >40 dB: 1.3/1,000
41–80 dB: 1.0/1,000
>80 dB: 0.3/1,000
Unilateral >40 dB: 0.4/1,000
41–80 dB: 0.2/1,000
>80 dB: 0.1/1,000
UK
UNHS – well and at-risk babies
Bilateral >40 dB: 1.0/1,000
USA
UNHS – well and at-risk babies
Sensorineural/conductive: 1.5/1,000b
41–70 dB: 0.8/1,000
>70 dB: 0.7/1,000
Australia
Other methods – health and/or
education records
Bilateral: >40 dB: 1.1/1,000
>60 dB: 0.5/1,000
>90 dB: 0.2/1,000
Austria
Other methods – health and/or
education records
Congenital >40 dB: 1.3/1,000b
41–69 dB: 0.7/1,000
70–94 dB: 0.3/1,000
>94 dB: 0.3/1,000
UK
Other methods – health and/or
education records
Congenital: 0.8/1,000b
UK
Other methods – health and/or
education records
Bilateral >40 dB: 1.2/1,000
UK
Other methods – health and/or
education records
Congenital (bilateral): 1.8/1,000
UK
Other methods – source of records
not stated
Unilateral >40 dB: 1.2/1,000
Universal neonatal hearing screening
37
Italy
UNHS – well and at-risk babies
Sensorineural >56 dB: 1.9/1,000b
UK
Other methods – health and/or
education records
Sensorineural/mixed and congenital/
progressive: >50 dB: 1.1/1,000b
51–64 dB: 0.3/1,000
65–79 dB: 0.3/1,000
80–94 dB: 0.2/1,000
>94 dB: 0.3/1,000
Cyprus
Other methods – health and/or
education records
Congenital (bilateral): >50 dB: 1.2/1,000
51–69 dB: 0.3/1,000
70–94 dB: 0.4/1,000
>94 dB: 0.5/1,000
>50 dB HL
UNHS = universal neonatal hearing screening. a Rounded estimates – raw data are provided in Appendix F. b Author did not distinguish
between unilateral and bilateral hearing loss.
Thirteen studies reported on the prevalence of PCHI greater than 40 dB HL. This was
made up of seven studies assessing universal neonatal hearing screening (UNHS)
programs, one of which was a quasi-randomised controlled trial (UNHS vs no screening)
(Kennedy et al 2005). From the five studies that presented data on bilateral hearing
impairment, prevalence in well and at-risk babies was found to range between 1.0 and 1.7
per 1,000. Two of these seven studies also reported the prevalence of unilateral hearing
loss in the same population and found it to range between 0.4 and 1.5 per 1,000.
A further six studies reported on the prevalence of PCHI greater than 40 dB HL as
estimated by retrospective examinations of health and/or education records. Three of
these reported prevalence data on bilateral hearing loss and this varied between 1.1 to 1.8
per 1,000. Unilateral hearing loss was reported in one UK study using records of which
the type was not specified, and it found a prevalence of 1.2 per 1,000 (Neary et al 2003).
Three studies reported prevalence data for PCHI greater than 50 dB HL. Davis and
Wood (1992) studied a cohort of children in the Nottingham Health District, UK, born
between 1983 and 1986. In this population the prevalence of sensorineural/mixed and
congenital progressive hearing loss was 1.1 per 1,000. Similar results were obtained in
Cyprus, where a cohort of children born between January 1979 and December 1990 had
congenital hearing loss at the rate of 1.2 per 1,000 children (Hadjikakou & Bamford
2000). Finally, in a case series of babies born in the hospital of Sciacca, Cyprus between
2003 and 2004 and screened as part of a universal neonatal hearing screening program,
the prevalence of sensorineural hearing loss greater than 56 db HL was 1.87 per 1,000
babies (Martines et al 2007).
The heterogeneity in prevalence rates, ranging from 0.7–1.8 infants per 1,000 with
bilateral PCHI at all levels of hearing loss, and 0.2–1.5 infants per 1,000 with unilateral
impairment at all levels of hearing loss, is unremarkable given the different methods used
to measure PCHI, the different case definitions and the varying nature and locations of
the sampled population.
38
Universal neonatal hearing screening
Summary
Given the range of prevalence rates, a single estimate for each level of hearing impairment
was calculated from the median of the available data. The median prevalence of bilateral and
unilateral permanent childhood hearing impairment (PCHI) >35 dB HL is estimated to be 1.3
and 0.6 per 1,000 infants respectively. The median prevalence of bilateral and unilateral PCHI
> 40 dB HL is estimated to be 1.3 and 1.2 per 1,000 infants respectively. Lastly, the
prevalence of bilateral PCHI > 50 dB is estimated to be 1.2 per 1,000 infants.
Universal neonatal hearing screening
39
Prevalence of permanent childhood hearing impairment in Australia
Two studies provided data on the prevalence of congenital permanent childhood hearing
impairment (PCHI) in neonates and infants in Australia (Bailey et al 2002; Russ et al
2003). From these studies the estimated prevalence of moderate to profound bilateral
PCHI ranges from 0.7–1.1 per 1,000 infants. This is lower than the median prevalence
rate estimated from the international literature. Populations at risk of hearing impairment,
such as indigenous, rural or immigrant Australians, are possibly under-represented in the
available Australian studies. Population-level PCHI prevalence data are currently not
available for Australia, although this may change once the results of the New South
Wales Statewide Infant Screening – Hearing (SWISH) program are published.
Using the unrounded median prevalence rate of PCHI greater than 35 dB HL
(1.25/1,000) from the international literature as a proxy for the Australian rate, and an
estimated 259,800 registered births per year in Australia (ABS 2005), approximately 325
children would be born every year with moderate to profound bilateral PCHI. Unilateral
PCHI of similar severity would occur in an additional 156 children born per year.
Overall, this would result in 481 Australian children born each year with either unilateral
or bilateral moderate to profound permanent hearing impairment.
Data available from Australian Hearing, the body that provides government subsidised
hearing assessment services, hearing aid fitting and hearing rehabilitation for all
Australian children, indicates that the above estimate is reasonable (Australian Hearing
2007). Rates of hearing aid fittings in children born in the years 1986 to 2001 (the years
when there is likely to be reasonably complete data available, including ascertainment of
mild hearing loss through school entry screening) ranged from 2.11–3.43 per 1,000 (or
between 520 and 878 children per year). These data are, however, probably an overestimation due to the inclusion of cases of aided mild hearing loss, and acquired or
progressive hearing losses that would not be identified through a universal neonatal
hearing screening program.
Australian Hearing also provided data on the number of children who have had hearing
amplified artificially within the first 12 months of life (Australian Hearing 2007). When
converted to rates per 1,000 births and graphed, the trend in improvement of early (<6
months) identification of hearing loss over a 16-year period is evident (Figure 4). This is
probably due to improvements in the identification of children at risk of hearing
impairment (ie targeted neonatal hearing screening).
Australian Hearing data indicates that 83–94 per cent of children who had their hearing
amplified received their hearing aid after the age of 12 months. 2 It is unclear, however,
how many of these children had acquired hearing loss as opposed to having undetected
congenital hearing loss. It is also unclear what additional benefit a program of universal
neonatal hearing screening in Australia would provide in terms of identifying children
with PCHI in the neonatal period.
2
40
Based on total hearing aid fittings for children born in the years 1986 to 2001. (Australian Hearing 2007).
Universal neonatal hearing screening
Figure 4
Age at first hearing aid fitting for children born between 1986 and 2005
0.3
Fittings per 1,000 births
0.25
0.2
0.15
0.1
0.05
0
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
Year
< 3 months
3 to < 6 months
6 to < 12 months
Source: Adelaide Health Technology Assessment (AHTA), Discipline of Public Health, University of Adelaide. Data
provided by Australian Hearing
Summary
Credible estimates suggest that 325 Australian children are born every year with moderate to
profound bilateral PCHI. Unilateral PCHI of similar severity is estimated to occur in an
additional 156 newborns each year. Overall, it is estimated that 481 Australian children are
born annually with either unilateral or bilateral, moderate to profound PCHI.
Universal neonatal hearing screening
41
Safety of universal neonatal hearing screening
Universal neonatal hearing screening (UNHS) was assessed in terms of possible
psychosocial and physical harms that may result from any aspect of the screening
process.
Studies addressing these issues were assessed for inclusion according to the criteria
delineated in Box 4. As the aim was to assess possible harms resulting from the screening
of neonates, rather than the diagnosis of PCHI, studies with a population of only true
positives were excluded (neonates with PCHI). Harms from a true positive diagnosis of
PCHI are likely to be similar regardless of the diagnostic method. With UNHS, any
harms are likely to occur earlier in the development of the infant which may or may not
be beneficial.
Box 4
Study selection criteria for assessing safety
Research questions
Is universal hearing screening safe for neonates and infants?
– Are the otoacoustic emissions (OAE) and automated auditory brainstem response (AABR) tests physically safe for the
individual being tested?
– What are the harmful consequences of early diagnosis and management of permanent childhood hearing impairment
(PCHI)?
Selection criteria
Population
Intervention
Comparator(s)
Outcomes
Study design
Search period
Language
Inclusion criteria
Neonates and infants ≤6 months of age undergoing testing for PCHI, and/or their parents/
caregivers.
Universal, including targeted, neonatal hearing screening using either the OAE or AABR testing
methods.
No universal neonatal hearing screening.
Adverse psychological, psychosocial or physical health outcomes associated with the testing
procedure, the diagnosis and/or the treatment/management options.
Randomised or non-randomised controlled trials, cohort studies, registers or systematic reviews
of these study designs. In the event that the evidence-base lacked these study designs, casecontrol or cross-sectional studies were acceptable, although low quality, alternatives.
No restriction.
Studies in languages other than English were only translated and included if they represented a
higher level of evidence than that available in the English language evidence-base.
Physical harms
There were no reported cases of physical harm occasioned by UNHS in any of the
available studies.
Psychosocial harms
One of the important aspects of assessing safety is to determine the impact of hearing
screening results on the psychosocial health of the infant and the caregiver. Incorrect
diagnoses (false positives) may prove to have psychological consequences (eg anxiety) for
the parents or caregivers of babies who have been tested, particularly given that this
added stressor would occur in the early postnatal period. The impact of false negatives
also need to be assessed. The lack of accurate diagnosis and false reassurance may have
serious psychosocial consequences for the infants that have been tested, and their
families, particularly given the possible lengthy period before confirmation of permanent
42
Universal neonatal hearing screening
childhood hearing impairment (PCHI). These potential adverse consequences would not
occur without UNHS.
Altogether, five controlled studies (level III-2) assessed the psychosocial safety of a
UNHS program (Crockett et al 2005; Crockett et al 2006; Kennedy 1999; Kolski et al
2007; Watkin et al 1998) (see Table 9). These studies were of generally poor to average
quality, with likely selection bias due to poor response rates or participation in the
studies. One study was a nested case-control study, while the other four were cohort
studies. Four studies used questionnaires to determine the anxiety of mothers whose
babies screened positive for hearing impairment, using the validated (in different
populations), Spielberger State Trait Anxiety Inventory (Crockett et al 2005; Crockett et
al 2006; Kennedy 1999). One of the studies modified the inventory to only include 10
items (Watkin et al 1998). The remaining study used semi-structured interviews, using the
Montgomery Åsberg depression rating scale and the anxiety subscale of the Edinburgh
post-natal depression scale (Kolski et al 2007).
Neonatal hearing screening was compared against the health visitor distraction test
(performed when the infant was over 6 months old) in an average quality cohort study
(Crockett et al 2005). Anxiety levels in mothers of babies who screened positive were
compared at 3 weeks and again at 6 months after screening. Overall state anxiety levels
were between 29.0 and 36.9 on the Spielberger State-Trait Anxiety Inventory, which is in
the normal range, and well below the ‘clinical cut-off’ of 49 out of 80, so further
comparisons were not made. When satisfaction was compared between mothers of
infants who were referred for further testing (from an unvalidated single item question),
mothers of those who underwent neonatal hearing screening were significantly more
satisfied than mothers of infants who received the distraction test (p<0.05, Cohen’s
d=0.72).
Two further studies (Kolski et al 2007; Watkin et al 1998) compared anxiety levels of
mothers whose babies received the screening test against mothers who gave birth in
hospitals where screening was not available. Babies were age-matched. Kolski et al (2007)
found no significant difference between maternal anxiety, depression levels, or any effect
on the quality of early interactions when a negative screen was compared against no
screening. Similarly, when a positive screen result was compared against a control group
with no screening, Watkin et al (1998) found no statistically significant differences
between anxiety levels.
Four poor to average quality studies compared anxiety levels in mothers of infants who
were referred for further testing, against a control group of mothers whose infants
screened negative (Crockett et al 2005; Crockett et al 2006; Kennedy 1999; Kolski et al
2007). Levels of anxiety were low in the three studies that reported mean values on the
Spielberger State-Trait Anxiety Inventory. Kolski et al (2007) reported that a positive
screening test significantly increased levels of anxiety, depression and reduced the quality
of early interactions, but the clinical significance of these differences was unclear as no
raw data were provided. Crockett et al (2006) found that the more screening stages that
infants are required to undergo (as the potential seriousness of the results increase), the
more anxious the mothers became. However, the difference between these levels of
anxiety was not statistically significant. The authors proposed that the low response rate
to the questionnaire and the delay in sending questionnaires (sent 3 weeks after the
completion of screening) may have been responsible for the overall low levels of anxiety
reported. Overall, the more knowledge the mothers had about what a referral for further
Universal neonatal hearing screening
43
tests meant, the less anxious (F(1,323)=6.8, p<0.01) or worried (F(1,332)=8.0, p<0.01)
they were about the need for them.
When Kolski et al (2007) compared two different hearing screening protocols, one
performed in newborns, and another strategy at 2 months of age, no significant
differences were found between the anxiety (F(1,139)=0.53, p>0.05) or depression levels
(F(1,139=2.97, p>0.05.) in mothers of infants receiving either negative or positive
screening results. However, when an analysis of variance was performed according to
time of screening, a positive screening result had more impact on anxiety (F(3,139)=3.5,
p=0.01) and depression levels (F(3,139)=3.1, p=0.03) when the screening test was in the
initial week of life, as compared with 2 months later. Raw data were not presented, so it is
unclear whether any of the anxiety levels were outside the normal range, and thus
whether there was any clinically important difference between groups on maternal anxiety
levels.
One study (Kennedy 1999) did find a statistically significant difference in ‘concern for
baby’ but the screen positive group actually presented with less anxiety than the screen
negative group (t = -2.0, df = 148, p = 0.04). Mean anxiety scores for the control group
were the same or higher than the screen positive group in this study. Anxiety scores in
both groups were low overall. Only one study assessed the impact of the screen on
mothers’ attitudes to their babies (Kennedy 1999). This controlled study found that
mothers of screen positives and screen negatives had very low negative attitudes to their
babies and that there was no difference in mean attitudinal scores between the groups.
Information from these controlled studies was supplemented by five descriptive, crosssectional surveys (Clemens et al 2000; Hergils & Hergils 2000; Tatli et al 2007; Vohr et al
2001; Weichbold & Welzl Mueller 2001) providing information on safety outcomes for
the universal screening arm alone, and predominantly assessing the anxiety of parents of
screen positive babies (see Table 10). The cross-sectional surveys were generally of poor
quality, with anxiety measured using non-validated tools. Two of the studies were
possibly affected by recall bias as the questionnaire was administered approximately 5
months after screening (Clemens et al 2000; Hergils & Hergils 2000). One study may also
have been affected by selection bias due to a poor response rate to the survey (Clemens
et al 2000). These studies indicate that parental anxiety is low (3.8–7.0%) at the initial
hearing screen (Hergils & Hergils 2000; Tatli et al 2007; Vohr et al 2001). Parents of
babies with normal hearing who screened positive for hearing impairment at the initial
test (ie false positives) experienced varying levels of anxiety, with 80 per cent ‘worried’
(Clemens et al 2000), between 17 and 33 per cent ‘worried or very worried’ (Vohr et al
2001) and 14 per cent ‘considerably or very concerned’ (Weichbold & Welzl Mueller
2001). Lasting anxiety as the result of a false positive screen was reported in 14 per cent
of parents in one study, although the anxiety was characterised as mild (Clemens et al
2000). Parental anxiety in mothers of screen positive babies referred for diagnostic testing
was higher (21%) (Weichbold & Welzl Mueller 2001).
There was no evidence available regarding the psychological effects of false reassurance.
44
Universal neonatal hearing screening
Summary
The data available on the safety of universal neonatal hearing screening were limited and of
poor to average quality. The main outcome reported was parental anxiety concerning:
(1) the screen;
(2) a false positive result – a large consideration given the high false alarm rate; and
(3) a screen positive result.
Levels of moderate to severe anxiety were predominantly low in all three groups. There were
no clinically significant differences in anxiety between parents of screen positive and screen
negative babies, or between parents of screen positive and unscreened babies (level III-2
interventional evidence). No clinically significant differences were found between levels of
anxiety or worry about their baby’s hearing when the state of mothers of infants screened by
UNHS were compared against mothers of infants ≥ 6 months screened by a behavioural test.
More satisfaction was expressed after the UNHS than the distraction test.
It has been suggested that screen status or anxiety may have an impact on the parental
relationship with the child, but in the one controlled study (level III-2 interventional evidence)
that reported on parental attitudes to the child, no differences were ascertained. A positive
screen was associated with statistically poorer quality of early interactions, but the clinical
importance of this finding is unknown (level III-2 interventional evidence). The available
studies did not report on the psychological effects of false reassurance nor the psychosocial
impact of a true-positive diagnosis, nor mention physical harms occasioned during the
screening process.
Universal neonatal hearing screening
45
Table 9
Study
Safety of universal neonatal hearing screening (comparative studies)
Level of
evidence
Study
qualitya
Screen
setting
Population
Outcomes
Critical
appraisalb
Mean ± SD
in UNHS
group
Mean ± SD
in control
group
Mean
Statistic, df, p
difference value
[95%CI]
Controlled studies comparing neonatal screening vs behavioural screening or no screening
(Crockett
et al 2005)
III-2
Cohort
study
(Kolski et
al 2007)
III-2
Cohort
study
QS = 3.5/6
Response
rate = 48%
for HDVT,
48-49% for
neonatal
hearing
screen
Recall bias
possiblee
6 maternity
hospitals,
health
visitor
clinics and
general
practice
surgeries
England
48 mothers of infants
who underwent health
visitor distraction test (21
were positive, 27 were
negative)
42 mothers of neonates
who received UNHS (16
were positive, 26 were
negative)
UNHS versus HVDT in
screen positive infants
Parental anxiety –
Anxiety state
Scale: 20 (low) –
80 (high)
Clin I = unable to
be determined
R = 3/5
3 weeks
36.9 ± 11.9
6 months
30.4 ± 8.2
3 weeks
33.2 ± 10.1
6 months
35.4 ± 12.8
3 weeks
3.7
6 months
-5.0
within normal
range so further
comparisons not
performed
Worry about baby’s
hearing
Scale: 0 (not at all
worried) – 7
(extremely worried)
(unvalidated)
Clin I = unable to
be determined
R = 5/5
3 weeks
2.6 ± 2.1
6 months
1.6 ± 1.2
3 weeks
3.3 ± 2.0
6 months
2.1 ± 1.6
3 weeks
-0.7
6 months
-0.5
p>0.05
QS = 1.5/6
Response
rate = not
stated
No recall
biasc
Maternity
hospital,
University
Hospital of
Picardy
115 mothers of well
babies who were
screened, 58 by 1st
strategy (at birth), 57 by
2nd strategy (2 months)
28 mothers of
unscreened babies
Screen negative
versus
unscreened
Parental anxiety
(1st strategy)
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
t = 0.39, df = 43
p > 0.05
Parental anxiety
(2nd strategy)
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
t = 0.69, df = 25
p > 0.05
Post-partum
depression (1st
strategy)
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
t = 0.25, df = 43
p > 0.05
Post-partum
depression (2nd
strategy)
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
t = 0.78, df = 25
p > 0.05
Quality of early
interactions (1st
strategy)
Clin I = unable to
be determined
R = 5/5
N/A
N/A
N/A
t = 0.03, df = 43
p > 0.05
Quality of early
interactions (2nd
strategy)
Clin I = unable to
be determined
R = 5/5
N/A
N/A
N/A
t = 0.95, df = 25
p > 0.05
Amiens,
France
46
Universal neonatal hearing screening
Study
Level of
evidence
Study
qualitya
Screen
setting
Population
Outcomes
Critical
appraisalb
Mean ± SD
in UNHS
group
Mean ± SD
in control
group
Mean
Statistic, df, p
difference value
[95%CI]
(Watkin et
al 1998)
III-2
QS = 3.5/6
Response
rate = 60%
No recall
biasd
Whipps
Cross
Hospital
57 mothers of rescreen babies
versus
61 mothers of
unscreened babies
Parental anxiety
Anxiety state
Clin I = 4/4
R = 3/5
14.8±3.9
16.3±5.2
-1.6
[-3.3, 0.1]
t = -1.8
df = 116
p = 0.07
Anxiety trait
Clin I = 4/4
R = 3/5
16.5±4.5
17.5±4.9
-1.0
[-2.7, 0.7]
t = -1.1
df = 116
p = 0.25
Parental anxiety
after HVDT –
Anxiety state
Scale: 20 (low) –
80 (high)
Clin I = unable to
be determined
R = 3/5
3 weeks
post test
33.2 ± 10.1
6 months
post test
35.4 ± 12.8
3 weeks post
test
29.0 ± 11.1
6 months
post test
32.8 ± 12.4
3 weeks
post test
4.2
6 months
post test
2.6
means within
normal range, so
not compared
further
Parental anxiety
after UNHS –
Anxiety state
Scale: 20 (low) –
80 (high)
Clin I = unable to
be determined
R = 3/5
3 weeks
post test
36.9 ± 11.9
6 months
post test
30.4 ± 8.2
3 weeks post
test
31.8 ± 11.1
6 months
post test
32.6 ± 8.9
3 weeks
post test
5.1
6 months
post test
-2.2
means within
normal range, so
not compared
further
Worry about baby’s
hearing after HVDT
Scale: 0 (not at all
worried) – 7
(extremely worried)
(unvalidated)
Clin I = unable to
be determined
R = 5/5
3 weeks
post test
3.3 ± 2.0
6 months
post test
2.1 ± 1.6
3 weeks post
test
1.1 ± 0.3
6 months
post test
1.1 ± 0.3
3 weeks
post test
2.2
6 months
post test
1.0
3 weeks post
test
p < 0.01
6 months post
test
p < 0.05
Cohort
London,
UK
Controlled studies comparing screen positive with screen negative
(Crockett
et al 2005)
III-2
Cohort
study
QS = 3.5/6
Response
rate = 48%
for HDVT,
48-49% for
neonatal
hearing
screen
Recall bias
possiblee
6 maternity
hospitals,
health
visitor
clinics and
general
practice
surgeries
England
Universal neonatal hearing screening
48 mothers of infants
who underwent health
visitor distraction test (21
were positive, 27 were
negative)
42 mothers of neonates
who received hearing
screening (16 were
positive, 26 were
negative)
Screen positive versus
screen negative
47
Study
(Crockett
et al 2006)
Level of
evidence
Level III-2
Cohort
study
(Kennedy
1999)
(Kennedy
1999)
48
III-2
Nested
case-
Study
qualitya
QS = 4/6
Response
rate = 53%
Possible
recall biasf
QS = 2.5/5
Response
rate = 75%
Possible
Screen
setting
Hospitals
participatin
g in UNHS
pilot
program
England
4 maternity
hospitals
Wessex,
Outcomes
Critical
appraisalb
Mean ± SD
in UNHS
group
Mean ± SD
in control
group
Mean
Statistic, df, p
difference value
[95%CI]
Worry about baby’s
hearing after
UNHS
Scale: 0 (not at all
worried) – 7
(extremely worried)
(unvalidated)
Clin I = unable to
be determined
R = 5/5
3 weeks
post test
2.6 ± 2.1
6 months
post test
1.6 ± 1.2
3 weeks post
test
1.1 ± 0.3
6 months
post test
1.1 ± 0.3
3 weeks
post test
1.5
6 months
post test
0.5
3 weeks post
test
p < 0.01
6 months post
test
p = 0.10
344 mothers of infants
were screened
Group 1: clear
responses in both ears
from 1st or 2nd stage
OAE
Group 2: not clear
responses in one or both
ears at 1st or 2nd stage
OAE but clear on AABR
Group 3: not clear
responses in one ear on
AABR and referred for
possible unilateral
hearing loss
Group 4: not clear
responses in either ear
on AABR and referred
for possible bilateral
hearing loss
Trend analysis of
between groups
Parental anxiety–
Anxiety state
Scale: 20 (low) –
80 (high)
Clin I = unable to
be determined
R = 3/5
Group 2:
32.7 ± 12.1
Group 3:
34.0 ± 9.5
Group 4:
35.7 ± 12.8
Group 1:
32.0 ± 11.1
N/A
F(3,327) = 1.5
p = 0.22
Worry about baby’s
hearing
Scale: 0 (not at all
worried) – 7
(extremely worried)
(unvalidated)
Clin I = unable to
be determined
R = 5/5
Group 2:
1.4 ± 1.0
Group 3:
2.7 ± 1.9
Group 4:
3.1 ± 2.2
Group 1:
1.3 ± 1.0
N/A
F(3, 337)= 26.6
p < 0.01
Mothers of 150 low-risk
babies
75 screen positive
Negative attitude
to baby (validated)
Scale: 0 (low) – 21
(high)
Clin I = 4/4
R = 1/5
4±3
4±3
0
[-1.0, 1.0]
t=0
df = 148
p = 1.00
Population
Universal neonatal hearing screening
Study
(Kolski et
al 2007)
Level of
evidence
Study
qualitya
Screen
setting
Population
Outcomes
Critical
appraisalb
Mean ± SD
in UNHS
group
Mean ± SD
in control
group
Mean
Statistic, df, p
difference value
[95%CI]
control
recall biasg
UK
versus
75 screen negative
Concern for baby
Scale: 0 (low) – 8
(high)
Clin I =3/4
R = 5/5
2±3
3±3
-1
[-2.0, 0.0]
t = -2.0
df = 148
p = 0.04
III-2
QS = 1.5/6
Response
rate = not
stated
No recall
biasc
Maternity
hospital,
University
Hospital of
Picardy
115 mothers of well
babies who were
screened, 58 by 1st
strategy (at birth), 57 by
2nd strategy (2 months)
Screen positive
versus
Screen negative
Parental anxiety
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
F(3,139) = 11.3
p < 0.01
Post-partum
depression
Clin I = unable to
be determined
R = 3/5
N/A
N/A
N/A
F(3,139) = 11.9
p < 0.01
Quality of early
interactions
Clin I = unable to
be determined
R = 5/5
N/A
N/A
N/A
F(3,139) = 6.8
p < 0.01
Cohort
study
Amiens,
France
a This includes a quality score (QS) derived from the NHMRC (2000)) checklist (see Appendix E) with a high score indicating good quality; b determination of statistical precision (SP), rank scores for the clinical importance (Clin I) of
the benefit/harm (with 1 ranked as highly clinically important and 4 ranked as clinically unimportant), and rank scores for the relevance (R) of the evidence (with 1 ranked as a highly relevant outcome and 5 as an unproven surrogate
outcome); c interviews conducted after first stage screen, and prior to second stage; d Parent Stress Index administered by telephone 1 month after discharge of screen negatives and just prior to retest of screen positives; e
Questionnaires sent 3 weeks and 6 months after completion of screening; f questionnaire completed 4 weeks after completion of screening; g questionnaire administered 2-12 months after screening. HVDT= Health Visitor Distraction
Test; UNHS=universal neonatal hearing screening, N/A=not available.
Universal neonatal hearing screening
49
Table 10
Safety of universal neonatal hearing screening (noncomparative studies)
Study
Level of
evidence
Study quality
Screen setting
Population
Outcomes
(Clemens et
al 2000)
Crosssectional
survey
Response rate = 64%
Possible recall biasa
Women’s Hospital,
Greensboro
North Carolina, USA
49 parents of 76 well babies
with normal hearing who
screened positive
Parental anxiety – ‘worried’
80.0
Lasting anxiety
14.0c
(Hergils &
Hergils
2000)
Crosssectional
survey
Response rate = 95%
Possible recall biasb
University Hospital of
Linköping
Linköping, Sweden
Parents of 83 well babies who
were screened
Parental anxiety
7.0d
(Tatli et al
2007)
Crosssectional
survey
Response rate = 78%
Possible recall biase
Dokuz Eylul University
Hospital, Izmir
Turkey
Parents of 711 well and at-risk
babies
Parental anxiety about screening
program – ‘a little concern’
3.4
Parental anxiety prior about screening
program – ‘moderate concern’
0.4
Parents of
screened babies
(%)
Parents
of screen
positives
(%)
Parental anxiety about positive result–
‘a little concern’
57.1
Parental anxiety about positive result –
‘moderate or greater concern’
7.8
(Vohr et al
2001)
Crosssectional
survey
Response rate = 85%
No recall biasf
Women & Infants
Hospital
Rhode Island, USA
307 mothers of initial screen
babies
40 mothers of rescreen babies
Parental anxiety – ‘worried / very
worried’
(Weichbold &
Welzl Mueller
2001)
Crosssectional
survey
Response rate = 84%
No recall bias
University Hospital
Innsbruck, Austria
85 mothers of well babies with
normal hearing who screened
positive
43 mothers of screen positives
referred for diagnostic testing
Parental anxiety – ‘considerably / very
concerned’
1997 screen: 3.8
1999 screen: 4.0
1997 rescreen: 33.0
1999 rescreen: 17.0
Screen
positives:
21.0
False
alarms:
14.0
survey administered at mean of 4.9 months after screening [range: 2–13 months]; b questionnaire administered 5–6 months after screening; c predominantly ‘mild’ anxiety, unclear how long after the screening this measure of
anxiety was taken; d most were parents of children that needed to be retested; e Unclear when mothers were interviewed; f Iinterviews conducted after first stage screen, and prior to second stage
a
50
Universal neonatal hearing screening
How accurate are the screening tests?
Test accuracy is crucial to the successful implementation of any screening program.
Expert opinion suggests that there are four factors that can be sources of error when
conducting hearing screening tests – the infant, the screener, the equipment and the
environment. Invalid test results can occur when:
(1) The infant:
•
is not asleep or settled, so internal noise and muscle movement affect the
test;
•
is not positioned optimally such that the ear canal is blocked, or testing is
impeded by the position of the head, pillow or mother’s arm; or
•
is not swaddled or wrapped and moves during the test, causing the insert
probe or muffins/couplers to move.
(2) The screener:
•
is unfamiliar with the test equipment and inexperienced at determining
whether the test result is valid;
•
is inexperienced at handling infants;
•
positions the insert probe or muffins/couplers inadequately or places the
electrodes poorly (creating impedance problems); or
•
allows insufficient time for testing.
(3) The equipment:
•
malfunctions;
•
is calibrated incorrectly; or
•
has an occluded probe tip.
(4) The environment:
•
has too much background noise; or
•
causes electrical interference (eg monitors affect both automated auditory
brainstem response (AABR) and otoacoustic emissions (OAE) tests).
The ‘Guidelines for using screening devices’ that were developed as part of the Western
Australian universal neonatal hearing screening program (Bailey 2003) are reproduced in
Appendix H.
Universal neonatal hearing screening
51
In order to evaluate the diagnostic accuracy of the tests routinely used to screen for
hearing impairment in neonates and infants in an evidence-based manner, criteria for
selecting studies that assessed diagnostic accuracy were delineated (Box 5).
Box 5
Study selection criteria for diagnostic accuracy
Research question
What is the diagnostic accuracy of the tests for permanent childhood hearing impairment when conducted on the neonate
or infant?
Selection criteria
Population
Intervention
Reference standard(s)
Outcomes
Study design
Search period
Language
Inclusion criteria
Children ≤6 years of agea who have not experienced particular diseases or traumas associated
with hearing impairment between the neonatal testsb for permanent childhood hearing
impairment (PCHI) and the later tests for hearing loss.
The otoacoustic emissions (TEOAE or DPOAE) test or the automated auditory brainstem
response (AABR) test for PCHI (≥35 dB) performed at ≤6 months of age.
Medical or behavioural assessment (including pure tone audiogram), tympanometry, steady state
evoked potential (SSEP) testing and/or conventional or diagnostic auditory brainstem response
(ABR) testing performed at ≤6 years of age.
Sensitivity and specificity (and therefore rates of false positives and negatives), positive and
negative predictive values (and therefore false alarm and reassurance rates).
Cross-sectional studies where allc neonates and infants are cross-classified on the test and
reference standard (including longitudinal studies). Case-control diagnostic studies were only
acceptable if cross-sectional studies were not available.
The OAE and AABR tests have only been in common usage since 1980. Studies published
before 1980 were not included.
Studies in languages other than English were only translated and included if they represented a
higher level of evidence than that available in the English language evidence-base.
This age range was selected to accommodate the reference standard which may often be performed at pre-school or school age. Studies
were excluded if children were selected solely on the basis of risk factors or indications – so as to avoid spectrum bias (Brenner & Gefeller
1997); b to mitigate possible difficulties with confounding caused by the delay between neonatal and later tests ie progressive deterioration,
disease or trauma that would worsen hearing for the later (reference standard) tests; c to prevent ‘work-up or verification bias.
a
Studies were excluded that selected children for hearing screening solely on the basis of
risk factors or indications. These studies were excluded partly because of the attendant
bias in estimating the efficacy of a test on a sample that does not contain the whole
spectrum of ill and healthy babies. They were also excluded because children at risk of
permanent childhood hearing impairment (PCHI) are often tested diagnostically in the
first instance, rather than screened.
As it was a requirement that all neonates and infants should be cross-classified on the
test and reference standard, and that only an appropriate spectrum of infants should be
included, the only levels of evidence (as defined in Table 5) available for inclusion were
levels I, II, III-1, III-2, and III-3.
There were many studies available that compared the screening tests to an acceptable
reference standard in high-risk populations of infants. There were also several studies
that compared the two screening tests to each other, rather than to an established
reference standard. There were very few studies, however, that assessed the accuracy of
the neonatal hearing screening tools in predominantly healthy infants – the target
population in a universal screening situation – and compared them to an established
reference standard (eg diagnostic auditory brainstem response testing).
52
Universal neonatal hearing screening
The ideal method for assessing the diagnostic accuracy of universal hearing screening
tests would be to test a random or consecutive sample of neonates and infants under 6
months of age with either the otoacoustic emissions (OAE) test or automated auditory
brainstem response (AABR) test, and then immediately test them on a reference standard
such as conventional or diagnostic ABR. The diagnostic accuracy of the tests can then be
determined using the classic 2 x 2 table, whereby the results of the screening test are
cross-classified against the results of the reference standard (Armitage et al 2002; Deeks
2001) and Bayes’ Theorem is applied:
Hearing status (based on reference standard)
Screening test
PCHI
Normal
Test +
True positive
False positive
Total positive
Test -
False negative
True negative
Total negative
Total with PCHI
Total without PCHI
The sensitivity of a hearing screening test is therefore calculated as the proportion of
infants with permanent childhood hearing impairment (PCHI) who have positive
screening test results:
Sensitivity = True positive / Total with PCHI
The specificity of a hearing screening test is calculated as the proportion of infants
without PCHI who have normal screening test results:
Specificity = True negative / Total without PCHI
The false positive rate of a hearing screening test is calculated as the proportion of
infants without PCHI who have positive screening test results:
False positive rate = False positive / Total without PCHI
Similarly the false negative rate of a hearing screening test is calculated as the proportion
of infants with PCHI who have negative screening results:
False negative rate = False negative / Total with PCHI
The false alarm rate of a hearing screening test is calculated as the proportion of infants
with positive screening results who do not have PCHI:
False alarm rate = False positive / Total test positive
The false reassurance rate of a hearing screening test is calculated as the proportion of
infants with negative screening results who have PCHI:
False reassurance rate = False negative / Total test negative
The positive predictive value of the test(s) in an average screening population (well and
‘at-risk’ babies) estimates the likelihood that an infant has PCHI on the basis of a positive
screening test result. The positive predictive value of a test is sensitive to the prevalence
of the condition in the population being tested (Deeks 2001). Therefore, if the
Universal neonatal hearing screening
53
prevalence of the condition is low, the positive predictive value of the test will be poor.
The formula for calculating the positive predictive value (PPV) is:
PPV =
sensitivity x prevalence of PCHI
(sensitivity x prevalence of PCHI) + (1 – specificity x prevalence of not having PCHI)
The prevalence of PCHI used in these calculations is the median estimated prevalence
rate determined in the previous section, that is 1.25/1,000.
It is particularly important that the cross-classification on the screening and reference
tests occurs with as little time-lag as possible. This is due to the nature of PCHI. Hearing
impairment may be congenital (which is the target of a neonatal screening program) or an
acquired or progressive condition. To accurately assess the diagnostic capabilities of a
neonatal screening test, therefore, the sensitivity and specificity of the test should not be
influenced by conditions that affect hearing in the postnatal period. It was fortunate that
the studies assessed as suitable for inclusion in this review conducted the crossclassification of infants on the screening test and reference standard at approximately the
same time.
Five studies met the inclusion criteria (Table 11). Profiles of these studies are provided in
Appendix F. Three of the studies compared a transient otoacoustic emissions (TEOAE)
test with a conventional auditory brainstem response (ABR) test (Jacobson & Jacobson
1994; McNellis & Klein 1997; Smyth et al 1999). One study compared the accuracy of
TEOAE to tympanometry (Ho et al 2002) and one study compared the automated
auditory brainstem response (AABR) screening test with a conventional ABR test
(Schauseil-Zipf & Von Wedel 1988). The latter study was originally in German and has
been translated into English for this assessment. There were no studies available that
compared the screening distortion product otoacoustic emissions (DPOAE) test with a
relevant reference standard.
The ability of TEOAE testing – in a one-stage screen – to accurately identify permanent
childhood hearing impairment (PCHI) in neonates and infants varied widely in the
included studies, with sensitivity ranging from 50 to 100 per cent when compared to
conventional ABR testing. This variation appears to be largely a result of the conditions
under which the testing was conducted. Both studies that utilised a ‘quiet’ – although not
sound-proofed – environment for testing elicited sensitivity results of 100 per cent,
including the better quality study produced by Smyth and colleagues (1999). Jacobson
and Jacobson (1994), however, determined 50 per cent sensitivity for TEOAE to detect
PCHI under ‘real world’ ambient noise conditions (ie within the nursery). In this study
failures included those children who were tested but could not produce an OAE in the
allocated screening time (45 minutes), mainly due to excessive noise levels. The
comparator was a combination of AABR and ABR testing.
Given the likely low sensitivity of TEOAE testing under ‘real world’ noise conditions,
and the concomitant large proportion of infants (50% identified by Jacobson and
Jacobson, 1994) with PCHI who may not be identified (false negatives) under these
conditions, TEOAE testing should only occur in environments that are quiet or possibly
sound-proofed. It is critical that false negatives are avoided as this false reassurance may
have considerable impact on when the children are re-tested, as well as the effectiveness
of the eventual rehabilitation.
54
Universal neonatal hearing screening
The specificity of TEOAE is also variable, although the better quality study by Smyth
and colleagues (1999) indicates that it can accurately determine a lack of hearing
impairment in normal infants under quiet conditions in 92 per cent of cases. The number
of false positive results was 3 out of 36 (8.3%) infants (Smyth et al 1999). In the study
performed under normal noise conditions the false positive rate was 48 per cent
(Jacobson & Jacobson 1994). The positive predictive value of an initial TEOAE
screening test under the best (quiet) conditions is very low (1.5%) – meaning that a
failure on an initial TEOAE test would accurately predict PCHI in only one to two
infants out of 100 identified by the test with the condition. This is probably a
consequence of the frequency of transient losses in newborns (ear occlusion), as well as
the low prevalence of PCHI in the general population.
In terms of identifying conductive hearing loss, TEOAE testing was found to have 100
per cent sensitivity and specificity, as compared to tympanometry, in one study of infants
who had no cerumen occlusion of the ear (Ho et al 2002).
Universal neonatal hearing screening
55
Table 11
Diagnostic accuracy of hearing screening tests
Study
Diagnostic
level of
evidence
Quality
scorea
Population
Setting
Sensitivity
(%)
[95%CI]
Specificity
(%)
[95%CI]
PPV (%)b
10/14
n = 119
babies at
33–41
weeks of
age (238
ears)
Norfolk,
Virginia,
USA
50.0
[15.7,84.3]
52.3
[45.4,59.1]
0.13
TEOAE vs conventional ABR
(Jacobson &
Jacobson
1994)
III-2c d
56% at risk
44% well
(McNellis &
Klein 1997)
III-2c
10/14
n = 50
healthy, lowrisk full-term
babies (100
ears)
Charleston,
South
Carolina,
USA
100.0
[15.8,100.0]
62.2
[51.9,71.8]
0.32
(Smyth et al
1999)
III-1
11/14
n = 37
normal, fullterm well
babies (74
ears)
without risk
factors
Brisbane,
Queensland
Australia
100.0
[2.5,100.0]
91.7
[77.5,98.3]
1.49
III-2
10/14
n = 33
normal and
at-risk
babies <6
months of
age,
excluding
those with
cerumen
occlusion of
the ear (total
of 29 ears)
Community
screening
clinics (29)
in
Minnesota,
USA
100.0
[54.1,100.0]
100.0
[85.2,100.0]
100.00
III-2c
9/14
n = 50
babies (100
ears)
Women and
children’s
clinic,
Cologne,
Germany
80.0
[44.4,97.5]
95.6
[89.0,98.8]
2.22
TEOAE vs tympanometry
(Ho et al
2002)
AABR vs ABR
(SchauseilZipf & Von
Wedel 1988)
50% at risk
50% well
PPV = positive predictive value; TEOAE = transient evoked otoacoustic emissions test; AABR = automated auditory brainstem response
test; ABR = conventional auditory brainstem response test. a Quadas checklist appraising quality of studies of diagnostic accuracy, see
Appendix E; b calculated using median prevalence estimate of PCHI as 1.25/1,000; c assumption that selection was non-consecutive as
‘consecutive’ not mentioned in text; d reference standard included combination of conventional ABR and AABR.
56
Universal neonatal hearing screening
The ability of the automated auditory brainstem response (AABR) test – in a one-stage
screen – to accurately identify PCHI in neonates and infants was compared to
conventional ABR testing in the one average quality study available that cross-classified
infants on the two tests (Schauseil-Zipf & Von Wedel 1988). The specificity (or the
likelihood of detecting normal hearing in a normal hearing individual) of AABR was
particularly good, given that its primary use is as a screening tool on a population of
predominantly healthy infants. However, the trade-off between sensitivity and specificity
means that the test has good, although not excellent, sensitivity at detecting PCHI and
thus some false negatives may result (20% in this study). As mentioned earlier, it is
critical that false negatives are avoided as they may have a significant impact on when
children are re-tested, as well as on eventual rehabilitation. The AABR equipment used in
this study was the earliest version available on the market and it is possible that newer
models have improved sensitivity. Expert opinion indicates that later models of the
AABR may, in fact, have improved diagnostic accuracy. Unfortunately, empirical
evidence on the accuracy of the newer AABR models at detecting hearing impairment in
predominantly healthy neonates is not yet available. In this study on the earliest AABR
test, the positive predictive value is still very low (2.2%), although marginally better than
TEOAE testing conducted under quiet conditions.
Universal neonatal hearing screening
57
Summary
Average quality, diagnostic level III-2 evidence suggests that the accuracy of transient evoked
otoacoustic emissions (TEOAE) testing appears to depend on the level of local ambient noise
(and therefore ear-probe fit and the testing environment), as well as the condition of infant ears
(eg whether occluded by vernix or wax) at testing (Ho et al 2002; Jacobson & Jacobson 1994;
McNellis & Klein 1997). If these factors are addressed adequately, diagnostic accuracy of the
test is very good (up to 100% sensitivity), although even under the best conditions the rate of
false positives can still be quite high (8%). The positive predictive value of an initial TEOAE test
is very low, with 1.5 per cent of children who screen positive for hearing impairment receiving
diagnostic confirmation. This is probably a consequence of the frequency of transient losses in
newborns (ear occlusion), as well as the low prevalence of permanent childhood hearing
impairment (PCHI) in the general population.
Based on one study, the specificity of the early model automated auditory brainstem response
(AABR) test is particularly good, given that its primary use is as a screening tool on a
population of healthy infants. However, the trade-off between sensitivity and specificity means
that the test has good, although not excellent, sensitivity at detecting PCHI and thus some false
negatives may result. The positive predictive value of an AABR test is also very low (2.2%),
although marginally better than a TEOAE test conducted under quiet conditions. Evidencebased assessment of more recent versions of the AABR test is required.
The number of false positives associated with either test could be reduced with the introduction
of a second-stage or third-stage screen of initial failures prior to diagnostic testing. This may,
however, result in unnecessary caregiver anxiety and added costs and delays in rehabilitation.
False negatives are not likely to be picked up until the child is older and this false reassurance
may lengthen the time until diagnostic assessment and thus the child’s rehabilitation.
58
Universal neonatal hearing screening
Is it effective to screen all neonates for hearing impairment?
An assessment of the effectiveness of a screening program is usually modelled as a
systematic review of intervention studies. That is, studies are sought that directly
compare the impact of screening versus not screening in the general population. The
effectiveness of the screening program (as a whole) at preventing, or allowing early
treatment of, potential adverse outcomes associated with a condition is then assessed.
Box 6 delineates the criteria for including studies to assess the effectiveness of a universal
neonatal hearing screening (UNHS) program.
Box 6
Study selection criteria for assessing effectiveness
Research questions
1. Does universal neonatal hearing screening, and the finding of a positive and/or negative test, affect the clinical
management or treatment options available to permanently hearing-impaired infants?
2. Does universal neonatal hearing screening, and therefore possible alterations in clinical management, have an impact
on the adverse outcomes associated with permanent childhood hearing impairment?
Selection criteria
Population
Intervention
Comparator(s)
Outcomes
Study design
Search period
Language
Inclusion criteria
Neonates and infants ≤6 months of age.
Universal neonatal hearing screening using either the otoacoustic emissions (OAE) or automated
auditory brainstem response (AABR) testing methodsa.
Not universal neonatal hearing screening.
Primary – screening yield, rate and quality of language acquisition, behaviour, family functioning,
communication ability / social functioning, educational achievement, employment status,
socioeconomic status, quality of life.
Secondary – age of referral for diagnostic testing, age of permanent childhood hearing
impairment diagnosis, age receiving therapeutic intervention.
Randomised or non-randomised controlled trials or cohort studies or systematic reviews of these
study designs. Case-control studies were acceptable only for the long-term (primary) outcomes.
Uncontrolled studies that provided, at minimum, screening yield data were included to provide
supplementary descriptive data on screening parameters.
The OAE and AABR tests have only been in common usage since 1980. Studies published
before 1980 were not included.
Studies in languages other than English were only translated and included if they represented a
higher level of evidence than that available in the English language evidence-base.
excluding studies solely concerned with targeted screening, ie children tested on the basis of existing risk factors or indications for permanent
childhood hearing impairment (PCHI), such as i) admission to a neonatal intensive care unit; (ii) prolonged usage of aminoglycosides; (iii)
family history of hearing impairment; (iv) intrauterine or perinatal infection (either suspected or confirmed); (v) birthweight less than 1.5 kg; (vi)
craniofacial deformity: (vii) birth asphyxia; (viii) chromosomal abnormality, including Down syndrome; or (ix) exchange transfusion or
intrauterine transfusion
a
Universal neonatal hearing screening
59
Research question 1:
Does universal neonatal hearing screening, and the finding of a positive
and/or negative test, affect the clinical management or treatment options
available to permanently hearing-impaired infants?
In order to determine whether the clinical management of permanent childhood hearing
impairment (PCHI) is altered by the introduction of universal neonatal hearing screening
(UNHS), secondary or surrogate outcome data were extracted from the controlled
studies. These included age at referral, age at PCHI diagnosis and age at treatment or
management.
Given the paucity of controlled studies available, information from uncontrolled studies
of screening programs – full text only – was collated to provide descriptive,
supplementary data. Inclusion of these descriptive studies was limited to only those that
provided data on screening yields. Additional information was also extracted on failure
rates (referrals), false alarm rates (incorrectly testing positive for hearing impairment),
coverage and loss to follow-up for different screening protocols and populations. These
studies could not, however, assist with the determination of the effectiveness of UNHS
compared to targeted (‘at-risk’ population) screening or not screening at all.
Altogether, five controlled studies (Table 12) assessed the effectiveness of a UNHS
program in terms of its impact on clinical management of hearing-impaired infants
(Kennedy et al 1998; Kennedy et al 2005; Kennedy et al 2006; Nekahm et al 2001a;
Neumann et al 2006; Weichbold et al 2006; Yoshinaga-Itano et al 2001). The study
conducted by Kennedy and colleagues was a quasi-randomised (alternate allocation)
controlled trial comparing periods with and without UNHS at four different hospitals
over 3 years (Kennedy et al 1998). This trial was conducted by the Wessex Universal
Neonatal Hearing Screening Trial Group and was of average quality (QS = 3/6) overall,
and there is a possibility that bias may have influenced the results through possible
contamination from one screening period to another within each hospital. This is likely
to dilute any effect of screening, however, rather than over-estimate its effect. The study
was under-powered for some of the outcomes, so it is possible that clinically relevant
differences have been missed or are statistically imprecise. Bias may have had an impact
on the results through the lack of proper randomisation and concealment of allocation to
the screening intervention. Statistical analysis also did not take into account the clustering
of the data.
Two subsequent studies from the same group reported results on 8 years of follow-up of
the original birth cohort of babies in the Wessex controlled trial (Kennedy et al 2005;
Kennedy et al 2006). High quality was achieved for the latter study based on the critical
appraisal checklists (see Appendix E). The most recent of these studies, a prospective
cohort study, also included a large birth cohort from the Greater London region
(Kennedy et al 2006)
Kennedy et al (2005) reported an 8 year follow-up of results in a research letter which
was primarily focused on publishing an estimate of the effect of UNHS on the
proportion of all true cases of PCHI ≥40dB HL that had early referral. Identifying an
accurate estimate requires long-term follow-up so that false negatives and true
60
Universal neonatal hearing screening
progressive PCHI cases can be correctly ascertained, thus this was not discussed in the
original published trial report.
The second follow-up study (Kennedy et al 2006) consisted of neonates from eight
districts: four districts from the Wessex controlled trial and four from the Greater
London group. Nonetheless, outcome measurement of identification and diagnosis was
comparable across the groups. Primarily of interest in this long-term study was
communication ability and language acquisition in the children. Researchers blinded to
the hearing history of the children assessed the children by means of four different
measures: Test for Reception Grammar, the British Picture Vocabulary Scale (receptive
language), the Renfrew Bus Story Test (expressive language) and Raven’s Progressive
Matrices Test (nonverbal abilities). Completeness of ascertainment was reported to be
over 95 per cent in both the Wessex and London groups. It was also apparent that
participants and non participants were similar with respect to age, sex and severity of
hearing loss.
The remaining four studies analysed, retrospectively, cohorts of children with hearing
impairment, and assessed whether UNHS affected the time of diagnosis of PCHI and
the age at which management was initiated. The study of Tyrolean children born in time
periods with and without UNHS was of good quality (QS = 5.5/6) (Nekahm et al
2001a). The results are likely to have been affected by bias, due to the lack of
randomisation and allocation concealment, as well as by confounding as the cohorts of
children were recruited over different time periods. It is possible, therefore, that
differences in time of diagnosis are attributable to some factor other than screening (eg
heightened parental or clinician awareness concerning certain risk factors). This group of
researchers also conducted an average quality retrospective follow-up study of children
born since 1990 and registered at an Austrian Ear, Nose and Throat (ENT) department
or institution for the hearing impaired (Weichbold et al 2006). These included 15
institutions out of a total of 35 Austrian institutions thus expanding the population
captured in their earlier study which only included Tyrolean children born between 1980
and 1999.
The study of German children, born in hospitals with and without UNHS was of good
quality (QS=5/6). This study compared two groups: one receiving Universal Neonatal
Hearing Screening and the other group receiving opportunistic testing outside of a
screening program, who originated from different populations in Germany (Neumann et
al 2006). The 17, 349 screened neonates came from the state of Hessen only, while the
unscreened neonates arose from a database of the Hessian and Thuringian population.
The analysis did not control for demographic and clinical differences in the two groups
thus the two groups were not entirely comparable.
The fourth retrospective cohort study conducted by Yoshinaga-Itano et al (2000 & 2001)
in the United States was also considered to be of good quality (QS = 5.5/6) (Yoshinaga
Itano et al 2000; Yoshinaga-Itano et al 2001) and bias and confounding are likely to have
had limited effect. Bias may have been introduced through unblinded assessment of
some of the outcomes. Confounding was well controlled with the use of a matched-pairs
design. Children were matched on age, severity of PCHI and cognitive ability – all factors
that had been associated with the identification, diagnosis and management of PCHI in
previous studies. This may, however, have limited the external validity of the study’s
results. Other potential confounders such as gender, ethnicity, other disabilities, mode of
Universal neonatal hearing screening
61
communication and education level of the primary caregiver were not distributed
differently between the screening and not-screening groups.
Table 12 provides a summary of data from the controlled studies on each of the
following screening outcomes: coverage, absolute and incremental yield, age at referral,
age at PCHI diagnosis and age at management or rehabilitation. This is supplemented in
Table 13, Table 14, and Table 15 by 56 uncontrolled, descriptive studies. These studies
provide qualitative information on screen protocol (1-stage, 2-stage, 3-stage), coverage,
failure rates (referrals), loss to follow-up (LTFU), false alarms and yield for the universal
screening arm alone. Several studies are mentioned repeatedly in the tables as they have
undertaken different screening protocols in different settings therefore their results have
been put in the appropriate sections of the table. All information presented in these
tables has been calculated according to intention-to-screen principles – data on LTFU,
however, are also presented so that compliance with the screening protocol can be
determined.
62
Universal neonatal hearing screening
Table 12
Effectiveness of universal neonatal hearing screening for secondary outcomes (controlled studies)
Study
Level of
evidence
Study
quality a
Setting
Population
Screen fail
criterion
Outcomes
4 hospitals
in Wessex,
UK
53,781 well
and at-risk
babies
n = 25,609
born during
UNHS
periods
n = 28,172
born during
periods
without
UNHS
Bilateral
failure to
produce
emission
spectrum
of
significant
gain
across 3 of
5 testing
frequency
bands
Coverage
Study
duration: 3
years
AABR fail
at ≥35 dB
HL
Critical
appraisalb
UNHS
rate
(%)
Control Number
rate (%) needed to
screen (NNS)/
diagnose
(NND)c
Relative
risk
[95% CI]
Adjusted
odds ratiod
[95% CI]
0.03
2.9
[1.4, 6.3]
19.0
[3.2, 111.0]
2-stage: TEOAE–AABR – dx audiology
(Kennedy et
al 1998)
III-1
QS = 3/6
(Kennedy et
al 2005) (8
yr follow up)
Control
group
received
health
visitor
distraction
test
Universal neonatal hearing screening
Referrale
<6 months
83.1
SP = ave.
Clin I = 1/4
R = 2/5
Power = 77%
False alarm
rate
False
negative
0.09
NNS
1,619
[955, 5297]
1.5
Clin I =4/4
R =2/5
False
reassurance
rate
6.5
17.1f
0.4
[0.1, 1.7]
0.005
Yieldg
SP = ave.
Clin I = 2/4
R = 2/5
Power = 58%
0.09
0.04
NNS
1,970
[1063, 13459]
2.3
[1.1, 4.7]
Diagnosis
<10 months
SP = NS
Clin I = 4/4
R = 2/5
Power = 23%
0.06
0.04
NNS
3,706
[1546, 9337]
1.8
[0.8, 3.9]
5.0
[1.0, 23.0]
63
Table 12 (cont.)
Study
(cont.)
(Kennedy et
al 2006)
Effectiveness of universal neonatal hearing screening for secondary outcomes (controlled studies)
Level of
evidence
III-2
Study
quality a
QS=5/6
Setting
4 hospitals
in Wessex,
UK
4 hospitals
in Greater
London
Population
68714
infants born
with UNHS
88,019
infants born
without
UNHS
168 PCHI
children
identified
overall
Screen fail
criterion
Wessex
Subgroup:
as above
Greater
London
Subgroup
Fail criteria
not stated
ABR fail
≥40 dB HL
Outcomes
Critical
appraisalb
UNHS
rate (%)
Control
rate (%)
Number
needed to
screen (NNS)/
diagnose
(NND)c
Relative
risk
[95% CI]
Adjusted
odds ratiod
[95% CI]
Management
<10 months
SP = NS
Clin I = 2/4
R = 1/5
Power = 41%
0.06
0.02
NNS
2,965
[1458, 86207]
2.4
[1.0, 5.8]
8.0
[1.2, 51.0]
Confirm-ation < 9 mths
Clin I = 1/4
R = 2/5
0.06
0.02
NNS
2,500
[1667, 5000]
3.3
[1.8, 5.8]
120 included
(UNHS=61;
NonUNHS=59)
64
Universal neonatal hearing screening
Table 12 (cont)
Study
(cont.)
Effectiveness of universal neonatal hearing screening for secondary outcomes (controlled studies)
Level of
evidence
Study
quality a
Setting
Population
Screen fail
criterion
Outcomes
Critical
appraisalb
UNHS
rate (%)
Control
rate (%)
Number
needed to
screen
(NNS)/
diagnose
(NND)c
Relative
risk
[95% CI]
SP = good
Clin I = 1/4
R = 2/5
Power = 92%
50.0
9.8
NND
3
[2, 5]
5.1
[2.1, 12.4]
Mean
4.5 mths
Hessen
Mean
25.7 mths
Median
17.8 mths
Thuringia
Median
52.0 mths
-
-
Adjusted
odds ratiod
[95% CI]
2-stage: TEOAE–TEOAE – dx audiology
(Nekahm et
al 2001a)
III-2
QS = 5.5/6
Tyrol,
Austria
91 Tyrolean
children with
PCHI, born
1990–1999
UNHS =
1995–1999
Not UNHS =
1990–1994
Study
duration: 10
years,
retrospective
Not stated
Diagnosis <6
months
(Neumann et
al 2006)
III-2
QS=5/6
46
maternity
clinics and
3 NICU’s
17, 349 well
and at-risk
babies in
2005 UNHS
group
98 Hessian
and 355
German
PCHI in nonUNHS group
TEOAE ≥
30dB
AABR ≥
35dB
Age at
diagnosis
Hessen,
Germany
and
Thuringia,
Germany
UNHS=
1995-2005
Not UNHS=
1990-2005
Universal neonatal hearing screening
Median
3.1 mths
-
Germany
Mean
39.0 mths
Median
33.0 mths
65
Table 12 (cont)
Study
(cont.)
Effectiveness of universal neonatal hearing screening for secondary outcomes (controlled studies)
Level of
evidence
Study
quality a
Setting
Population
Screen fail
criterion
UNHS
rate (%)
Control
rate (%)
Number
needed to
screen
(NNS)/
diagnose
(NND)c
Relative
risk
[95% CI]
Adjusted
odds
ratiod
[95% CI]
Mean
4.8 mths
Median
3.5 mths
Mean
29.1
Median
21.0 mths
-
-
-
Mean
9.7 mths
Mean
46 mths
NND
3 months (%)
35
2
6 months( %)
69
6
12 months (%)
81
12
4
[3, 4]
2
[2, 2]
2
[2, 2]
Outcomes
Critical
appraisalb
Age at
intervention
(Weichbold
et al 2006)
III-2
QS=4.5/6
Innsbruck,
Austria
321 children
with PCHI,
born 19902003
Not stated
Age at
confirmation
Age at
management
6 months (%)
12 months (%)
66
Clin I =1/4
R = 2/5
Clin I =1/4
R = 2/5
17.5
[5.7, 55.7]
11.5
[6.2, 22.4]
6.8
[4.5, 10.7]
NND
61
4
76
9
2
[2, 3]
2
[2, 2]
15.3
[7.1, 34.7]
8.4
[5.0, 13.9]
Universal neonatal hearing screening
Table 12 (cont)
Study
(cont.)
Effectiveness of universal neonatal hearing screening for secondary outcomes (controlled studies)
Level of
evidence
Study
quality a
Setting
Population
Screen fail
criterion
Outcomes
Critical
appraisalb
UNHS
rate (%)
Control
rate (%)
Number
needed to
screen
(NNS)/
diagnose
(NND)c
Relative
risk
[95% CI]
Colorado,
USA
25 matched
pairs of
children with
bilateral
PCHI born in
hospitals
with and
without
UNHS
Study
duration: 5
years,
retrospective
Bilateral fail
≥35 dB
Diagnosis <6
months
SP = excellent
Clin I = 1/4
R = 2/5
Power = 100%
84.0
8.0
NND
1
[1, 2]
10.5
[2.7, 40.1]
Adjusted
odds
ratiod
[95% CI]
1-stage: AABR – dx audiology
(YoshinagaItano et al
2001)
III-2
QS = 5.5/6
UNHS = universal neonatal hearing screening; TEOAE = transient otoacoustic emissions test; AABR = automated auditory brainstem response test; NICU=neonatal intensive care unit; dx = diagnostic; PCHI = permanent
childhood hearing impairment; NS = not significant at p<0.05. a This is a quality score (QS) derived from NHMRC (2000) checklist (see Appendix E) with a high score indicating good quality; b determination of statistical
precision (SP), rank scores for the clinical importance (Clin I) of the benefit/harm (with 1 ranked as highly clinically important and 4 ranked as clinically unimportant), and rank scores for the relevance (R) of the evidence
(with 1 ranked as a highly relevant outcome and 5 as an unproven surrogate outcome); c NNS=number needed to screen-= number of infants who need to be screened in order to have one infant be referred/ diagnosed
or managed earlier, NND= number needed to diagnose= number of infants with PCHI who would need to be screened, as opposed to not screened, in order to have one infant diagnosed with PCHI before the age of 6
months; d adjusted for severity of PCHI – logistic regression results reported by the author; e referral to an audiologist for diagnostic assessment; fpatients who were underwent distraction test not UNHS, g children
confirmed with PCHI after diagnostic assessment – these data not complete until youngest children (18 months old) are approximately 5 years of age;
Universal neonatal hearing screening
67
Coverage
The largest quasi-randomised screening trial to date – the Wessex Controlled Trial –
attempted to screen 25,609 infants using a two-stage screening protocol with program
coverage of 83 per cent (Kennedy et al 1998). The largest uncontrolled screening
program in Rhode Island, USA, screened 99 per cent of 53,121 infants using a different
two-stage protocol (Vohr et al 1998). The latter program’s excellent coverage is likely to
have been assisted by state legislation mandating universal neonatal hearing screening
(UNHS). Bamford and colleagues (2005) also reported on the screened population of 23
areas of England, reaching 97.5 per cent coverage, however, while it is stated that there
are about 120, 000 births per annum, the exact number of screened infants is unclear
(Bamford et al 2005). Coverage does not appear to vary significantly according to either
the size of the source population being screened or the screening protocol used. A large
proportion of hearing screening programs that reported on coverage managed to screen
over 90 per cent of infants in their catchment area.
The lowest coverage of 45.0 per cent was seen in a private rural hospital in Gauteng,
South Africa. This study was conducted over a 4 year period, during which time the
hearing screening service was subsidised for 22 months as part of the hospital birthing
package. This resulted in a coverage of 75 per cent. The subsequent 26 month period,
when the service was no longer subsidised, achieved coverage of only 20 per cent, giving
an overall rate of 45 per cent. Nonetheless, there were several other studies (k=7) that
also achieved coverage rates less than 80 per cent. Reasons for this included initial
screens occurring between and one and two months after the baby was born. It is
possible that there were accessibility issues and/or that resources available for these
programs to ensure the ‘capture’ of infants at this age for outpatient or health clinic
appointments for a stand-alone screen were limited. Additionally, the majority of these
studies were located in rural areas or small clinics.
Referrals and false alarms
The highest level of evidence available (Kennedy et al 1998; Kennedy et al 2006)
indicated that infants who receive universal neonatal hearing screening are nearly three
times more likely [RR = 2.9, 95%CI 1.4, 6.3] to be referred for diagnostic testing within 6
months than infants who are not screened universally. When adjusted for the severity of
hearing impairment, the odds of referral improve up to 19 times [OR = 19.0, 95%CI 3.2,
111.0]. In practical terms this means that 1,619 infants would need to be universally
screened for hearing impairment, as compared to not screening, to ensure the referral for
diagnostic testing of one infant under the age of 6 months (Table 12).
There was considerable variation (1.4–30.3%) in the referral rates from initial transient
evoked otoacoustic emissions (TEOAE) screening in the uncontrolled studies Table 13,
Table 14 and Table 15). No systematic differences could be ascertained according to the
population screened (ie well babies or well and ‘at-risk’ babies) or to the time of
screening. Bilateral fail screening criteria contributed to lower initial referral rates (1.1–
12.8%), although some studies that referred according to unilateral fail criteria also
managed low referral rates. Most of the variation in TEOAE referral rates can be
attributed to the different definitions of a TEOAE fail (ie different emission response
levels at different numbers and types of frequency bands, with different reproducibility
thresholds), or were possibly a consequence of the different environmental conditions
(ambient noise levels) under which the testing was conducted.
68
Universal neonatal hearing screening
There was considerably less variation (1.2-13.0%) in the referral rates from initial
automated auditory brainstem response (AABR) screens. Rates did not vary substantially
according to the population screened or to the screen fail criteria used Table 13, Table
14, and Table 15). AABR uses a standard referral threshold of >35 db HL, which may
have contributed to the homogeneity of the results.
Diagnostic referrals in uncontrolled studies again varied widely as a consequence of
TEOAE testing. After a one-stage TEOAE screen, nearly one-quarter (20.4%) of infants
screened were referred for diagnostic testing (Table 13) (Hahn et al 1999). Ng (2004)
used a 1-stage DPOAE and found 3.4 per cent of infants were referred for diagnostic
testing. The two-stage TEOAE screening resulted in 2–46 per cent of first screen failures
being referred for diagnostic testing (Table 12). The majority of these studies obtained
less than a 34 per cent referral rate with one outlier to this, specifically the second
strategy conducted by Kolski et al (2007) which found a referral rate of 46.2 per cent.
This strategy was based on first screens taking place when the newborn was two months
old and re-screens taking place at 3 months old. The authors believed that this did not
have an impact on the referral rates obtained, however it is unclear why the results are
disparate. Three-stage TEOAE screening resulted in 8 per cent of second screen failures
receiving referrals for diagnostic testing (Table 15).
One-stage AABR screening resulted in consistently low referral rates (4.5–5.0%) for
diagnostic testing (Table 13). After two-stage AABR screening, 14.6-59.2 per cent of
initial screen failures were referred for diagnostic testing. The high referral rate reported
in the study by Iwasaki et al. (2003) may be explained by the different population
assessed, ie Japanese infants and also it was unclear how many were well babies or at-risk.
After three-stage AABR screening one study reported nearly one-quarter (24.2%) of
second screen failures being referred for diagnostic testing Table 15.
Studies using a combination of different screening tests reported the lowest consistent
referral rates for one- and two-stage screen failures (0.6–7.7%) (Table 13, Table 14). A
three-stage mixed screening protocol resulted in 16.5-45.5 per cent of screen failures
being referred (Table 15).
As indicated by the diagnostic accuracy section of this report, TEOAE screening has a
high false alarm rate. Incorrect (positive) initial screens were determined in 16–97 per
cent of TEOAE failures. In general, false alarms were seen in approximately half of the
infants receiving an initial TEOAE screen. These were infants who received unnecessary
re-screening. Similarly, 0-93 per cent of those referred for diagnostic testing from a
TEOAE screen were found to have normal hearing – including one study that reported
no diagnostic false alarms (Owen et al 2001). Overall, the high TEOAE referral rate and
the large proportion of false alarms meant that a considerable number of infants
screened were incorrectly identified as having possible hearing impairment. With a
TEOAE screen protocol, unnecessary re-screening (ie false alarm) occurred in 1–11.5 per
cent of all infants screened, whilst unnecessary diagnostic testing occurred in 0–10 per
cent.
False alarms caused through an AABR screening protocol, and requiring re-screening,
occurred in 41–100 per cent of initial screen failures – although in only 0.7–7 per cent of
the total infants screened. Unnecessary diagnostic testing occurred in 0–100 per cent of
AABR screen failures, which equates to 0–5 per cent of the total population screened. In
Universal neonatal hearing screening
69
terms of the number of false alarms, AABR screen protocols performed better than
TEOAE screen protocols.
70
Universal neonatal hearing screening
Table 13
Study
Descriptive (uncontrolled) studies of 1-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila
criterion
Coverage
(%)
Stage 1
Failure
rate
(%)
LTFUb
(%)
FAc
rate
(%)
Yieldd
(/1000)
1-stage: TEOAE – dx audiology e
(Hahn et
al 1999)
Hospital
388 well
babies
Not
stated
Unilateral
fail if
TEOAE did
not have
≥60%
reproducibility and
≥80%
stimulus
stability
-
20.4
53.2
46.8
0.0
5 months
Unilateral
or bilateral
failure >40
dB HL
98.9
3.4
-
56.8
5.6
Münster,
Germany
Level IV
QS=2/5
1-stage: DPOAE-dx audiology (3 re-screens)
(Ng et al
2004)
Tsan Yuk
hospital
Hong Kong
Level IV
QS=4/5
1064
infants born
between
May October
1999
1-stage: AABR – dx audiology
(Downs
1995)
17
hospitals in
Colorado,
USA
Level IV
QS=2/5
14,494 well
and at-risk
babies
Not
stated
Bilateral or
unilateral
fail ≥35 dB
HL
-
5.0
0.0
93.5
2.6
(Iley &
Addis
2000)
York district
hospital
48 babies
4 days
Unilateral
AABR fail
≥35 dB HL
91.7
4.5
0.0
100.0f
0.0
North
Yorkshire,
UK
Level IV
QS=3.5/5
Universal neonatal hearing screening
Diagnostic
fail ≥40 dB
71
Table 13 (cont.)
Study
Descriptive (uncontrolled) studies of 1-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila
criterion
Coverage
(%)
Stage 1
Yieldd
(/1000)
Failure
rate
(%)
LTFUb
(%)
FAc
rate
(%)
-
3.6
0.0
100.0
0.0
66.6
-
-
-
22.1
1-stage: TEOAE+AABR – dx audiology e
(Hahn et
al 1999)
Hospital
55 well
babies
Not
stated
Münster,
Germany
Level IV
QS=2/5
Unilateral
fail if
TEOAE did
not have
≥60%
reproducibility and
≥80%
stimulus
stability
Unilateral
AABR fail:
>35 dB HL
(Chiong et
al 2007)
Several
communities
in a rural
area,
Bulacan
province,
Philippines
724 babies
2 years
10
months
Fail not
stated
Level IV
QS=2/5
TEOAE = transient evoked otoacoustic emissions test; AABR = automated auditory brainstem response test; dx = diagnostic. a Definition of the
fail is provided in the study profiles in Appendix F; b loss to follow-up of failures after the screen; c false alarm rate – calculated as the number of
infants falsely identified with hearing impairment divided by the total number of infants testing positive for hearing impairment ; d children with
PCHI after diagnostic assessment – transient conductive hearing losses excluded where possible; e includes diagnostic ABR or any of the other
diagnostic tests performed separately or as a test battery; f both cases were actually mild losses that fell between the screening and diagnostic
hearing loss thresholds.
72
Universal neonatal hearing screening
Table 14
Study
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
48.3
47.9
3.8
0.0
77.8
Yielde
(/1000)
2-stage: TEOAE–TEOAE – dx audiology f
(Aidan et al
1999)
Hospital
1,727 well and
at-risk babies
18 months
Bilateral or
unilateral fail
>40 dB HL
82.3
16.7
1,492 well
babies
4 years
Bilateral or
unilateral fail
>40 dB HL
-
19.7
-
319 well and atrisk babies
1 year
Bilateral or
unilateral fail
>40 dB HL
75.0
18.3
-
4,196 well and
at-risk babies
3 years
Failg if absent
TEOAE for ≥2
of 4 frequency
bands
90.6
1.7
Paris, France
Level IV
QS = 5/5
(Bantock &
Croxson
1998)
Hospital
London, UK
Level IV
QS = 4/5
Communitybased health
centre (~7% of
population)
44.6
-
-
1.4
-
-
0.7
2.3
-
-
0.0
34.2
0.0
60.0
2.4
London, UK
(Chapchap &
Segre 2001)
Hospital
Israelita Albert
Einstein
Sao Paulo,
Brazil
Level IV
QS = 3.5/5
Universal neonatal hearing screening
17.8
47.9
73
Table 14 (cont.)
Study
(Daemers et
al 1996)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
St Augustinus
Hospital
Population
Study
duration
907 well babies
born in 1993
and 1994
2 years
Bilateral or
unilateral fail if
TEOAE absent
for 4 frequency
bands
532 well and atrisk infants
(448 well
babies; 84 atrisk)
Not stated
Fail if TEOAE
reproducibility
<50% with n<3
frequencies
with intensity >3
dB SPL (sound
pressure level)
2,012 well and
at-risk babies
born in 1999
1 year
Bilateral fail if
TEOAE absent
at 3 of 4
frequencies
Antwerp,
Belgium
Level IV
QS = 3/5
(De Capua et
al 2003)
University of
Siena
Italy
Level IV
QS=4.5/5
(Govaerts et
al 2001)
St Augustinus
Hospital
Antwerp,
Belgium
Level IV
QS = 4/5
(Habib &
Abdelgaffar
2005)
Dr. Soliman
Fakeeh
Hospital
Jeddah, Saudi
Arabia
Level IV
QS=4/5
74
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Yielde
(/1000)
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
18.9
25.1
56.1
18.7
59.4
34.4
1.1
11.7
-
-
20.9
-
-
5.6
99.4
1.4
0.0
78.6
21.4
0.0
33.3
2.0
91.7
8.7
0.0
71.2
28.8
-
92.7
1.8
-
-
Dx ABR
bilateral fail >40
dB
11, 986 nonhigh-risk
neonates
8 years
Failure to
produce more
than 50%
reproducibility
and response
amplitude at
least 1dB SPL
per octave
Universal neonatal hearing screening
Table 14 (cont.)
Study
(Hatzopoulos
et al 2007)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Main maternity
hospital of
Tirana
Tirana,
Albania
Population
Study
duration
Air Force
community
hospital
Temple
University
Hospital, North
Philadelphia
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
1 year
Fail to produce
≥ 70%
reproducibility,
and a ≥6 dB
SNR (signal-tonoise ratio) for
at least 2 out of
5 frequency
bands
-
13.6
50.0
42.3
7.7
12.5
75.0
1.3
639 well babies
6 months
Bilateral fail at
first screen /
unilateral fail at
re-screen if
TEOAE <80%
reproducib-ility
at 2.4, 3.2, 4.0
kHz
98.1
9.6
0.0
81.7
18.3
0.0
45.5
1.6
2,137 well and
at-risk babies
1 year
Bilateral or
unilateral fail if
TEOAE absent
at signal-tonoise ratio ≥3
dB at 3 of 4
frequency
bands
95.0
8.2
36.5
Maryland,
USA
Level IV
QS=5/5
(Isaacson
2000)
Coverage
(%)
1,561 well and
at-risk babies
(463 well &
1,098 NICU
babies)
Level IV
QS=3.5/5
(Huynh et al
1996)
Faila criterion
Pennsylvania
USA
Level IV
QS=3.5/5
Universal neonatal hearing screening
-
-
-
-
6.9
75
Table 14 (cont.)
Study
(Jakubikova
et al 2003)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Two
gynaecology
and
neonatology
departments in
Bratislava
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
3048 newborns
(approx 45%
at-risk)
Not stated
Fail criteria not
stated
100
4.9
19.3
60.7
20.0
0.0
23.3
6.9
2,537 well and
at-risk babies
14 months
Unilateral or
bilateral failure
to produce
emission
spectrum of
significant gain
across testing
frequency
range >30 dB
HL
90.2
8.9
12.3
64.7
23.0
8.5
80.9
0.4
Department of
Pathological
Newborn and
Intensive Care
Unit of
Children’s
University,
Bratislava
Slovak
Republic
Level IV
QS=3.5/5
(Kanne et al
1999)
Madigan Army
Medical
Centre,
Tacoma
Washington
USA
Level IV
QS=4.5/5
76
Universal neonatal hearing screening
Table 14 (cont.)
Study
(Khairi et al
2005)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Hospital
Universiti
Sains
Malaysia
Population
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
401 newborns
(202 well
babies, 199 atrisk)
Study
duration:
Feb-March
2000 &
Feb-May
2001
Fail not stated
-
7.7
29.0
54.8
16.1
0.0
20.0
10.0
32,080 live
births (well and
at-risk) in 2003
1 year
Not stated
66.6
10.7
-
88.5
11.5
20.0
32.7
1.2
Maternity
hospital
Strategy 1:
3202 newborns
Not stated
Bilateral fail
Strategy 1:
95.7
1.1
5.9
64.7
29.4
10.0
50.0
1.3
France
Level IV
QS=3.5/5
Strategy 2:
2588 2 month
old babies
3.1
36.5
17.3
46.2
4.2
79.2
Level IV
QS=4/5
(Khandekar
et al 2006)
Study
duration
Hospitals in
Oman, Turkey
Level IV
QS=4.5/5
(Kolski et al
2007)
Strategy 2:
64.2
Study duration:
not stated
Universal neonatal hearing screening
77
Table 14 (cont.)
Study
(Lin et al
2004)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
2 hospitals & 4
obstetric
clinics
Tainan,
Taiwan
Population
5,938 neonates
Study
duration
2 years 9
months
Failure to
produce 4 pairs
of alternating
positive &
negative peaks
Coverage
(%)
Stage 1
Stage 2
Yielde
(/1000)
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
59.3
9.0
26.1
56.8
17.0
-
-
1.5
Diagnostic
ABR:
Failure to
produce wave V
latency within
developmental
norms in
response to
35dB nHL clicks
Level IV
QS=4/5
(Low et al
2005)h
Singapore
General
Hospital,
Singapore
Level IV
QS=4/5
2,973 well and
at-risk babies
2 years
Fail not stated
99.7
-
-
-
2.4
16.7
40.0
1.2
(Martines et
al 2007)
Sciacca
Hospital
1,068 well and
at-risk babies
born during
2003-2004
942 well babies
126 at-risk
babies
2 years
Fail criterion not
stated
89.7
At-risk
11.1
At-risk
0.0
At-risk
57.1
At-risk
42.9
At-risk
0.0
At-risk
-
1.9
Italy
Level IV
QS=3.5/5
78
Faila criterion
Universal neonatal hearing screening
Table 14 (cont.)
Study
(McPherson
et al 1998)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Eight
community
health clinics,
Northern
Brisbane
Population
1,305 children,
1.5–2.5 months
of age
Study
duration
30 months
Local health
centres and
homes in
urban and
rural settings
Maternity
Hospital
Stara Zagora
Bulgaria
Level IV
QS=4/5
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
93.3
10.8
42.7
24.4
32.8
27.9
20.9
2.5
683 well babies 1 year
registered at
participating
health centres in
1999
Bilateral or
unilateral fail if
TEOAE
response <28
dB, or response
correlation
<98%, or
signal-to-noise
ratio did not
reach target
level for 3
wavebands
98.8
4.3
0.0
96.6
3.4
0.0
0.0
1.5
1,750 well and
at-risk babies
Bilateral fail >30
dB HL
95.5
12.8
0.0
89.7
10.3
0.0
86.4
1.8
West
Gloucestershire, UK
Level IV
QS=3.5/5
(Rouev et al
2004)
Coverage
(%)
Unilateral or
bilateral fail if
TEOAE <3 dB
above the noise
floor and at
least halfway
across the test
frequency
bands of 2–3
kHz and 3–4
kHz
Queensland,
Australia
Level IV
QS=2/5
(Owen et al
2001)
Faila criterion
Universal neonatal hearing screening
329 days
79
Table 14 (cont.)
Study
(Swanepoel
et al 2007)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Private
hospital in
urban Gauteng
South Africa
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
6,241 well and
at-risk babies
4 years
Bilateral and
unilateral
sensorineural
hearing loss of
≥ 35dB
45.0
11.1
-
31.5
-
-
-
1.0
711 well and atrisk babies
(475 well
babies, 236
NICU babies)
18 months
Unilateral fail
-
3.9
21.4
60.7
17.9
0.0
40.0
4.2
Level IV
QS=4/5
(Tatli et al
2007)
Dokuz Eylul
University
Hospital, Izmir
Turkey
Level IV
QS=4.5/5
(Tsuchiya et
al 2006)
Kumamoto
University
Hospital
Kumamoto
Japan
Level IV
QS=4/5
8979 well and
at-risk babies
born during
1999-2004
5 years
Fail ≥ 35dB
64.8
0.4
0.0
16.2
83.8
3.2
16.1
2.9
(Vohr et al
1998)
8 maternity
hospitals
53,121 well and
at-risk babies
born 1993–
1996
4 years
Unilateral or
bilateral fail if no
TEOAE
response at 2–4
kHz with 75%
reproducib-ility
99.1
10.2
15.2
72.2
12.5
3.0
80.6
2.1
Rhode Island,
USA
Level IV
QS=5/5
80
ABR fail if wave
V not present
>30 dB nHL
Universal neonatal hearing screening
Table 14 (cont.)
Study
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Failure
rate (%)
(Watkin &
Baldwin
1999)
Whipps Cross
hospital
Rijeka
University
Hospital
Yielde
(/1000)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
-
-
11.1
65.6
1.7
28,890 babies
6 years
Bilateral fail on
initial screen
and re-screen,
although
unilateral fail
allowed if
obvious
parental anxiety
ABR fail >40dB
nHL in better
hearing ear
87.2
6,019 neonates
26 months
Unilateral or
bilateral fail on
first screen
98.8
6.0
6.6
69.8
23.6
13.9
69.8
2.3
18 months
Unilaterali or
bilateral fail ≥35
dB HL
96.2
2.7
8.5
73.5
3.7
0.0
47.8
0.9
London, UK
Level IV
QS=4/5
(Zaputovic et
al 2005)
LTFUb
(%)
Stage 2
Study duration:
26 months
-
-
Croatia
Level IV
QS=3.5/5
2-stage: TEOAE–AABR – dx audiology
(Bailey et al
2002)
5 maternity
hospitals
13,214 well and
at-risk babies
Perth,
Australia
Level IV
QS=5/5
Universal neonatal hearing screening
81
Table 14 (cont.)
Study
(Bamford et
al 2005)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
23 areas of
England
Level IV
QS=4/5
(Brennan
2004)
Instititution in
Illinois
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
About 120,000
births per
annum (well
and at-risk)
2 years
Not stated
97.5
-
-
-
-
-
-
Bilateral
1/1000
Unilateral
0.64/1000
Not stated
Not stated
Not stated
>98.0
-
-
-
3.0
-
-
-
217 well and atrisk babies
4 weeks
Unilateral or
bilateral AABR
fail ≥35 dB HL
93.0
30.3
13.1
85.2
1.6
0.0
100.0
0.0
33, 873 well
and at-risk
babies born
from January
2004 to March
2006
26 months
Bilateral fail
≥35 dB HL -40
db HL
depending on
screening
device used
(could not be
modified)
92.4
1.3
2.5
89.8
7.7
0.0
14.7
0.8
USA
Level IV
QS=2.5/5
(Hunter et al
1994a)
Princess Anne
Hospital
Southamp-ton,
UK
Level IV
QS=5/5
(Leveque et
al 2007)
17 maternity
wards, private
and public,
and one NICU
in
ChampagneArdenne
France
Level IV
QS=4.5/5
82
Universal neonatal hearing screening
Table 14 (cont.)
Study
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
(Low et al
2005)
National
University
Hospital,
Singapore
Level IV
QS=4/5
4,849 well and
at-risk babies
2 years
Fail not stated
99.8
-
-
-
0.6
3.2
12.9
5.4
(Molini et al
2004)
Hospital
2,425 full-term
newborns (well
and at-risk
babies)
Study 1: 17
months
Study 2: 9
months
Failure to
produce at least
3 of the 4
frequency
bands centred
at 1600, 2400,
3200 and 4000
Hz.
Study 1
94.2
Study 1:
8.9
Study 1:
17.4
Study 1:
77.5
Study 1:
5.1
Study 1:
0.0
Study 1:
0.0
Study 1 & 2:
5.0
Study 2:
72.4
Study 2:
10.1
Study 2:
20.2
Study 2:
74.2
Study 2:
5.6
Study 2:
0.0
Study 2:
0.0
17, 349 well
and at-risk
babies in 2005
1 year
TEOAE
≥ 30dB HL
AABR
≥ 35dB HL
95.2
2-stage
7.1
-
-
-
-
-
Italy
Level IV
QS=4/5
(Neumann et
al 2006)
Germany
Level IV
QS=4/5
2.8
1 stage
AABR
1.9
1 stage
TEOAE
5.2
Total:
3.0
Universal neonatal hearing screening
83
Table 14 (cont.)
Study
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Population
Study
duration
Faila criterion
Coverage
(%)
Stage 1
Stage 2
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
Yielde
(/1000)
2-stage: DPOAE–AABR – dx audiology
(Cox & Toro
2001)
Boston
Medical
Centre
1,713 well and
at-risk j babies
1 year
Fail not statedg
-
-
-
-
-
13.3
36.7
2.9
4,437
newborns
(315 NICU,
4122 nonNICU)
11 months
Fail not stated
84.6
18.7
43.0
50.3
6.7
38.1
32.7
3.6
1 year
Fail >35 dB HL
99.5
2.1
17.4
68.0
14.6
6.7
33.3
1.8
Massachusetts, USA
Level IV
QS=5/5
(Mukari et al
2006)
Hospital
University
Kebangsaan
Malaysia
(HUKM)
Level IV
QS=4/5
2-stage: AABR–AABR – dx audiology
(Clemens et
al 2000)
84
Women’s
Hospital
North
Carolina, USA
Level IV
QS=3/5
5,034 well
babies k
Universal neonatal hearing screening
Table 14 (cont.)
Study
(Connolly et
al 2005)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
University of
Mississippi
Population
17,602 well and
at-risk babies
Study
duration
5 years
Mississippi,
USA
SeireiHamamatsu
General
Hospital &
SeireiMikatahara
General
Hospital
Hamamatsu,
Japan
Fail
>35 dB nHL
Coverage
(%)
Stage 1
Stage 2
Yielde
(/1000)
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
100
-
-
-
-
-
-
4.4
Several rescreens at
stage 1 and
stage 2
Level IV
QS=4.5/5
(Iwasaki et al
2003)
Faila criterion
4092 infants
2 years
Fail if likelihood
ratio is less
than 160 after
15000 sweeps
99.8
1.2
0.0
40.8
59.2
0.0
48.3
3.7
28,273 well and
at-risk babies
2 years
Fail not stated
99.8
-
-
-
-
19.6
8.5
3.0
Level IV
QS=4.5/5
(Low et al
2005)
KK Women’s
and Children’s
Hospital,
Singapore
Level IV
QS=4/5
Universal neonatal hearing screening
85
Table 14 (cont.)
Study
(Mason &
Herrmann
1998)
Descriptive (uncontrolled) studies of 2-stage universal neonatal hearing screening
Setting
Medical
Center
Population
10,773 well and
at-risk babies
Study
duration
5 years
Honolulu,
Hawaii
Level IV
QS=4/5
(OudesluysMurphy &
Harlaar 1997)
Faila criterion
Bilateral AABR
fail ≥35 dB HL
Coverage
(%)
Stage 1
Stage 2
Yielde
(/1000)
Failure
rate (%)
LTFUb
(%)
FAc rate
(%)
Failure
rated (%)
LTFU
(%)
FA rate
(%)
96.3
4.0
-
-
-
-
-
1.4
29.7
51.4
18.9
14.3
57.1
3.5
Diagnostic fail
>35 dB nHL
Community well 288 well and atrisk babies
baby clinic,
Barendrecht
1 year
Unilateral or
bilateral fail >35
dB HL
98.6
13.0
217 well babies
born in 1999
and 2000
1 year
Not stated
78.8
-
-
-
-
-
-
5.8
3,066 well and
at-risk babies
2 years
ABR unilateral
or bilateral fail
>35-40 dB HL
99.9
-
-
-
-
-
-
2.0
Netherlands
Level IV
QS=5/5
(Rao et al
2002)
5 small rural
hospitals,
central
Minnesota
USA
Level IV
QS=2.5/5
(Yee-Arellano
et al 2006)
Private
hospital in San
Pedro Garza
Garcia
Mexico
Level IV
QS=4/5
86
Universal neonatal hearing screening
TEOAE = transient evoked otoacoustic emissions test; AABR = automated auditory brainstem response test; dx = diagnostic; DPOAE = distortion product otoacoustic emissions test; SPL=sound
pressure level; NICU=neonatal intensive care unit; a Definition of the fail is provided in the study profiles in Appendix F; b loss to follow-up of failures after the screen; c false alarm rate – calculated as
the number of infants falsely identified with hearing impairment divided by the total number of infants testing positive for hearing impairment ; d referral to an audiologist for diagnostic assessment; e
children with PCHI after diagnostic assessment – transient conductive hearing losses excluded where possible; f includes diagnostic ABR or any of the other diagnostic tests performed separately or
as a test battery; g bilateral or unilateral fail not stated; h the type of OAE is not clearly specified; i bilateral fail in 40 babies, until criteria changed; j all NICU babies received a 1-stage AABR screen
followed by diagnostic referral; k approximately half of the babies that failed the initial AABR received a re-screen AABR in stage 1.
Universal neonatal hearing screening
87
Table 15
Study
Descriptive (uncontrolled) studies of 3-stage (or more) universal neonatal hearing screening
Setting
Population
Study
duration
Faila
criterion
Coverage
(%)
Stage 1
Stage 2
Stage 3
Yielde
(/1000)
Failure
rate (%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFU
(%)
FA
rate
(%)
Failure
rated
(%)
LTFU
(%)
FA
rate
(%)
0.0
12.5
2.2
3-stage: AABR–AABR–AABR – dx audiology f
(Clemens
& Davis
2001)
Women’s
Hospital
North
Carolina,
USA
Level IV
QS=5/5
3,144 well
babies
6 months
Fail >35 dB
HLg
99.9
4.2
4.6
70.2
25.2
0.0
75.8
24.2
Unilateral
or bilateral
fail >35 dB
HL
91.0
11.2
28.2
47.1
24.7
48.8
31.3
20.0
-
-
1.6
26.9
18.9
57.5
23.5
-
-
-
27.0
31.3
5.9
3-stage: AABR–AABR–TEOAE – dx audiology
(Messner
et al 2001)
Lucile
Packard
Children’s
Hospital,
Stanford
6,340 well
babies born
1998–1999
16
months
California,
USA
Level IV
QS=5/5
3-stage: TEOAE–ABR–TEOAE + ABR – dx audiology
(Clarkson
et al 1994)
(Maxon et
al 1993)
88
Women
and Infants
Hospital,
Providence
Rhode
Island,
USA
Level IV
QS=2.5/5
1,850 well
and at-risk
babies born
1990–1991
6 months
Fail ≥60 dB
HL referred
for dx ABR;
<60 dB HL
referred for
behavioural
audiologic
evaluationg
-
Universal neonatal hearing screening
Table 15 (cont.)
Study
(cont)
Setting
Descriptive (uncontrolled) studies of 3-stage (or more) universal neonatal hearing screening
Population
Study
duration
Faila
criterion
Coverage
(%)
Stage 1
Stage 2
Stage 3
Failure
rate (%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFU
(%)
FA
rate
(%)
Failure
rated
(%)
LTFU
(%)
FA
rate
(%)
-
-
-
5.8
18.6
73.4
8.0
0.0
0.0
Yielde
(/1000)
3-stage: TEOAE–TEOAE–TEOAE – dx audiology
(Lin et al
2005)
Mackay
Memorial
Hospital
18, 260
well babies
born 19982004
Taipei,
Taiwan
Level IV
QS=4/5
5 years 2
months
Bilateral or
unilateral
failure to
produce
TEOAE of
(1) ≥5 dB
in 3 of 5
frequency
bands or
(2) ≥3 dB
in 4 of 5
frequency
bands
-
4.5
Dx ABR
failure to
produce a
repeatable
wave V at
35 dB nHL
unilaterally
or
bilaterally
Universal neonatal hearing screening
89
Table 15 (cont.)
Study
(cont)
Descriptive (uncontrolled) studies of 3-stage (or more) universal neonatal hearing screening
Setting
Population
Study
duration
Faila criterion
Coverag
e (%)
Stage 1
Stage 2
Stage 3
Failure
rate (%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFU
(%)
FA
rate
(%)
Failure
rated
(%)
LTFU
(%)
FA
rate
(%)
Yielde
(/1000)
3-stage: TEOAE–TEOAE–ABR – dx ABR
(Martines
et al 2007)
Sciacca
Hospital
Italy
Level IV
QS=3.5/5
(Pastorino
et al 2005)
Instituti
Clinici di
Perfeziona
mento
Milan, Italy
Level IV
QS=4/5
90
1,068 well
and at-risk
babies
born during
2003-2004
942 well
babies
126 at-risk
babies
2 years
Fail criterion
not stated
89.7
Wellbabies
4.7
Wellbabies
0.0
Wellbabies
75.0
Wellbabies
25.0
Wellbabies
0.0
Wellbabies
54.5
Wellbabies
45.5
Wellbabies
0.0
Wellbabies
20.0
1.9
19777 well
and at-risk
babies
Not
stated
Screening:
Failure to
produce ≥
70% total
reproducibility
Failure to
produce
≥50% in the
1.6-kHz band
and 70% in
2,4-, 3.2- and
4-kHz bands
Diagnostic
ABR ≥40 dB
HL
-
Well
babies
Well
babies
Well
babies
Well
babies
Well
babies
Well
babies
Well
babies
Well
babies
Well
babies
2.4
12.8
62.8
24.4
15.7
67.8
16.5
0.0
0.0
Well &
at-risk
babies
3.2
3-stage for
well babies
1-stage for
at-risk
babies
however
only
screening
yield was
obtainable
Universal neonatal hearing screening
Table 15 (cont.)
Study
(cont)
Setting
Descriptive (uncontrolled) studies of 3-stage (or more) universal neonatal hearing screening
Population
Study
duration
Faila
criterion
Coverage
(%)
Stage 1
Stage 2
Stage 3
Failure
rate (%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFU
(%)
FA
rate
(%)
Failure
rated
(%)
LTFU
(%)
FA
rate
(%)
6.4
0.0
71.6
28.4
-
-
-
-
-
Yielde
(/1000)
3-stage: TEOAE–AABR–TEOAE– dx ABR
(Lin et al
2005)
Mackay
Memorial
Hospital
Taipei,
Taiwan
3,013 well
babies
born FebDecember
2004
Level IV
QS=4/5
10
months
Bilateral or
unilateral
failure to
produce
TEOAE of
(1) ≥5 dB in
3 of 5
frequency
bands or
(2) ≥3 dB in
4 of 5
frequency
bands
-
3.0
Dx ABR –
failure to
produce a
repeatable
wave V at
35 dB nHL
unilaterally
or
bilaterally
AABR = automated auditory brainstem response test; dx = diagnostic; TEOAE = transient evoked otoacoustic emissions test. a Definition of the fail is provided in the study profiles in Appendix F; b loss to follow-up of failures
after the screen; c false alarm rate – calculated as the number of infants falsely identified with hearing impairment divided by the total number of infants testing positive for hearing impairment ; d referral to an audiologist for
diagnostic assessment; e children with PCHI after diagnostic assessment – transient conductive hearing losses excluded where possible; f includes diagnostic ABR or any of the other diagnostic tests performed separately or
as a test battery; g bilateral or unilateral fail not stated.
Universal neonatal hearing screening
91
Table 15 (cont.)
Descriptive (uncontrolled) studies of 3-stage (or more) universal neonatal hearing screening
Multi-stage: TEOAE–TEOAE–TEOAE-TEOAE (4 to 7 re-screens) – dx audiology
Study
(Shoup
et al
2005)
Setting
Large
public
hospital
Population
48, 211
well and atrisk babies
Study
duration
3 years
Faila
criterion
Fail not
stated
Covera
ge (%)
-
Stage 1
Stage 2
Stage 3
Stage 4
Yielde
(/1000)
Failure
rate
(%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFUb
(%)
FAc
rate
(%)
Failure
rate
(%)
LTFUb
(%)
FAc
rate
(%)
Failur
e rate
(%)
LTFUb
(%)
FAc
rate
(%)
3.6
-
-
34.0
-
-
50.7
-
-
58.2
-
-
3.7
Dallas,
Texas
Level IV
QS=4.5/5
AABR = automated auditory brainstem response test; dx = diagnostic; TEOAE = transient evoked otoacoustic emissions test. a Definition of the fail is provided in the study profiles in Appendix F; b loss to follow-up of failures
after the screen; c false alarm rate – calculated as the number of infants falsely identified with hearing impairment divided by the total number of infants testing positive for hearing impairment ; d referral to an audiologist for
diagnostic assessment; e children with PCHI after diagnostic assessment – transient conductive hearing losses excluded where possible; f includes diagnostic ABR or any of the other diagnostic tests performed separately or
as a test battery; g bilateral or unilateral fail not stated.
.
92
Universal neonatal hearing screening
Loss to follow-up
Losses to follow-up (LTFU) of initial screen failures ranged from 0–53 per cent in the
uncontrolled screening programs that reported on this outcome (Table 13, Table 14,
Table 15). The largest LTFU (36.5–53.2%) after the initial screen occurred in studies
where there was commonly a long delay before re-screening (2 weeks to 2 months
(Aidan et al 1999; Kolski et al 2007; McPherson et al 1998; Mukari et al 2006) or before
diagnostic testing (4 months) in the case of a one-stage screening protocol (Hahn et al
1999). In only one of these studies (Isaacson 2000) was there an immediate re-screen.
This hospital-based program was conducted in an economically depressed inner-city area
and it is likely that the 37 per cent LTFU was related to very early hospital discharge
rates.
The majority of hospital-based studies experienced no LTFU from diagnostic referrals.
However, the largest LTFU from diagnostic referral did occur with a hospital-based
screening program in Belgium (Daemers et al 1996) – 59 per cent of infants referred for
diagnostic testing at 3 months of age were not assessed. The authors attribute this to a
lack of parental and health professional awareness and commitment. It is probable that
the longer it takes for diagnostic testing to occur, the less likely it is that there will be
compliance in attendance unless, for example, reminder and educational strategies are put
in place. Two of the four community-based programs experienced large LTFU – in the
range of 30–43 per cent after the initial screen and 14–28 per cent after diagnostic
referral (McPherson et al 1998; Oudesluys-Murphy & Harlaar 1997). However, two
community-based studies managed no LTFU, in a two-stage screening program (Chiong
et al 2007; Owen et al 2001).
Absolute and incremental yield
Yield is defined as the number of cases of permanent childhood hearing impairment
(PCHI) ultimately identified in the screened population. This relates primarily to
permanent sensorineural or conductive hearing impairment as opposed to transient
conductive losses. The latter were reported in the evidence-base and generally had a
much higher yield. These were not, however, included in the total presented yield figures
as there is insufficient evidence to indicate that transient conductive losses have an
impact on the language and learning outcomes of young children. Conductive losses
were included in the yield figures when they were described as permanent losses.
The highest level of evidence available (Kennedy et al 1998) reported that infants born
during periods of universal neonatal hearing screening (UNHS) are 2.3 times more likely
[RR = 2.3, 95%CI 1.1, 4.7] to receive a diagnosis of PCHI than infants born in periods
without universal hearing screening (Table 12) The absolute increase in benefit is small
(an extra five children identified per 10,000) because of the low prevalence of the
condition. This means that 1,970 infants [95%CI 1,063, 13,459] would need to be
universally screened for hearing impairment, as compared to not screening, to ensure the
diagnosis of one infant with PCHI. It was unclear in this study whether diagnosis before
the age of 10 months occurred more frequently with or without UNHS, as there was a
lack of statistical power for this outcome. Controlling for the severity of infant hearing
impairment resulted in a trend towards a five-fold increase in the odds of diagnosis
(before 10 months of age) during periods of hearing screening.
Universal neonatal hearing screening
93
Three average to good quality retrospective cohort studies (level III-2 screening
evidence) determined that children with PCHI were more likely to be diagnosed before
the age of 6 months when born during periods of, or in hospitals with, UNHS than
children who were not exposed to such screening programs (Table 12) (Nekahm et al
2001a; Weichbold et al 2006; Yoshinaga-Itano et al 2001). The larger of these good
quality studies indicated that diagnosis of bilateral PCHI before the age of 6 months was
5.1 times more likely to occur in children who were born in hospitals with UNHS [RR =
5.1, 95%CI 2.1, 12.4]. This means that for every three children [95%CI 2, 5] with
bilateral PCHI born in an Austrian hospital with a screening program, one additional
child would be diagnosed with PCHI before the age of six months than if born in a
hospital without a screening program (Nekahm et al 2001a). The average quality study
conducted by the same researchers, covering a larger population of Austria, also found
that the early diagnosis of infants at three months of age was nearly 18 times more likely
when born in a hospital with, as opposed to without, universal hearing screening [RR
17.5, 95% CI 5.7, 55.7] (Weichbold et al 2006).
In uncontrolled studies the yield of infants with PCHI from UNHS programs generally
ranged from 1/1,000 to 3/1,000. This yield was fairly consistent no matter the type of
screening protocol used (Table 13, Table 14, Table 15). Exceptionally high yields were
seen in the studies by Chiong et al (2007) and Khairi et al (2005), reporting yields of 22.1
and 10.0/1, 000, respectively. These two studies sampled a non-Caucasian population.
The study by Chiong and colleagues, reporting the highest screening yield, was
conducted in underprivileged rural communities in the Philippines, with poor antenatal
care. It is possible that selection of this population translated into higher rates of ‘at-risk’
babies. Variation in yields in these uncontrolled studies may have occurred due to
differences in LTFU, the proportion of well and ‘at-risk’ babies, and the size of the
population being sampled.
Age at management
The Wessex Controlled Trial provided the highest level of available evidence (level III-1
screening evidence) to assess the effect of universal neonatal hearing screening on the
age at which children with permanent childhood hearing impairment (PCHI) receive
treatment and/or rehabilitation, compared to not screening (Table 12). This study found
a trend that indicated screening may increase the likelihood of PCHI management before
the age of 10 months by 2.4 times [RR = 2.4, 95%CI 1.0, 5.8]. The study was underpowered for this outcome but when the authors controlled for the severity of PCHI, the
odds of early management through screening increased eight-fold [OR = 8.0, 95%CI 1.2,
51.0]. The absolute increase in benefit, over not screening, is early PCHI management
(<10 months of age) for an approximate additional 4 children in 10,000. In real terms
this translates into screening 2,965 children [95%CI 1458, 86207] for hearing
impairment, compared to not screening, to ensure the early management (<10 months of
age) of one infant with PCHI. A retrospective cohort study conducted by Weichbold et
al (2006) (level III-2 screening evidence) also assessed age at management but only had
the population of true hearing impaired children available, thus the number needed to
screen amongst the whole population could not be calculated.
94
Universal neonatal hearing screening
Summary – Does universal neonatal hearing screening, and the finding of a positive
and/or negative test, affect the clinical management or treatment options available to
permanently hearing impaired infants?
Referral for definitive diagnostic testing, actual permanent childhood hearing impairment
(PCHI) diagnosis, and management of PCHI commonly occurs earlier and more frequently
with universal neonatal hearing screening (UNHS) than without it.
Level III-1 screening evidence indicates that referring an infant for diagnostic testing before
the age of 6 months is nearly three times more likely [RR = 2.9, 95%CI 1.4, 6.3] (19 times
when controlling for the severity of hearing impairment) with UNHS than when universal
screening is not available. Infants born during periods of UNHS are twice as likely to receive
a diagnosis of PCHI, than infants born in periods without universal screening [RR = 2.3,
95%CI 1.1, 4.7]. The absolute increase in benefit is small, however – an extra five children
identified per 10,000 – because of the low prevalence of the condition. There is also an
indication that screening may increase the likelihood of PCHI management before the age of
10 months by nearly two-and-a-half times [RR = 2.4, 95%CI 1.0, 5.8] (eight times when
controlling for the severity of PCHI). Similar, but more precise, results were reported in
studies of a lower level of evidence (III-2 screening evidence).
Descriptive data indicates that the majority of UNHS programs manage to screen over 90
per cent of infants in their catchment area. These programs are largely hospital-based, with
initial screening occurring prior to discharge. Community-based studies also obtain very
good coverage when screening is ‘piggy-backed’ onto other health or immunisation checks
at the health clinic or when it occurs in the home. Losses to follow-up commonly occur when
there is a long delay prior to re-screening or diagnostic testing of the infant, or when infants
and mothers are discharged early from the hospital.
Uncontrolled studies suggest that given the higher referral rate from transient evoked
otoacoustic emissions screen protocols, the number of false alarms associated with these
programs is higher (up to approximately 10%) than programs using automated auditory
brainstem response screen protocols (up to approximately 6%). It is, however, possible to
maintain low false alarm rates using either type of screening protocol.
Universal neonatal hearing screening
95
Does universal neonatal hearing screening, and therefore possible
alterations in clinical management, have an impact on the adverse
outcomes associated with permanent childhood hearing impairment?
Data on the longer term impact of a screening program on primary outcomes – rate and
quality of language acquisition, behaviour, family functioning, communication ability /
social functioning, educational achievement, employment status, socioeconomic status
and quality of life – were available in limited form in two studies (Kennedy et al 2006;
Yoshinaga Itano et al 2000). While Yoshinaga-Itano and colleagues conducted a
retrospective matched-pair cohort study, Kennedy et al followed the birth cohort
reported in their 1998 Wessex trial prospectively until schooling age (level III-2 screening
evidence). Both studies compared the quality of language acquisition and communication
ability in children with permanent hearing impairment born in hospitals with and without
universal neonatal hearing screening (UNHS) (Kennedy et al 2006; Yoshinaga Itano et al
2000). Neither of these good quality studies, however, reported on the longer term
outcomes, such as educational achievement and employment status, because the cohorts
in question were not old enough.
Language acquisition
Language is a system representing concepts through the use of words (Owens 1996).
Assessment of language ability incorporates the measurement of several distinct features
of language. For example, receptive language deals with the extent of a child’s
vocabulary, including the number of words understood, and an understanding of
requests, parts of speech and semantic categories. In contrast, expressive language is
more concerned with the way in which language is used, such as the expression and use
of words and parts of speech (White & White 1987). Studies included in this assessment
used a variety of evaluation tools to measure the various aspects of language.
Yoshinaga-Itano and colleagues (2000) assessed the expressive language, receptive
language and total language ability of children with bilateral permanent childhood hearing
impairment (PCHI) and determined whether there were any differences in linguistic
ability according to whether the children were born in hospitals with or without UNHS.
The sample size was small, consisting of 25 matched pairs, and some information on the
outcome measures was incomplete. Similar outcomes were measured in a second cohort
study however it was a much larger study with 120 children which included 61 who
received UNHS and 59 without UNHS (Kennedy et al 2006).
Linguistic ability was measured differently across the two studies. Yoshinaga-Itano and
colleagues measured it using a language quotient derived from the Minnesota Child
Development Inventory. This inventory has apparently been validated using other
special-needs children but has not been validated for a hearing-impaired paediatric
population. Further, the language quotient is based on assessment by the parent or
caregiver of the child – who, of course, was aware of the screening status of the child –
rather than through an independent, blinded assessment. Only those children without
significant cognitive delays (cognitive quotient >70) were included in the analysis. The
children were matched on chronological age, severity of PCHI and cognitive ability, thus
nullifying the effect of confounders known to have an impact on linguistic ability. Other
potential confounders such as gender, ethnicity, other disabilities, mode of
communication and education level of the primary caregiver were not distributed
differently between the screening and not-screening groups.
96
Universal neonatal hearing screening
Children in the screened group were found to have, on average, higher language quotient
scores for expressive language (UNHS = 82.9±18.5 ‡; not UNHS = 62.1±21.5; t[24] =
5.53, p<0.001), receptive language (UNHS = 81.5±18.5; not UNHS = 66.8±20.0; t[24] =
4.21, p<0.001) and total language (UNHS = 82.2±16.5; not UNHS = 64.4±19.5; t[24] =
5.39, p<0.001) than children in the unscreened group. Assuming that a 20 per cent
difference in mean scores on the Minnesota Child Development Inventory is clinically
relevant, the comparative gain in expressive and total language for the screened group
was clinically significant and the gain in receptive language approached clinical
significance.
In terms of expressive vocabulary, as measured by the MacArthur Communicative
Development Inventory, the majority of children in the screened group had larger, more
expressive vocabularies than their matched pairs in the unscreened group (nonparametric
Wilcoxon signed ranks test, Z = 3.72, p<0.001). Again, these results are based on
unblinded assessment.
Kennedy et al (2006) used two different tests for receptive language: the Test for
Reception of Grammar and the British Picture Vocabulary Scale. An aggregate score,
however, was calculated to combine the results obtained. Children in the screened group
were found to have higher scores for receptive language (adjusted mean difference=0.56
[95% CI 0.03, 1.08], p=0.04), however the clinical significance of this difference is
unclear due to the lack of data provided.
For determination of expressive language acquisition, an aggregate score was also
calculated, this time because two components of the Renfrew Bus Story Test were used
(sentence information and 5 longest sentences). This aspect of language ability, however,
did not show a significant difference between those exposed and not exposed to
universal hearing screening (adjusted mean difference 0.30 [95% CI -0.22, 0.81], p=0.25).
Communication ability
Speech is the neuromuscular process of producing sounds for communication. It is the
verbal means of transmission of language (Owens 1996). Language and speech are
distinctly different.
Yoshinaga-Itano and colleagues (2000) assessed the speech of those children with
bilateral permanent childhood hearing impairment (PCHI) born in hospitals with and
without universal neonatal hearing screening (UNHS). Phonological experts observed
video-taped parent–child interactions and evaluated the speech intelligibility and number
of intelligible words produced by the children. The phonological expert was blinded to
the screening status of the child (Yoshinaga-Itano, pers comm 2003), so it is unlikely that
observer bias affected the results. Median speech intelligibility for the screened group
was rated as ‘speech is very hard to understand’, whilst for the matched pairs in the
unscreened group it was rated as ‘always / almost always unintelligible’. The two
distributions of speech intelligibility ratings were significantly different (nonparametric
Wilcoxon signed ranks test, Z = 2.43, p = 0.015).
‡
Mean ± standard deviation, related t-test with degrees of freedom in brackets.
Universal neonatal hearing screening
97
On the other hand, Kennedy and colleagues (2006) measured communication ability by
means of principal caregiver completion of a speech scale from the Children’s
Communication Checklist. This assessment was also blinded, as the researcher
conducting the interview was unaware of the child’s early hearing or audiologic history
and often the caregiver was the mother. This method was distinctly different to that used
by Yoshinaga-Itano and colleagues (2000). Mean z scores for children who had or did
not have universal neonatal hearing screening were calculated using 63 English-speaking
children with normal hearing, matched for place of birth and age at assessment. These
were used to measure unadjusted and adjusted mean differences (based on severity of
hearing impairment and maternal education) between the two groups. There were no
statistically significant differences in measures of speech between those exposed to
universal newborn hearing screening and those not exposed to screening (adjusted mean
difference=0.12 [95% CI -0.46, 0.71], p=0.68). There needs to be consideration that
assessment was based on parental or professional report as opposed to direct infant
measurement and therefore may not be as sensitive as the results from the study by
Yoshinaga-Itano and colleagues.
As UNHS programs were predominantly introduced in the mid- to late-1990s, it is
unlikely that information on the longer term, but highly relevant, outcomes (ie
educational and employment status) will be reported in the peer-reviewed literature for
another decade or so.
98
Universal neonatal hearing screening
Summary – Does universal neonatal hearing screening, and therefore possible
alterations in clinical management, have an impact on the adverse outcomes
associated with permanent childhood hearing impairment?
There is very limited information available on the effect of universal neonatal hearing
screening (UNHS) on primary or patient-relevant outcomes. Two good quality studies (level
III-2 screening evidence) were identified, although one was a small, matched-pair
retrospective cohort study that assessed the impact of screening on language acquisition and
communication ability.
The results indicate that children with bilateral permanent childhood hearing impairment
(PCHI) born in hospitals with UNHS have better receptive language abilities than children with
PCHI born in hospitals without screening. There were contradictory results on expressive
language acquisition with the larger, more recent study finding no statistically significant or
clinically important differences. Differences seen in the smaller retrospective study need to be
considered with caution as assessment of the language outcome measures was unblinded to
the screening status of the child.
Conversely, the independent, blinded assessment conducted by Yoshinaga-Itano and
colleagues of communication ability lends credence to the finding that the superior speech
intelligibility of some children with bilateral PCHI is associated with birth in a screening
hospital. Whilst findings were not consistent with Kennedy et al.( 2006), this may be a result
of parental assessment, rather than direct assessment of children’s communication.
Information on the impact of UNHS programs on the longer term outcomes (ie educational
and employment status) is unlikely to be reported in the peer-reviewed literature for another
decade or so.
Universal neonatal hearing screening
99
What are the economic considerations?
The purpose of economic evaluation is to assist decision-makers in ensuring that
society’s ultimately scarce resources are allocated to those activities from which it will get
the most value. That is, it seeks to enhance economic efficiency.
The aim of this economic evaluation has been to systematically review the evidence for
the costs and effectiveness of a nationwide universal neonatal screening program for
permanent childhood hearing impairment (PCHI) compared to the alternative, no
universal screening program, under Australian conditions. This information was
synthesised in terms of the best current estimates, with an indication of the extent of
uncertainty entailed.
The perspective for the present evaluation was that of Australian society overall. Cost
data covered all resources directly used in establishing and implementing the screening
program. Costs were estimated for the screening program, for subsequent diagnostic
investigations and treatments, and for the longer term implications for patients and their
families and were reported in 2003 Australian dollars. Only the literature review for
UNHS has been updated to 2007. The original 2003 economic analysis has not been
revised. Both short-term and long-term health outcomes were considered.
Quality of relevant economic evaluations
The quality of the economic evaluations of universal neonatal hearing screening (UNHS)
identified in the literature search was assessed according to a checklist previously
developed by the National Health and Medical Research Council (Appendix E). Quality
was expressed in terms of its internal validity as a score out of 10 (which is an attribute of
the original study design). The external validity, or generalisability (which depends on the
setting), was expressed as a score out of 6. Description of the evaluated studies and their
resulting quality scores can be found in Appendix F.
Unit costs
Possible costs of universal neonatal screening for permanent childhood hearing
impairment
Notional costings of UNHS were calculated using a costing model derived from a
template developed by the Medical Services Advisory Committee (MSAC 2000).
Developers of established screening programs such as the Western Australian UNHS
program were consulted to assist with the identification of all non-trivial costs associated
with creating a screening program. Costs of implementing a program can be evaluated
for two possible treatment pathways. Present information suggests that either automated
auditory brainstem response (AABR) or otoacoustic emissions (OAE) tests can be used
in the initial screening in either a one- or two-stage process (Figure 2). Neonates who fail
the first test are either referred directly for medical assessment (1-stage) or are tested
again with either AABR or OAE (2-stage). Both of these pathways are explored in this
report.
100
Universal neonatal hearing screening
The costs of items of service reported for a given year from other countries were first
expressed in Australian dollars for the same year using Purchasing Power Parity
calculations developed by the Organisation for Economic Cooperation and
Development (OECD) (OECD 2003). Conversion to Australian dollars for the first half
of 2003 was done using the Consumer Price Index (CPI) (ABS 2004).
Reporting of the economic evaluation
The results were expressed in terms of the incremental cost-effectiveness ratio (ICER),
which was calculated as shown below:
ICER =
=
increment in cost
increment in effectiveness
cost of UNHS – cost of not universal screening
health outcome from UNHS – health outcome from no universal screening program
The health outcomes of interest are both short- and long-term. The primary short-term
outcomes are the cases of permanent childhood hearing impairment (PCHI) identified
and the cases of PCHI treated (viz when hearing aids are fitted or cochlear devices
implanted). From these outcomes the following incremental cost-effectiveness ratios
were reported:
• cost per case of PCHI identified
• cost per case of PCHI treated.
Where provided in the identified literature, the process measure of cost per infant
screened was also reported.
The long-term outcomes of interest relate to language and social development, quality of
life, education level and employability. The achievement of satisfactory long-term
outcomes may require the allocation of resources to community services and educational
support.
Universal neonatal hearing screening
101
Relevant cost items
When evaluating the cost-effectiveness of a program from a societal perspective, direct
costs must be included from three key areas (Drummond et al 1997) and indirect costs
must also be assessed:
Direct costs
1) Health care sector
Health care sector costs of implementing a universal neonatal hearing screening (UNHS)
program would include: the purchase of screening equipment (AABR and OAE
technology); consumables; the training and deployment of personnel for screening,
diagnostics and administration; and the cost of follow-up diagnostic assessment (Bailey et
al 2002).
2) Patient and family
Costs to patient and family of permanent childhood hearing impairment (PCHI) could
be travel costs and time off work.
3) Other sectors
Non-medical costs that may be incurred by the identification of children with hearing
impairment would be special education and rehabilitation (Mohr et al 2000).
The relevant direct costs of a neonatal hearing screening program are itemised in Table
19.
Indirect costs
The loss to society of the productivity of persons who are unable, or are limited in their
capacity, to work.
102
Universal neonatal hearing screening
Table 16
Relevant resource items for a neonatal hearing screening program
Resource items
Universal screening
Targeted screening
2-stage OAE and
AABR
2-stage AABR
1-stage AABR
OAE instrument
9
n/a
n/a
AABR instrument
9
9
9
Computer
9
9
9
OAE probes
9
n/a
n/a
ABR electrodes
n/a
9
9
OAE probe tips
9
n/a
n/a
Ear couplers/muffins
n/a
9
9
Cables
9
9
9
Machine calibration
9
9
9
Other consumables (eg paper)
9
9
9
Screener training
9
9
9
Overheads
9
9
9
Nurse
9
9
9
Audiologist
9
9
9
Administrative support
9
9
9
Follow-up
9
9
9
Patient and family
9
9
9
Other sectors
9
9
9
Capital equipment
Recurrent items
Personnel
AABR = automated auditory brainstem response test; OAE = otoacoustic emissions; 9 = relevant; n/a = not applicable.
Reported economic evaluations
A systematic and comprehensive literature search identified all published economic
analyses of the effectiveness of universal neonatal hearing screening (UNHS) up until
2003. Relevant papers were retrieved and subjected to critical appraisal, and
bibliographies were searched for additional pertinent literature. For a detailed analysis of
each of these studies see Appendix F.
Comparison of universal and targeted screening in the short term
Two studies determined that UNHS would, in the short term, be more expensive and
less cost-effective to operate than a targeted screening program. Disparity in the
incremental cost-effectiveness ratio between these two studies, for incremental cost per
case identified, appears to stem from the different referral rates used to determine the
costs of follow-up diagnostic assessment. In the study by Kemper and Downs (2000) the
assumed difference in referral rates from universal, as opposed to targeted, hearing
screening programs was much smaller than that assumed by Keren et al (2002). With
similar assumed differences in referral rates, the overall cost of diagnostic follow-up for
each of the programs presented by Kemper and Downs (2000) would be closer, reducing
the incremental cost.
Universal neonatal hearing screening
103
Comparison of modelled universal neonatal hearing screening programs in the
short term
In three modelled evaluations of UNHS protocols two- or three-stage otoacoustic
emissions (OAE) screening, along with two-stage OAE and automated auditory
brainstem response (AABR) screening, were recognised as the most cost-effective
methods in the short term for identifying permanent childhood hearing impairment
(PCHI). Differences in cost per child identified are substantial between studies, although
costs per child screened are similar (Boshuizen et al 2001; Kezirian et al 2001). This
could only be due to differences in referral rates and/or costs assumed for diagnostic
testing, neither of which can be compared from the information published. This is based
on evidence obtained prior to 2003.
Comparison of existing universal neonatal hearing screening programs in the
short term
From studies of existing UNHS protocols, the cost per child screened including followup was quite variable (Table F, Appendix C). Two-stage protocols ranged from $25 to
$80 per child screened while one-stage protocols ranged from $29 to $46 per child
screened. Estimates of the cost per child identified were similarly varied, ranging from
$6,700 to $23,000 for two-stage screening programs, and from $20,000 to $32,000 for
one-stage protocols.
In the only study comparing one-stage transient evoked otoacoustic emissions (TEOAE)
or AABR with two-stage TEOAE–AABR hearing screening, the cost per child screened
and cost per child identified with PCHI were roughly equivalent (Vohr et al 2001). In a
single study comparing two-stage OAE with two-stage AABR, the latter was considered
more cost-effective because referral rates were lower, despite the initial higher cost outlay
for AABR equipment (Lemons et al 2002).
Two included studies presented cost per child screened estimates that were substantially
lower than for comparable studies examined in this evaluation. In the one-stage TEOAE
study by Driscoll (Driscoll et al 2000) the screened sample were drawn from volunteer
walk-in subjects to a child health clinic (CHC). Cost per child screened here was
noticeably lower ($13), probably due to the higher prevalence of hearing impairment in
this population, since concerned parents are more likely to attend the CHC with their
child than parents of seemingly healthy infants. Similarly, the cost of using a two-stage
TEOAE protocol for screening was determined to be only $11 per child screened before
the costs of diagnostic assessment (Weirather et al 1997). In this study, however, the
initial purchase cost of screening equipment was not included.
On the basis of this evidence, it is difficult to recommend any one protocol for UNHS.
Comparison of universal and targeted screening in the long term
Only one study attempted to model the cost-effectiveness of UNHS over the long term
and from a societal perspective (Keren et al 2002). A hypothetical cohort of 80,000
newborns from one US state was used, consisting of 10,400 high-risk and 69,600 low-risk
infants. Under targeted screening, only infants with identified risk factors for congenital
deafness were screened. Analysis resulted in an incremental saving of $2 million per child
with normal language using targeted screening compared to having no screening
program, and a further saving of $1.2 million per child using UNHS. Modelling of long104
Universal neonatal hearing screening
term costs by Keren et al (2002) suggested that identifying a larger proportion of hearingimpaired infants at an early stage (ie ≤6 months of age) would result in savings in other
sectors (such as education and social welfare) that would far outweigh the initial expense.
Overall summary of published economic evaluation studies of neonatal hearing
screening
With the exception of two studies, all identified published research examining the costeffectiveness of existing or modelled universal neonatal hearing screening (UNHS) programs
are from an American perspective. In the only study from Australia, the study design did not
reflect the optimal conditions for a universal screening program, as the participation was
voluntary. Therefore, due to differences in costs and the structure of the US medical system,
the results obtained from the majority of the literature can be only suggestive of what might
occur under Australian conditions.
Short-term costs and benefits
From the literature it can be concluded that, in the short term, costs for the additional cases
identified and diagnosed by UNHS are greater per unit than those of targeted screening.
However, it is inappropriate to determine incremental cost-effectiveness without considering
long-term costs and cost savings. Further, modelled two-stage UNHS protocols appear to be
more cost-effective than modelled one-stage protocols, but this observation is not supported
by reports of existing UNHS programs using either one- or two-stage screening protocols.
Therefore, based on the available evidence, no determination of the most cost-effective
protocol for UNHS can be made.
Long-term costs and benefits
While it was established earlier in this report that UNHS would be more expensive to operate
than a targeted screening program, taking a societal perspective over the long term suggests
that identifying a larger proportion of hearing impaired infants at an early stage (ie ≤6 months
of age) would result in an overall cost-effective program. However, these societal costs are
based primarily on observation and expert opinion and have not been generated from
properly designed studies.
Universal neonatal hearing screening
105
Economic model
Decision analytic modelling can assist the policy making process by simulating the costeffectiveness of a program of universal neonatal hearing screening (UNHS) for
permanent childhood hearing impairment (PCHI) for Australia, compared to the current
situation where there is no universal screening program.
In 2003, targeted neonatal hearing screening was limited in Australia, and only New
South Wales approximated a statewide UNHS program. Depending on the health
service, PCHI was usually identified at various stages in an infant’s development (Figure
5) §.
Figure 5
2003 timeline of identification of PCHI in Australia in the absence of a UNHS
program **
in NICU
when "risk factors" identified
all newborns
tested
tested
parental concern
after discharge from birthing hospital
community nurse follow-up
school entry
test between 0 and 5 years
test_0_to_5_years
test between 0 and 5 years
test_0_to_5_years
test between 0 and 5 years
test_at_5_years
The varied situation across Australia poses a question as to the most appropriate
comparator against which to estimate the incremental cost-effectiveness ratio for UNHS.
Two comparators are available: no formal screening program and a targeted screening
program (ie identification of potential cases on the basis of risk factors alone, and
including both formal and informal identification methods).
Therefore, the use of decision analytic modelling was explored to simulate the following
decisions:
1. The respective costs, cost savings (in both the short and long term) and yield of
UNHS compared to no screening or targeted screening.
2. If a universal screening program was adopted, which two-stage protocol, AABR–
AABR or OAE–AABR, is the most cost-effective?
It is noted that a few infants may acquire their hearing loss after birth but during the neonatal hospital
stay, and still others will develop an acquired hearing loss after discharge.
§
**
106
‘Risk factors’ may be identified through either case finding or a formal screening program.
Universal neonatal hearing screening
The model was designed to calculate the incremental cost per additional case of PCHI
diagnosed by the age of 6 months, the result allowing calculation of the incremental cost
per additional person with normal language. The definitions of these ratios are:
Short-term ICER =
additional costs of universal screening and diagnostic follow-up
number of additional infants with PCHI identified by universal screening
Long-term costs /
cost savings =
costs of universal screening and diagnostic follow-up
minus savings on educational/rehabilitation services and productivity gain
Possible hearing testing programs for neonates
Three possible scenarios for a neonatal hearing screening program that a health authority
might provide are described in Figure 6:
1) no formal screening program: individual infants identified by clinical case finding.
2) a targeted program where newborns identified with a risk factor for permanent
childhood hearing impairment (PCHI) and thus deemed at increased risk of PCHI
(Box 7) would be tested by a one-stage screening protocol, progressing directly to a
diagnostic evaluation if they fail the test. In this program infants with PCHI but
without recognised risk factors would only be exposed by case finding
Box 7
Established risk factors for targeted neonatal hearing screening
Residence in neonatal intensive care unit / special care baby unit for ≥48 hours
Prolonged usage (>7 days) of aminoglycosides
Family history of permanent childhood deafness
Craniofacial abnormality noticeable at birth
Perinatal infection (either suspected or confirmed), eg toxoplasmosis, rubella, cytomegalovirus,
herpes or acquired meningitis
Birthweight <1.5 kilograms
Birth asphyxia
Chromosomal abnormality, including Down syndrome (Trisomy 21)
Exchange transfusion or intrauterine transfusion, eg hyperbilirubinaemia
Intracranial haemorrhage
Universal neonatal hearing screening
107
3) a universal neonatal hearing screening (UNHS) program in which all babies are
screened. This implies that all babies will enter a UNHS program but will follow a
protocol dependent on the identification of risk factors for PCHI (including
residence in the neonatal intensive care unit (NICU)). Infants with identified risk
factors would be screened using the one-stage protocol as above, while infants not
identified as having risk factors (‘well babies’) would still be tested, but using a twostage screening protocol.
Figure 6
Options for neonatal hearing screening programs
no screening program
case finding
case_finding
NICU
identified at risk
all newborns
other risk factors
targeted
well babies
NICU
one_stage
one stage screening
identified at risk
other risk factors
well babies
one stage screening
one_stage
case_finding
case finding
screening program
universal
one_stage
one
stage screening
one_stage
one
stage screening
two_stage
two stage screening
However, determining the costs involved in any formal neonatal hearing screening
program is complicated by the need for screening facilities to be available in different
locations, even within the same health service, in order to assure the complete coverage
of all births (Table 17). Having identified that there will be different costs for hospitals in
rural and remote regions, the costs for the urban setting have been calculated. It is
acknowledged that screening in rural and remote settings is likely to be more costly due
to diseconomies of scale and to parental costs of transporting a child to a diagnostic
facility.
108
Universal neonatal hearing screening
Table 17
Region-dependent design of UNHS program
Urban region
Birth location
Infant designation
Screening protocol used
Screening location a
Tertiary hospital
NICU
1-stage AABR
H
Community
hospital
Identified at risk
1-stage AABR
H
Well babies
2-stage OAE or AABR–AABR
H or H/C
Lost to first screen (eg early
discharge)b
2-stage OAE or AABR–AABR
C or H or H/C
Well babies
2-stage OAE or AABR–AABR
H or H/C
Identified at risk
1-stage AABR
H
2-stage OAE or AABR–AABR
C or H or H/C
Lost to first screen
Rural and remote region
Regional hospital
Community
hospital
Well babies
2-stage OAE or AABR–AABR
H or H/C
Identified at risk
1-stage AABR
H
Lost to first screen
2-stage OAE or AABR–AABR
C or H or H/C
Well babies
2-stage OAE or AABR–AABR
H or H/C
Identified at risk
1-stage AABR
H
Lost to first screen
2-stage OAE or AABR–AABR
C or H or H/C
H = hospital, C = community, H/C = 1st stage in hospital but 2nd stage performed in community. b It is assumed that those ‘lost to first screen’
would be tracked by screening personnel and followed up with testing in the community or as outpatients at the local hospital. NICU = neonatal
intensive care unit; AABR = automated auditory brainstem response test; OAE = otoacoustic emissions test.
a
Short-term costs
Short-term costs are those incurred in the identification of unilateral permanent
childhood hearing impairment by the age of 6 months. Calculations are based on a
hypothetical birth cohort of 4,000 births per annum. This number is typical of the annual
number of births recorded for a tertiary birthing hospital.
Individual resource costs are summarised in Table 18. At least two commercial
quotations were obtained for each item of capital equipment in 2003, and in each
instance a clinically useful life of 5 years accorded with the manufacturer’s estimate. The
equivalent annual cost was calculated according to the annuity method (Drummond et al
1997).
Universal neonatal hearing screening
109
Table 18
Identification of 2003 unit costs for neonatal hearing screening ††
Resource item
Cost ($A)
Range ($A)
Manufacturer
14,943–15,781
Madsen
Capital items
Algo 3i portable AABR
19,380
Accuscreen AABR
14,943
Natus
ABaer AABR
19,500
Biologic
ECHOCHECK portable TEOAE
6,350
Medtel
Accuscreen TE/DPOAE
12,088
12,088–13,229
Madsen
Accuscreen OAE/AABR
19,207
19,207–23,974
Madsen
ABaer OAE/AABR
29,900
Biologic
Probe (AABR/OAE, Accuscreen)
1,545
Madsen
Probe (AABR/OAE, ABaer)
1,246
Biologic
Recurrent costs – consumables
Probe tips (AABR, ABaer)
3
Probe tips (AABR, Accuscreen)
1
Biologic
Couplers or muffins (AABR, Algo 3i)
17
Couplers or muffins (AABR, ABaer)
16
Biologic
Electrode (AABR, ABaer)/infant
3
Biologic
1–2
Madsen
Natus
Electrode (AABR, Algo 3i)/infant
5
Natus
Electrode (AABR, Accuscreen)/infant
6
Madsen
Probe (OAE, ECHOCHECK)
688
Medtel
Probe tips (OAE, ECHOCHECK)
1
1–2
Medtel
($A)a
Range ($A)
Time (minutes/infant)
Resource item
Cost
Recurrent costs – wages including on
costs ($/hr)
Diagnostic evaluation/infant
Coordinator
25
9
Clerk
14
4
Screener
17
35
Audiologist
24
4
170b
a Unless otherwise stated, recurrent costs obtained from the Western Australia universal neonatal hearing screening project; b determined by
addition of MBS items # 11300, 11332, 11330, 11327. AABR = automated auditory brainstem response test; OAE = otoacoustic emissions test;
DPOAE = distortion product otoacoustic emissions test; TEOAE = transient evoked otoacoustic emissions test.
The costs of automated auditory brainstem response tests
Capital equipment
For the purpose of this analysis it is assumed that one machine is required for every
2,000 infants screened per annum ‡‡. In Australia, in 2003, the purchase price of an
The capital items recorded in this table are indicative only of what is used in Australia but does not
include all available products.
††
It is recognised that some hospitals may already have sufficient machines in place and therefore the cost
of targeted screening for some hospitals in the initial years will be nearer that calculated for the recurrent
cost items. On the other hand, in rural and remote regions, distance may preclude the sharing of machines
‡‡
110
Universal neonatal hearing screening
automated auditory brainstem response (AABR) machine was $17,950. With an expected
clinically useful lifetime of 5 years, the equivalent annual cost was $4,150 (annuity
method).
Presently, AABR technology allows for the use of either probe tips or ear couplers to
deliver the stimulus. If the option of using probe tips is chosen, then the 2003 purchase
price of one probe capable of both otoacoustic emissions (OAE) and AABR testing was
$1,400. Assuming a 5-year lifetime, each probe had an equivalent annual cost of $325.
Consumables
One probe tip with one set of electrodes was $7 per infant screened (Table 18).
No probe is required when using ear couplers. One ear coupler with one set of
electrodes was $21 per infant screened.
The costs of otoacoustic emissions tests
Capital equipment
Again it is assumed that one machine is required for every 2,000 infants screened per
year. Based on two quotes of single function (ie transient evoked otoacoustic emissions
(TEOAE) or distortion product otoacoustic emissions (DPOAE) only) the average
purchase price of a single OAE machine in 2003 was $7,000. With an expected clinically
useful lifetime of 5 years for each machine, the equivalent annual cost of one OAE
machine was $1,625. Equipment is available that can perform both TEOAE and
DPOAE at almost twice the purchase price (Table 18).
One probe for OAE measurements was $688 (Table 18). Assuming a 5-year clinically
useful lifetime, each probe had an equivalent annual cost of $159.
Consumables
One tip per screen was $1.00.
Screener salary
From the Western Australian UNHS project it was estimated that the average time taken
per screen was 0.6 hours. The hourly salary of a designated screener was reported to be
$17 (nursing staff would receive much higher remuneration) and therefore the cost for
the screener’s time would be $10 per screen.
Diagnostic follow-up
The cost to the MBS of one diagnostic follow-up appointment is $170 per infant. This
was calculated from the combined benefits charged for four diagnostic tests (see Box 8).
between hospitals; in this situation a greater number of machines may be required unless this can be offset
by a mobile screening service (which will incur travel costs).
Universal neonatal hearing screening
111
Box 8
MBS items for diagnostic assessment in paediatric audiometry
MBS # 11300 Brainstem evoked response audiometry.
Fee $156.05, Benefit $117.05
MBS #11332 Otoacoustic emissions audiometry for the detection of permanent congenital hearing
impairment performed by, or on behalf of a specialist or consultant physician, on an infant who is at
risk due to one or more of the following factors:
admission to a neonatal intensive care unit
family history of hearing impairment
intrauterine or perinatal infection (either suspected or confirmed)
birthweight less than 1.5 kg
craniofacial deformity
birth asphyxia
chromosomal abnormality, including Down syndrome
exchange transfusion,
where the patient is referred by another medical practitioner, and where middle ear pathology has
been excluded by specialist opinion.
Fee $47.45, Benefit $35.60
MBS #11330 Impedance audiogram where the patient is not referred by a medical practitioner – one
examination in any 4-week period.
Fee $6.40, Benefit $4.80
MBS #11327 Impedance audiogram involving tympanometry and measurement of static compliance
and acoustic reflex performed by, or on behalf of, a specialist in the practice of his or her specialty,
where the patient is referred by a medical practitioner – being a service associated with a service to
which item 11309, 1132, 11315 or 11318 applies.
Fee $16.00, Benefit $12.00
112
Universal neonatal hearing screening
Cost-effectiveness of screening protocols
No screening protocol (case finding)
In an Australian study by Russ and colleagues (2002) using a screening questionnaire to
identify risk factors for permanent childhood hearing impairment (PCHI), 4.4 per cent of
a birth cohort of 64,116 infants in 1993 were referred for diagnostic assessment.
Seventeen of 120 infants, later confirmed with PCHI were identified before the age of 6
months. Out of the total birth cohort, the yield was 0.3/1,000 infants. Therefore, if no
screening protocol was implemented it was assumed that this rate would be even lower.
Using the above referral rate and yield as a maximum in a hypothetical birth cohort of
4,000, 176 infants would be referred for audiologic assessment and only one child per
year would be identified with PCHI before 6 months of age. (Russ et al 2002)
Summary: Cost of no formal screening program
With no formal screening method, it would cost a maximum of $30,000 per 4,000 newborns
per year for one child with permanent childhood hearing impairment to be identified before 6
months of age. In a total population of 250,000 newborns per annum throughout Australia, 68
children would be identified by 6 months of age, at a total 2003 cost of $1,870,000 per annum.
Targeted neonatal hearing screening
In a program of targeted neonatal hearing screening, only infants with recognised risk
factors would be directed to a hearing test. Ideally this would occur prior to the discharge
of the infant from hospital. Figure 7 describes the decision analytic model for
determining the short-term costs and effectiveness of a targeted program of hearing
screening in neonates (ie based on risk factors for PCHI being identified prior to hospital
discharge).
Figure 7
Decision model for targeted hearing screening §§
PCHI (case finding)
no identified risk factors
no PCHI
PCHI
dx evaluation
all newborns
no PCHI
fail
lost to dx evaluation
tested AABR
pass
identified at risk or in NICU
lost to test
§§ Case finding assumes that any neonates with PCHI may be discovered through parental concern, health
visitor testing or testing at school entry.
Universal neonatal hearing screening
113
This model acknowledges that some newborns at high risk of PCHI may go unidentified
and therefore screening would not be performed prior to discharge. Further, this model
also recognises that a percentage of infants who have no risk factors for PCHI would
have congenital hearing loss.
In the absence of a full systematic review to determine the best available transition
probabilities for targeted screening protocols, a search of PubMed, Embase and the
MSAC’s existing literature database was performed to identify any studies that used
automated acoustic brainstem response (AABR) on an identified ‘at-risk’ population.
Present technology allows for the use of AABR with either probe tips similar to those
used for otoacoustic emissions (OAE) testing or the traditional ear couplers (or muffins).
For the purposes of this model, this distinction is not assumed to result in different
yields; however, it will affect both the cost per child screened and the cost per child
identified. Using transition probabilities from Table 20 and costs of resource items from
Table 18, the short-term outcomes cost per child screened and the cost per child
identified by 6 months of age were calculated as follows:
Yield of targeted screening protocol (one-stage AABR)
In a cohort of 4,000 newborns it would be expected that 8.1 per cent, or 342 infants, at
risk of PCHI (Kennedy et al 1998) would be identified and 314 infants (coverage 97%)
(Barsky Firkser & Sun 1997) would be screened. This would result in a yield of six (5.7)
infants with unilateral or bilateral PCHI identified before 6 months of age, or 1.5/1,000
infants.
In the remainder of the birth cohort it is assumed that those infants otherwise identified
by case finding would be identified by the screening method as they are assumed to be
the most obvious cases at risk of PCHI. Therefore, no infants in the remainder of the
cohort would be identified using a case finding protocol.
However, from the cohort of 4,000 infants, 14 will be lost either to initial testing (10) or
to diagnostic evaluation (4). Assuming that they would have failed testing at the same
probability as those that were tested, it would be expected that less than one of those lost
to initial testing would have PCHI, but that two of the four lost to diagnostic follow-up
would eventually be diagnosed for PCHI. However, for the purposes of this model, it is
assumed that these infants would not be identified by 6 months of age.
Program management costs for one-stage AABR and follow-up
In the absence of any available costing studies on one-stage AABR followed by
diagnostic assessment, costs of administrative support per child screened were
determined from individual reported data from the Western Australian UNHS project. It
is assumed that one full-time program coordinator would be required for enrolment and
tracking of any sized program up to 12,000 prospective neonates per year at an annual
2003 reported salary of $60,840 including on-costs. Assistance for data entry was
estimated at $1.20 per infant or $4,800 per year for a cohort of 4,000 infants in 2003.
Incremental cost of targeted screening
Therefore, from a birth cohort of 4,000 infants per year, the incremental cost of
screening 320 infants identified as at risk of PCHI using a targeted screening program (1114
Universal neonatal hearing screening
stage AABR) is shown in Table 19. Eleven infants are referred on to diagnostic
assessment using this protocol.
Table 19
2003 cost of targeted screening by method of AABR delivery for a cohort of 4,000 infants per
year
Item
Cost per method of AABR delivery ($A)
Probe tips
Couplers
AABR
4,150
4,150
Probe
325
n/a
AABR probe tips plus electrodes (@ $7 per screen)
2,240
n/a
AABR couplers plus electrodes (@ $21 per screen)
n/a
6,720
Screeners’ time (@ $10 per screen)
3,200
3,200
Program coordinator (1 FTE)
60,840
60,840
Data entry
4,800
4,800
1,870
1,870
77,425
81,580
242
255
Capital
Consumables
Administrative support
Diagnostic assessment
Total cost
Cost per child screened
n/a = not applicable; AABR = automated auditory brainstem response test; FTE = full time equivalent.
Universal neonatal hearing screening
115
Table 20
Transitional probabilities for an Australian targeted screening program
Transition
Base case
probability
Source for base
case
Justification for
selection
Range for
sensitivity
analysis
Source for
range
Protocol 1-stage AABR
Percentage at risk
0.081
Kennedy et al
1998
Only study including
all identified at risk
n/a
Probability screened
0.970
0.030
Only study that
described coverage
0.970–1.000
Probability lost to
screen
Barsky-Firkser &
Sun 1997
0.048
Mason &
Herrmann 1998
Unilateral PCHI,
AABR threshold of 35
dB HL in NICU
population
If screened:
probability initial
screen positive
probability initial
screen negative
If screen positive:
probability receive
diagnostic follow-up
probability lost to
diagnostic follow-up
If followed up
diagnostically:
probability
diagnosed PCHI
probability
diagnosed not
PCHI
0.952
0.730
McClelland et al
1992
Unilateral PCHI, ABR
threshold of 30 dB HL
in SCBU
McClelland et al
1992
Unilateral PCHI, ABR
threshold of 30 dB HL
in SCBU
0.000–0.030
0.048–0.210
0.790–0.952
Barsky-Firkser
& Sun 1997
Mason &
Herrmann
1998;
McClelland et al
1992; Norton et
al 2000
0.270
0.516
0.484
AABR = automated auditory brainstem response test; PCHI = permanent childhood hearing impairment; NICU = neonatal intensive care unit;
dB = decibels; HL = hearing loss; SCBU = special care baby unit.
Summary: cost of a targeted screening program
A yield of six infants with permanent childhood hearing impairment before 6 months of age is
obtained from a birth cohort of 4,000 per year, using a targeted screening program at a total
2003 cost of approximately $79,500 per year and a cost per infant screened of approximately
$249.
116
Universal neonatal hearing screening
Universal neonatal hearing screening
In a program of universal neonatal hearing screening (UNHS) all newborns are directed
to hearing testing through one of two pathways (Figure 8). As in targeted screening
programs, all newborns identified as having one of the risk factors indicated in Box 7
undergo one-stage automated acoustic brainstem response (AABR) testing and follow up
with diagnostic evaluation in the event of a failure. As distinct from targeted screening,
UNHS generally directs all other newborns, designated as ‘well’, through a two-stage
testing procedure of either AABR–AABR or otoacoustic emissions (OAE)–AABR
testing followed by diagnostic evaluation.
Yield of universal screening protocol (two-stage AABR–AABR or OAE–AABR)
Based on transition probabilities presented in Table 21 and Table 22 UNHS on a cohort
of 4,000 infants per annum would identify:
•
six (5.7) infants with unilateral or bilateral permanent childhood hearing
impairment (PCHI) before 6 months of age using risk factors as an indicator for
one-stage AABR screening; and
•
five (5.2) infants with unilateral or bilateral PCHI by 6 months of age using twostage AABR, or seven (6.7) infants using two-stage OAE–AABR in the ‘well’
baby population.
This gives a total prevalence of unilateral and bilateral PCHI in this cohort of 2.7–
3.1/1,000.
Program management costs
The 2003 costs for administrative support were calculated by first assuming that the
cohort would require a full-time program coordinator at a cost of $60,840 (source:
Western Australian UNHS project). This cost was included in a total cost of $93,584 for
managing a UNHS program of 12,530 births per year, leaving a total cost of other
administration items of $32,744. This value was divided by the number of infants
managed in the Western Australian UNHS project, giving a cost of $2.60 per infant or
$10,400 per cohort of 4,000 infants (Table 23) in 2003.
Universal neonatal hearing screening
117
Figure 8
Decision model for universal neonatal hearing screening
lost to test
identified at risk (AABR)
pass
test
lost to dx
fail (dx test)
no PCHI
dx test
PCHI
lost to test
all newborns
AABR
pass
test
lost to test
fail (AABR)
pass
test
all others "well babies"
lost to dx
fail (dx test)
no PCHI
dx test
PCHI
lost to test
OAE
pass
test
lost to test
fail (AABR)
pass
test
lost to dx
fail (dx test)
no PCHI
dx test
PCHI
118
Universal neonatal hearing screening
Table 21
Transitional probabilities for an Australian UNHS program (protocol A)
Transition
Base case
probability
Source for base
case
Justification for
selection
Range for
sensitivity
analysis
Source for
range
Expert opinion
based on: Bailey
et al 2002; Hunter
et al 1994
The mid-point of
coverage rates cited
by two hospital-based
studies (one
Australian) with
TEOAE–AABR screen
protocols
0.831–0.962
Bailey et al 2002;
Hunter et al
1994; Kennedy
et al 1998
Bantock &
Croxson 1998;
Daemers et al
1996; McPherson
et al 1998; Owen
et al 2001
Mean initial TEOAE
screen failure rate of
all four studies
utilising TEOAE as a
1st-stage screen on
well babies only and
reporting unilateral
and bilateral fail
criteria
Expert opinion
based on:
Daemers et al
1996; Molini et al
2001
The mid-point of LTFU
rates cited by two
hospital-based studies
utilising TEOAE as a
1st-stage screen on
well babies only
UNHS protocol A: 2-stage TEOAE–AABR
Probability screened
0.950
Probability lost to
screen
0.050
If screened:
probability initial
screen positive
probability initial
screen negative
If initial screen
positive:
probability
rescreened
probability lost to
rescreen
If rescreened:
probability rescreen
positive
probability rescreen
negative
If rescreen positive:
probability receive
diagnostic followup
probability lost to
diagnostic followup
0.134
0.866
0.800
0.200
Bailey et al 2002
0.037
0.963
0.933
Clemens et al
2000
0.067
Universal neonatal hearing screening
Australian
metropolitan hospitalbased study with
TEOAE–AABR screen
protocol. Relevant for
local conditionsa
Hospital-based study
on well babies only
with unilateral fail
criteria and
information on 2ndstage AABR failures
0.169–0.038
0.043–0.197
0.957–0.803
0.749–0.817
Bantock &
Croxson 1998;
Daemers et al
1996;
McPherson et al
1998; Owen et al
2001
Daemers et al
1996; Molini et al
2001
0.251–0.183
0.016–0.050
0.984–0.950
0.721–1.000
0.279–0.000
Expert opinion
based on: Bailey
et al 2002;
Hunter et al 1994
LTFU for
diagnostic testing
from a 2nd-stage
AABR referral:
Bailey et al 2002;
Cox & Toro
2001; Clemens
et al 2000;
Hunter et al
1994;
OudesluysMurphy &
Harlaar 1997;
Yoshida et al
2002
119
Table 21 (cont.)
Transitional probabilities for an Australian UNHS program (protocol A)
Transition
If followed up
diagnostically:
probability diagnosed
PCHIb
probability diagnosed
not PCHI
If initial screen
negative:
probability diagnosed
PCHI
probability diagnosed
not PCHI
If rescreen negative:
probability diagnosed
PCHI
probability diagnosed
not PCHI
If lost to diagnostic
follow-up
probability diagnosed
PCHI
probability diagnosed
not PCHI
Base case
probability
Source for base
case
Bailey et al 2002
0.521
0.478
0.000
Smyth et al 1999;
McNellis & Klein
1997
1.000
Expert opinion
0.001
0.999
Expert opinion
0.050
0.950
Justification for
selection
Australian metropolitan
hospital-based study
with TEOAE–AABR
screen protocol.
Relevant for local
conditionsa
Both studies on well
babies determined that
false negatives on
TEOAE as a first
screen are negligible if
the screen is
implemented under
quiet (ideal) conditions
Assumption that rate of
false negatives is not
negligible, based on:
Schauseil-Zipf & Von
Wedel 1988.
Child is unlikely to be
tested elsewhere and
receive confirmation of
PCHI within 6 months.
Range for
sensitivity
analysis
0.400–0.600
Source for range
Expert opinion –
10% variation each
side of point
estimate
0.600–0.400
0.000–0.500
0.990–0.970
0.000–0.020
Smyth et al 1999;
McNellis & Klein
1997; Jacobson &
Jacobson 1994
Large variation in
false negatives when
include study where
screening occurred
with normal ambient
noise
Expert opinion based
on: Schauseil-Zipf &
Von Wedel 1988
1.000–0.980
Expert opinion
0.030–0.070
0.970–0.930
TEOAE = transient evoked otoacoustic emissions test; AABR = automated auditory brainstem response test; LTFU = lost to follow-up.
a A limitation of using this source is that the study population included both well and at-risk babies – however, there are no studies available that
used this screen protocol on well babies alone; b assumption that this is diagnosis <6 months of age.
120
Universal neonatal hearing screening
Table 22
Transitional probabilities for an Australian UNHS program (protocol B)
Transition
Base case
probability
Source for base
case
Justification for
selection
Range for
sensitivity
analysis
Source for
range
Clemens et al 2000
Hospital-based study
on well babies only
with unilateral fail
criteria and AABR–
AABR screen protocol
0.788–1.000
Expert opinion
based on:
Clemens et al
2000;
Bretschneider
et al 2001;
Mason &
Herrmann
1998;
OudesluysMurphy &
Harlaar 1997;
Rao et al 2002;
Yoshida et al
2002
UNHS protocol B: 2-stage AABR–AABR
Probability screened
0.985
Probability lost to
screen
0.015
If screened:
probability initial
screen positive
probability initial
screen negative
If initial screen
positive:
probability
rescreened
probability lost to
rescreen
If rescreened:
probability
rescreen positive
probability
rescreen negative
If rescreen positive:
probability receive
diagnostic followup
probability lost to
diagnostic followup
Clemens et al 2000
0.021
0.979
Clemens et al 2000
0.830
0.170
Clemens et al 2000
0.146
0.854
Clemens et al 2000
0.933
0.067
Universal neonatal hearing screening
Hospital-based study
on well babies only
with unilateral fail
criteria and AABR–
AABR screen protocol
Hospital-based study
on well babies only
with unilateral fail
criteria and AABR–
AABR screen protocol
Hospital-based study
on well babies only with
unilateral fail criteria
and AABR–AABR
screen protocol
Hospital-based study
on well babies only with
unilateral fail criteria
and AABR–AABR
screen protocol
0.212–0.000
0.006–0.130
0.994–0.870
0.703–1.000
0.297–0.000
0.000–0.189
1.000–0.811
0.721–1.000
0.279–0.000
Clemens et al
2000; Mason &
Herrmann
1998;
OudesluysMurphy &
Harlaar 1997;
Yoshida et al
2002
Clemens et al
2000;
OudesluysMurphy &
Harlaar 1997;
Yoshida et al
2002
Clemens et al
2000;
OudesluysMurphy &
Harlaar 1997;
Yoshida et al
2002
LTFU for
diagnostic
testing from a
2nd-stage
AABR referral:
Bailey et al
2002; Cox &
Toro 2001;
Clemens et al
2000; Hunter
et al 1994;
OudesluysMurphy &
Harlaar 1997;
Yoshida et al
2002
121
Table 22 (cont.)
Transitional probabilities for an Australian UNHS program (protocol B)
Transition
If followed up
diagnostically:
probability diagnosed
PCHI
probability not
diagnosed PCHI
If initial screen negative:
probability diagnosed
PCHI
probability diagnosed
not PCHI
If rescreen negative:
probability diagnosed
PCHI
probability diagnosed
not PCHI
If lost to diagnostic
follow-up:
probability diagnosed
PCHI
probability diagnosed
not PCHI
Base case
probability
0.600
0.400
Source for base
case
Clemens et al 2000 Hospital-based study on
well babies only with
unilateral fail criteria and
AABR–AABR screen
protocol
Expert opinion
0.001
0.999
Expert opinion
0.001
0.999
Expert opinion
0.050
Justification for
selection
Assumption that rate of
false negatives is not
negligible based on:
Schauseil-Zipf & Von
Wedel 1988.
Assumption that rate of
false negatives is not
negligible based on:
Schauseil-Zipf & Von
Wedel 1988.
Child is unlikely to be
tested elsewhere and
receive confirmation of
PCHI within 6 months.
0.950
Range for
sensitivity
analysis
0.500–0.700
Source for
range
Expert opinion –
10% variation
each side of
point estimate
0.500–0.300
0.000–0.020
1.000–0.980
0.000–0.020
1.000–0.980
Expert opinion
based on:
Schauseil-Zipf &
Von Wedel 1988
Expert opinion
based on:
Schauseil-Zipf &
Von Wedel 1988
Expert opinion
0.030–0.070
0.970–0.930
AABR = automated auditory brainstem response test; PCHI = permanent childhood hearing impairment.
Cost of universal screening
In determining the cost-effectiveness of universal screening, several assumptions have
been made. For an annual birth cohort of 4,000 infants in 2003, it was assumed that two
AABR machines and two OAE (if the OAE–AABR option were used) units would be
required.
Further, the cost of UNHS compared to no screening involves the inclusion of the
targeted screening arm for those infants identified to be ‘at risk’ of PCHI. This value was
extracted from Table 23 and the administration support costs were subtracted as they
were already accounted for in the ‘well baby’ arm. This provides an added cost of
$11,785 using probe tips for one-stage AABR and $15,940 using couplers.
122
Universal neonatal hearing screening
Table 23
Incremental cost of UNHS by choice of 2-stage screening method
Item
Cost per method of 2-stage delivery ($A)
AABR–AABR
Well babies only
Probe tips
OAE–AABR
Couplers
Probe tips
Couplers
Capital
AABR
4,150
4,150
4,150
4,150
AABR probe
325
n/a
325
n/a
OAE
n/a
n/a
1,625
1,625
OAE probe
n/a
n/a
159
159
AABR probe tips plus electrodes (@ $7 per screen)
25,347
n/a
n/a
n/a
AABR couplers plus electrodes (@ $21 per screen)
n/a
76,041
n/a
n/a
OAE probe tips (@ $1 per screen)
n/a
n/a
3,492
3,492
Screeners’ time (@ $10 per screen)
36,210
36,210
34,920
34,920
Consumables
1st stage
2nd stage
AABR probe tips plus electrodes (@ $7 per screen)
441
n/a
2,618
n/a
AABR couplers plus electrodes (@ $21 per screen)
n/a
1,323
n/a
7,854
Screeners’ time (@ $10 per screen)
630
630
3,740
3,740
Program coordinator (1 FTE)
60,840
60,840
60,840
60,840
Other administrative costs
10,400
10,400
10,400
10,400
1,530
1,530
2,210
2,210
139,873
191,124
124,479
129,390
39
53
36
37
11,785
15,940
11,875
15,940
151,658
202,909
136,264
145,330
39
52
36
38
Administrative support
Diagnostic assessment
Well babies only
Total cost
Cost per child screened
At-risk infants only (from Table 22)
Total cost
All infants in universal screening protocol
Total cost
Cost per child screened
n/a = not applicable; AABR = automated auditory brainstem response test; OAE = otoacoustic emissions test; FTE = full time equivalent.
Summary: cost of universal neonatal hearing screening
A yield of 11–12 infants with unilateral or bilateral permanent childhood hearing impairment
identified before 6 months of age is obtained from this cohort using universal screening at a
total cost of $136,000 to $203,000 per year. The cost is dependent on the two-stage
screening method used – automated auditory brainstem response (AABR)–AABR or
otoacoustic emissions (OAE)–AABR – and the type of AABR method used (probe tips or
couplers). The 2003 cost per infant screened ranged from $36 to $52.
Universal neonatal hearing screening
123
Incremental cost-effectiveness ratios
Table 24 summarises total yields and the 2003 cost for each of the three screening
options, and Table 25 reports the incremental cost-effectiveness ratio (ICER).
Table 24
Summary cost-effectiveness of three screening options for PCHI in a birth cohort of 4,000
infants per year
Screening option
Number
screened
No screening
0
Number
referred
Yield
30,000
probe tips
244
77,400
couplers
255
81,600
probe tips
36
136,300
couplers
38
145,300
probe tips
39
151,700
couplers
52
202,900
314
1.1
11
5.7
Total program
cost ($A)
0
Targeted screening
176a
2003 cost/child
screened ($A)
UNHS
2-stage OAE–AABR
3,806
2-stage AABR
3,935
24
20
12.4
10.8
based on referral rate in study by Russ (Russ et al 2002). AABR = automated auditory brainstem response test; OAE = otoacoustic emissions
test; UNHS = universal neonatal hearing screening.
a
Table 25
ICER of three screening options for PCHI in a birth cohort of 4,000 infants per year
Screening option comparison
ICER (incremental cost per extra child identified in 2003
$AUD)
UNHS v no screening
2-stage OAE–AABR
2nd stage AABR using probe tips
9,300
2nd stage AABR using couplers
10,100
1st and 2nd stage AABR using probe tips
12,500
1st and 2nd stage AABR using couplers
17,600
2-stage AABR
UNHS v targeted screening
2-stage OAE–AABR
2nd stage AABR using probe tips
8,800
2nd stage AABR using couplers
9,500
1st and 2nd stage AABR using probe tips
14,600
1st and 2nd stage AABR using couplers
23,800
2-stage AABR
AABR = automated auditory brainstem response test; OAE = otoacoustic emissions test; UNHS = universal neonatal hearing screening.
124
Universal neonatal hearing screening
Summary: Cost-effectiveness of universal neonatal hearing screening for permanent
childhood hearing impairment
In the absence of any screening program for a birth cohort of 4,000 infants per year, clinical
case finding will identify one neonate with permanent childhood hearing impairment (PCHI)
before 6 months of age, at a total 2003 cost of $30,000. Dependent upon the type of
acoustic apparatus (probe tips or couplers) used in association with automated auditory
brainstem response (AABR), employment of a universal screening protocol will identify an
extra 10–11 infants with unilateral or bilateral PCHI before 6 months of age in this cohort, at
an incremental cost of $9,000 to $17,600 per infant. Compared to targeted screening,
universal neonatal hearing screening (UNHS) would identify an additional 5–7 infants with
unilateral or bilateral PCHI before 6 months of age in this cohort. The incremental cost of
UNHS compared to targeted screening would be $8,800 per additional infant identified using
two-stage otoacoustic emissions (OAE)–AABR, and from $14,600 to $23,800 using twostage AABR–AABR, dependent on the use of probe tips or couplers with AABR.
Long-term costs
After infants with permanent childhood hearing impairment (PCHI) have been
identified, it is considered best practice to implement both hearing amplification and
rehabilitation as soon as possible, followed where indicated by cochlear implantation
after the age of 12 months. For the purpose of this analysis, long-term is defined as the
remainder of the life-span following diagnosis of PCHI. Clinical anecdote and level III-2
evidence suggests that infants diagnosed by 6 months of age will develop normal
language skills if interventions are implemented immediately, while those identified after
6 months will have some form of language skill deficit compared to their normal hearing
peers. The following estimates are for the incremental long-term costs and cost savings if
a universal neonatal hearing screening (UNHS) program were to be implemented in an
environment where previously there was no organised screening program.
Assuming a national birth cohort of 250,000 infants per annum, adopting a UNHS
program Australia-wide in an environment where there was previously no organised
screening program would result in an incremental annual yield of 607 infants (2-stage
automated acoustic brainstem response (AABR)) with unilateral or bilateral hearing
impairment.
Studies of the long-term cost of hearing impairment
Little published literature is available that describes the long-term costs and benefits of
early intervention in infants with PCHI. Three studies have attempted to determine the
long-term costs of deafness (Downs 1994; Mohr et al 2000; Weinrich 1972).
Three decades ago Weinrich (1972) used census data and educational statistics to
estimate that the economic costs of deafness amounted to the equivalent in 2003
Australian dollars of more than $410,000 per person with hearing impairment over the
life of the person. This value was based only on the expected loss of lifetime earnings
due to the reduced probability of a hearing-impaired child reaching various educational
milestones. This information suggests that deaf infants lag behind their hearing peers
Universal neonatal hearing screening
125
educationally, fail more often to continue with their education, and graduate from school
and enter the work force at a later age.
Downs (1994) described more recent data showing that deaf members of American
society earned on average 30 per cent less than their hearing peers. Information from the
Internal Revenue Service indicated that 24 per cent of deaf college graduates reported no
income and that hearing-impaired secondary school graduates have twice the rate of
unemployment of the national average. The report argued that this disparity is due in part
to poorer language skills of deaf persons compared to their hearing peers. For example,
the average reading level of 15-year-old normal hearing students corresponded to school
Year 10, whereas their deaf peers had an average reading level at Year 3.
In a recent study the societal costs of severe to profound hearing loss were estimated in a
US population (Mohr et al 2000). This report estimated that the lifetime financial cost to
society, for each individual with prelingual onset of severe to profound hearing loss,
would be the equivalent in 2003 Australian dollars of over $1.5 million. Both direct and
indirect costs were included in this estimate. Direct medical costs covered diagnosis,
intervention (eg hearing aids) and periodic medical evaluations. Direct non-medical costs
included education and rehabilitation services (such as speech and language pathology).
The costs due to loss of productivity were also considered.
The study revealed that the largest component of cost (67%) that would be borne by
society due to severe to profound hearing loss would be the resulting loss in productivity
due to reduced language skills. The second largest cost to society (21%) was identified as
the consumption of special education resources that would be provided in an effort to
improve language skills. Finally, medical costs made up 11 per cent of the total cost to
society, with over half of this related to hearing aid use. However, this report does not
include moderately impaired individuals who may also experience some reduction in
language ability, require some form of social support such as education, and be at some
risk of lower productivity.
No Australian-published literature on the long-term costs of hearing impairment was
retrieved. Therefore, in order to estimate the associated long-term costs, stakeholders in
the fields of therapy and rehabilitation for childhood hearing impairment were contacted.
Table 28 is a summary of their cost estimates.
2003 costs of hearing treatment
It was assumed that, irrespective of whether the diagnosis of PCHI is made before or
after 6 months of age, infants will receive treatment such as hearing aids or cochlear
implants and will undergo some form of rehabilitation until the age of school entry. The
incremental cost of treatment of screened, compared to not screened, infants was
calculated on the basis of the difference between the earliest age for hearing aid fitting (3
months) and the average age at which hearing impairment is identified in the absence of
a UNHS program (21 months).
Information provided by Australian Hearing estimates that a hearing aid costs
approximately $125 plus $48 for four ear moulds per hearing aid in the first year
(Australian Hearing 2001) (Table 26). This treatment is provided by Australian Hearing,
with Commonwealth Government funding, free of charge at the point of service.
126
Universal neonatal hearing screening
Assessment, fitting and aftercare costs were based on Australian Hearing estimates of 27
hours of service by an audiologist over a 15-month period from month 3 to 18 (the point
where most infants with PCHI join the Australian Hearing program). Distributed evenly
over the period, this results in 16 hours in the first year and 11 hours in the second
(Table 26). Cost per hour for audiologist services was estimated at $32 (Western
Australian UNHS project).
Therefore, the total 2003 cost in the first year for hearing aids and moulds would be $173
per ear. Audiologist time for hearing assessment, fitting and aftercare would be
approximately $512 per child in the first year and $352 in the second year (Table 26).
Direct costs and cost savings to other sectors
Where costs will differ substantially is upon entry to school. For the purposes of this
evaluation, it is assumed that infants identified with PCHI by 6 months of age and
receiving immediate intervention and rehabilitation will develop language skills equivalent
to their hearing peers. This assumption, based on expert opinion because no definitive
evidence yet exists, is revisited in a threshold analysis (see Table 28). Therefore, in the
best case analysis it is assumed that when these infants enter mainstream classrooms they
will require no special education interventions. Those infants identified with PCHI after
6 months of age will have some greater degree of impairment in language skills compared
to their hearing peers and therefore will require some form of special education. It is
acknowledged that infants with limited cognitive ability will incur other costs irrespective
of the time of diagnosis of PCHI, but these costs will not be changed by UNHS.
Universal neonatal hearing screening
127
Table 26
Resources used (in 2003 $AUD) in therapy, rehabilitation and education of infants with PCHI
to 18 years of age (or school Year 12)
Resource item
Unit cost
($A)
# sessions
per year
Cost per
year ($A)
Source
Therapy
Hearing aid
Prosthesis
125a
Australian Hearing
Assessment, fitting and aftercare 1st year
512
MSAC estimate
Assessment, fitting and aftercare 2nd year
352
MSAC estimate
Ear moulds
12
Australian Hearing
Maintenance and overheads @ 30%
MSAC estimate
Cochlear implant
Surgical procedure, hospitalisation and prosthesis
22,984
AR-DRG D01Z
Rehabilitation
To age 5 years
Individual therapy session (per session)
90
45
Cora Barclay Centre
Parent–infant program group session (per session)
90
45
Cora Barclay Centre
Parent guidance individual session (per session)
45
40
Cora Barclay Centre
session)b
90
15
Cora Barclay Centre
Profound hearing impairment (per session)
120
80
Cora Barclay Centre
Unilateral hearing impairment program (per
From school entry to Year 12
Hearing specialist support for mainstream students:
Severe hearing impairment (per session)
120
40
Cora Barclay Centre
Moderate hearing impairment (per session)
120
20
Cora Barclay Centre
Mild hearing impairment (per session)
120
5
Cora Barclay Centre
session)c
120
4
Cora Barclay Centre
Unilateral hearing impairment (per
Incremental school budget increase per student
with hearing impairmentd
Reception to Year 2
19,769
DECS
Years 3 to 7
20,315
DECS
Years 8 to 10
18,747
DECS
Years 11 to 12
18,070
DECS
Based on an estimated cost of $62,500 for an extra 500 hearing aids(Australian Hearing 2001); b infants with unilateral hearing impairment
require fewer sessions per year and only attend the parent–infant program (15 sessions per year are assumed); c infants with unilateral hearing
impairment require fewer specialist sessions per year (approximately 4; d costs of extra teaching and capital resources. AR-DRG = Australian
Refined Diagnosis Related Groups (from the Australian Department of Health and Ageing), The Cora Barclay Centre for deaf and hearing
impaired, South Australia; DECS = South Australian Department of Education & Children’s Services.
a
Rehabilitation costs (identification to age 5)
Based on information provided by Australian Hearing, in 2003, the average age of
hearing aid fitting was 18 months (Australian Hearing 2001). It is expected that universal
screening will bring forward the time to hearing aid fitting to an average of 3 months of
age. Therefore, it is assumed that rehabilitation will also start 15 months earlier than is
presently occurring. The costs of rehabilitation of infants with unilateral or bilateral
hearing impairment before or after 6 months of age (ie with or without a UNHS
program) are shown in Table 27.
128
Universal neonatal hearing screening
Education cost savings (age 5 to school Year 12)
The current educational approach to children with permanent hearing impairment is
direct therapist intervention, with additional assistance to the classroom teachers. Upon
entering mainstream school, the child identified before the age of 6 months would incur
no extra educational costs in a best case scenario. The child identified with bilateral
PCHI later than 6 months of age (assumed to have some language skill deficit) would
require an extra $217,000 in hearing specialist and extra teaching support costs up to
Year 12 (see Table 27). Where the hearing impairment is unilateral, later identified infants
require fewer rehabilitation sessions and do not require extra teaching resources, thus
generating 2003 costs of an extra $4,700 up to Year 12, compared to their peers who are
identified early.
In a worst case scenario children identified before the age of 6 months would still require
some specialist support during their school career but to a lesser degree. Expert opinion
suggests that for a moderately impaired infant, a maximum of 20 visits per year would be
required from Reception to Year 2, 10 visits per year for Years 3–9, and only four visits
per year for Years 10–12. In such a case the cost per child of specialist support
(discounted at 5%), from Reception to Year 12, would amount to approximately $11,000,
and to an overall incremental cost in hearing specialist and extra teaching support of
$206,000 for infants identified with bilateral PCHI later than 6 months of age.
Universal neonatal hearing screening
129
Table 27
2003 cost (discounted @ 5% p.a.) of extra rehabilitation and education per child with
bilateral PCHI dependent on age of identification and whether unilateral or bilateral: best
case scenario
Resource item
Cost per
year ($A)
Cost to Year 12 ($A)
Age of identification and unilateral/bilateral
Unilateral
≤6 months
Unilateral
>6 monthsa
Bilateral
≤6 months
Bilateral
>6 monthsa
Rehabilitation
To age 5 years
Therapy
4,050
n/a
Parent–infant program
4,050
Parent guidance
1,800
Unilateral parent–infant program
1,350
Total
n/a
18,400
12,400
n/a
n/a
18,400
12,400
n/a
n/a
8,200
5,500
6,140
4,100
n/a
n/a
6,140
4,100
45,000
30,300
From school entry to Year 12
Hearing specialist support for
mainstream students:
Moderate bilateral
2,400
n/a
n/a
n/a
Unilateral
480
n/a
4,700
n/a
19,769
n/a
n/a
n/a
23,700b
Education support
Incremental school budget increase
per student with hearing impairment
Reception to Year 2
56,500
Years 3 to 7
20,315
n/a
n/a
n/a
79,800
Years 8 to 10
18,747
n/a
n/a
n/a
36,300
Years 11 to 12+
18,070
n/a
n/a
n/a
20,600
0
4,700
0
216,900
6,140
8,800
45,000
Total
Total cost of rehab and education
247,200b
n/a = not applicable. Assuming that the average infant did not start rehabilitation before 18 months of age (ie no costs first year and 1/2 of
second year); b total cost based on yearly cost of specialist support for student with moderate hearing impairment (ie $2,400)
a
Indirect cost savings (beyond school Year 12 or age 18)
Conceptually, the loss of productivity due to late diagnosis could be compared to average
weekly full-time adult ordinary time earnings discounted over the working lifespan. This
would value the long-term benefits of UNHS by the human capital method. This
estimate would depend critically upon the number unemployed as a consequence of the
delay in diagnosis of hearing impairment. There is no available Australian estimate of this
figure, but Downs in the US (Downs 1994) suggested this to be double the national
average unemployment rate. Taking the Australian unemployment figure to be 6 per cent
(ABS 2003) and ignoring hidden unemployment, for an annual cohort of 405 bilateral
hearing-impaired people, the incremental annual indirect cost would amount to $931.40 x
52 x (0.06 x 405) = $1.18 million. The 2003 value of the indirect costs over the working
lifespan (say age 18 to 60) and discounted at 5 per cent per annum would therefore
amount to more than $20 million. Even though this is merely a rough estimate, it
demonstrates that the potential long-term indirect cost savings from implementing a
UNHS program will far outweigh the direct costs.
Given the salience of indirect cost savings in the eventual decision whether or not to
implement and continue to support a national UNHS program, it is important that more
130
Universal neonatal hearing screening
valid and accurate estimates of the indirect costs of hearing and language skill impairment
be obtained for Australian conditions.
Net long-term savings
Table 28 presents the net savings, with and without indirect savings (due to increased
productivity), over the lifetime of one birth cohort of 250,000 infants, of identifying
PCHI before 6 months of age using a UNHS program, compared to either a targeted
screening program or no formal hearing screening program. These estimates are based
on the assumption that no specialist support is required for bilateral hearing-impaired
infants identified by 6 months of age and that 100 per cent of these infants would attain
normal language skills. When allowance is made for some specialist support for infants
with PCHI that are identified early, there is an overall reduction in the net savings by 16
per cent.
Table 28
Summary: Incremental yield and discounted 2003 costs ($million, discounted @ 5% p.a.) for
a UNHSa program over the lifetime of an Australian annual birth cohort of 250,000 infants
Items
UNHS compared to no organised
screening program
UNHS compared to targeted
screening
Indirect savings
excluded
Indirect savings
included
Indirect savings
excluded
Indirect savings
included
607
607
319
319
Incremental yield
Number of infants diagnosed before
age 6 months
Direct costs
Screening and diagnosis
$7.6
$7.6
$4.7
$4.7
Hearing treatment
$0.7
$0.7
$0.4
$0.4
Pre-school rehabilitation
$6.3
$6.3
$3.3
$3.3
Total direct costs
$14.6
$14.6
Proportion attaining normal
language skills b
$8.4
$8.4
100%
100%
100%
100%
$89 ($75)c
$89 ($75)c
$46.7 ($39.3)c
$46.7 ($39.3)c
Direct savingsb
Mainstream school services
Indirect savings
b
Productivity
n/a
$20.6
n/a
$10.8
Net savings
$74.4 ($60.4)c
$95 ($81)c
$38.3 ($30.9)c
$49.1 ($41.7)c
Based on the 2-stage AABR protocol; b under the assumption that 100% of children identified by 6 months of age will attain normal language
skills (equivalent to their hearing peers); c reduced savings with worst case scenario, n/a = not applicable.
a
One-way threshold analyses were performed to determine the effect on overall
discounted cost savings of variations in the percentage of infants with PCHI attaining
normal language skills. Compared to having no formal screening program, running a
UNHS program is cost saving until the proportion of infants identified with PCHI by 6
months of age and who attain language skills equivalent to their normal hearing peers, is
reduced below 17 percent when direct costs and savings are considered and below 14 per
cent with indirect savings included (Table 29). When the comparator is a targeted
screening program, this proportion is similarly low – at 18 percent when direct costs and
savings are considered and 15 per cent when indirect savings are included.
Universal neonatal hearing screening
131
Table 29
One-way threshold analysis: Incremental yield and discounted 2003 costs ($million,
discounted @ 5% p.a.) for a UNHSa program over the lifetime of an Australian annual birth
cohort of 250,000 infants
Items
UNHS compared to no organised
screening program
Indirect savings
excluded
Total direct costs
Proportion attaining normal
language skills b
UNHS compared to targeted screening
Indirect savings
included
Indirect savings
excluded
$14.6
Indirect savings
included
$8.4
100%
16.4%
100%
13.3%
100%
18%
100%
14.6%
$89
$14.6
$89
$11.8
$46.7
$8.4
$46.7
$6.8
Productivity
n/a
n/a
$20.6
$2.8
n/a
n/a
$10.8
$1.6
Net savings
$74.4
$0
$95
$0
$38.3
$0
$43.1
$0
Direct savings
Mainstream school services
Indirect savings b
Based on the 2-stage AABR protocol; b assuming varying percentages of children identified by 6 months of age will attain normal language
skills (equivalent to their hearing peers). n/a = not applicable.
a
Two-way threshold analysis was performed to assess the concurrent but independent effects
of variations in the proportion of infants with PCHI attaining normal language skills and in
the proportion of hearing-impaired adults who are unemployed. Compared to having no
formal screening program, running a UNHS program is also cost saving for the majority of
possible levels of unemployment amongst the hearing impaired (Figure 9).
132
Universal neonatal hearing screening
Two-way threshold analysis: combinations of proportions of hearing-impaired
persons who are unemployed and PCHI infants who attain normal language skills,
where a UNHS program for an Australian annual birth cohort will be less costly
over the lifetime than no organised screening program
proportion of hearing impaired persons who are
unemployed
Figure 9
1
0.9
0.8
0.7
Favourable (ie
UNHS is less costly
than no screening)
0.6
0.5
0.4
0.3
0.2
0.1
0
0
Unfavourable
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
proportion who attain normal language skills
Triangles (Δ) represent the threshold between favourable and unfavourable cost outcomes
Summary
For universal neonatal hearing screening (UNHS) of an Australian birth cohort, the
discounted direct and indirect cost savings over the lifetime far outweigh the costs of the
screening program and associated diagnosis (Table 30). Thus, over the long term, a
UNHS program will be cost saving for Australian society.
One-way threshold analysis indicates that these cost savings would be realised even if
only about 15 per cent of children identified by 6 months of age were to attain a language
level equivalent to that of their hearing peers. Two-way threshold analysis found that
unemployment rates among people identified with permanent childhood hearing
impairment had little negative effect on these cost savings and were robust, except at low
levels of unemployment amongst hearing-impaired persons.
Under these circumstances, short-term cost-effectiveness ratios are potentially
misleading. The short-term objective should be to maximise yield.
The above conclusions rely on a set of assumptions that are believed to reflect the most
likely Australian situation.
Universal neonatal hearing screening
133
Government expenditure
So far, the analysis has taken a societal opportunity cost perspective on screening
programs for permanent childhood hearing impairment (PCHI). It is also relevant to
consider how the programs will be financed, that is their impact on government
expenditure.
Assuming a national birth cohort of 250,000 infants per annum, the resulting
expenditures by all levels of government combined for the identification, treatment and
rehabilitation of hearing-impaired infants are summarised in Table 32.
Years 1 to 5 – the first five cohorts enter the program
Yield
According to the UNHS decision model, adopting a universal neonatal hearing screening
(UNHS) program Australia-wide in an environment where there was previously no
organised screening program would result in an incremental annual yield of 607 infants
(2-stage automated auditory brainstem response (AABR) testing) with unilateral or
bilateral hearing impairment (Table 24). Therefore, at the end of 5 years 3,035 infants
with unilateral or bilateral hearing impairment will be receiving treatment and
rehabilitation. This is a more conservative estimate than using two-stage otoacoustic
emissions (OAE)–AABR testing, where the incremental yield would be 707 infants per
annum.
Contributions to expenditure over the first 5 years
Expenditures incurred over the first 5 years of the program have been calculated in
constant 2003 dollars and using the lower incremental yield of 607 infants per year with
either unilateral or bilateral PCHI.
Screening and diagnosis
The incremental annual expenditure (compared to no screening) of $121,700 for a cohort
of 4,000 births (Table 30) extrapolates to an annual expenditure over 250,000 births for
screening and diagnosis of $7.6 million. It is anticipated that the screening program
would be provided without charge at the point of delivery, and co-payments for
diagnosis have been ignored.
Treatment — hearing aids (years 1 to 5)
After identification, infants are assumed to undergo immediate treatment and
rehabilitation. Initial treatment involves the fitting of hearing aids. For the purpose of
this evaluation, it is assumed that all identified infants (whether unilaterally or bilaterally
hearing impaired) are fitted with hearing amplification. Information provided by
Australian Hearing estimates that each hearing aid costs approximately $125 plus $48 for
ear moulds in the first year (Australian Hearing 2001). This treatment is provided by
Australian Hearing, with Commonwealth Government funding, free of charge at the
point of service. From Table 8 it is estimated that one-third of the identified cohort
(using the lower value of 607 infants per year) with PCHI would have unilateral
impairment (202) and the remainder would have bilateral impairment (405). Therefore,
total expenditure in the first year for hearing aids and moulds would be $175,000. In the
134
Universal neonatal hearing screening
second year of life the replacement cost of four ear moulds per hearing aid per infant
amounts to a total expenditure of $24,000.
Assessment, fitting and aftercare expenditures were based on Australian Hearing
estimates of 27 hours of service by an audiologist over a 15-month period from month 3
to 18 (the point where most infants with PCHI join the Australian Hearing program).
Distributed evenly over the period, this amounts to 16 hours in the first year of life and
11 hours in the second year (Table 26). The price per hour for audiologist service was
estimated at $32 (Western Australian UNHS project). Therefore, the total expenditure on
audiologist time for hearing assessment, fitting and aftercare would be approximately
$310,000 in the first year and $214,000 in the second year (Table 26), based on the 2003
costs.
Cochlear implants cannot normally be provided until 12 months of age (MBS 2002). It
was assumed that this treatment was carried out before 2 years of age and so the time
delay for implantation between those infants identified either early or late with PCHI
would be small, and the expenditure difference negligible.
Thus, the overall expenditure on treatment with hearing aids would be $485,000 for each
incoming cohort and $238,000 in their second year of life (Table 30).
Rehabilitation
Table 32 reports the expenditures on rehabilitation services for infants up to 5 years of
age. Combined annual expenditure for therapy, the parent–infant program and parent
guidance sessions amounts to $9,900 per child for infants with bilateral impairment and
$1,350 for infants with unilateral hearing impairment. Again, it is assumed that an extra
405 infants per year will be identified with bilateral hearing impairment and 202 with
unilateral hearing impairment in a birth cohort of 250,000 infants across Australia.
Rehabilitation expenditures specific to supporting children with PCHI in mainstream
schools are not included, as the initial cohort of children will not have entered school by
the fifth year of the expenditure projection. Therefore, the total expenditure per year per
cohort for rehabilitation is $4,280,000.
Transfer payments in the first 5 years
People with disabilities are eligible for specific allowances or pensions. For the purposes
of this analysis, it is assumed that all carers of children (<16 years) identified with PCHI
will be eligible for the Carer Allowance of $88 per fortnight (Table 33). Yearly transfer
payments are based on an identified population of 607 infants with unilateral or bilateral
hearing impairment.
Universal neonatal hearing screening
135
Table 30
Estimated additional government expenditure (all jurisdictions combined) over the first 8
years of a national program of UNHS (2003 costs in $’000)
Financial year
Expenditure items
1
2
3
4
5
6
7
8
Outlays ($’000)
Incoming cohort
Screening and diagnosis
7,600
7,600
7,600
7,600
7,600
7,600
7,600
7,600
Treatment (hearing aids)
485
485
485
485
485
485
485
485
Rehabilitation
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
Carer support payments
1,390
1,390
1,390
1,390
1,390
1,390
1,390
1,390
Second year of life
Rehabilitation
4,280
4,280
4,280
4,280
4,280
4,280
4,280
Treatment (hearing aid after care)
238
238
238
238
238
238
238
Carer support payments
640
640
640
640
640
640
640
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
4,280
31,753
31,753
31,753
31,753
Specialist speech and language support
(moderate impairment) in Reception
1,069
1,069
1,069
Educational support in Reception
8,006
8,006
8,006
Specialist speech and language support
(moderate impairment) in school Year 1
1,069
1,069
Educational support in school Year 1
8,006
8,006
Third year of life
Rehabilitation
Fourth year of life
Rehabilitation
Fifth year of life
Rehabilitation
Total annual outlays
13,755
18,913
23,193
27,473
Savings on special education ($’000)
Sixth year of life
Seventh year of life
Eighth year of life
Specialist speech and language support
(moderate impairment) in school Year 2
1,069
Educational support in school Year 2
8,006
Total savings on special education
Net expenditures
136
13,755
18,913
23,193
27,473
31,753
9,075
18,150
27,225
22,678
13,603
4,528
Universal neonatal hearing screening
Summary: total expenditure over the first 5 financial years of the program
Under the assumptions mentioned in the text using two-stage automated auditory brainstem
response (AABR) testing, the total expenditure of the Commonwealth, State and Territory
Governments of Australia combined to identify and care for all infants with unilateral or
bilateral permanent childhood hearing impairment (PCHI) over the first 5 financial years
following the introduction of a universal neonatal hearing screening (UNHS) program using
2003 costs, is estimated to be $13.8 million in the first year, peaking at $32 million in the fifth
year (Table 30).
Operating a UNHS program using two-stage otoacoustic emissions (OAE)–AABR testing,
the total expenditure would peak in the fifth year at $35 million (not shown) due to the higher
number of infants identified with PCHI.
Universal neonatal hearing screening
137
Years 6 to 18 – the first cohort enters mainstream school
Under the main assumption that all infants with unilateral or bilateral permanent
childhood hearing impairment (PCHI) identified by 6 months of age would attain the
same language skills as their hearing peers by mainstream school entry, the following
expenditures on specialist and educational support would be avoided in the best case
scenario. It should be noted that these expenditures are associated with interventions
designed to support hearing-impaired students in mainstream schools.
Expert opinion suggests that a worst case scenario would envisage that all PCHI children
identified early in infancy would receive some specialist support but to a lesser degree.
Under such a situation the net savings for specialist support would decrease by 16 per
cent.
Savings
Specialist speech and language support
From Table 30 the type and expenditure savings of this specialist speech and language
support would depend on the severity of the hearing impairment. The savings per
student with bilateral hearing impairment per year range from $600 for mild impairment
to $9,600 for profound impairment. For the purpose of this analysis, the value of $2,400
per year for moderate impairment is used. For each student with unilateral hearing
impairment, the yearly savings on specialist support is $480 (undiscounted, Table 30).
This amounts to an overall expenditure saving in the first year of $1.1 million
(undiscounted). Expenditure savings for severe and profound hearing loss would be
higher. When the second cohort enters Reception, the total expenditure saving of
eliminating the need for specialist support becomes $2.1 million (undiscounted, Table
32). As the program continues, there will eventually be a maximum of 13 cohorts in
school at any one time, providing an expenditure saving for foregone specialist support
of $11.4 million per year (undiscounted).
Educational support for students with hearing impairment
As with the specialist speech and language support, it is assumed that infants who are
identified with unilateral or bilateral PCHI by 6 months of age would not require
educational support. The expenditure savings of educational support vary depending on
the grade of the child, and are given as amounts that would be provided normally if the
infants did not attain the normal language skills of their hearing peers (Table 28).
From Reception to Year 2 each child with hearing impairment would normally generate
an educational expenditure of $19,769 per year. This would result in a total expenditure
for the Australian birth cohort of $8 million (undiscounted Table 28). When the second
cohort enters Reception and with the first cohort now in Year 1, the total expenditure of
educational support becomes over $16 million (undiscounted). With the implementation
of a UNHS program, these expenditures would become savings.
Transfer payments over remainder of lifetime
When this cohort reaches the age of 16, the Carer Allowance no longer applies and they
may instead become eligible for a disability support pension. For the purposes of this
analysis, only those hearing-impaired persons who are unemployed are assumed to be
eligible for the maximum pension allowance. Under the assumption that infants
138
Universal neonatal hearing screening
identified for PCHI by 6 months of age will have normal language skills, the
unemployment rate of this cohort would be the same as the Australian average. In the
absence of a universal screening program, it is assumed that the unemployment rate for
the hearing impaired would be double that of the national average (Downs 1994), which
in 2003 was about 6 per cent overall (ABS 2003). From an annual national cohort of 607,
this would amount to an additional 36 hearing-impaired people being unemployed. This
is probably a conservative estimate because the unemployment rate for all people aged
under 25 years is much higher than the overall rate; also, the estimate does not account
for the income and assets-tested partial payments that would be provided to those
persons with PCHI who are in limited employment.
Table 31
Savings on transfer payments per child with unilateral or bilateral hearing impairment
Per
fortnight
($A)
Per year
($A)
Additional
number of
persons
eligible
Disability support
pension (16–20
years)
398c
10,348
Disability support
pension (>21 years)
453c
11,778
Savings on transfer
payments
Rate per persona
Total annual expenditure for
each cohort ($A)b
Source
Identified ≤6
months
Identified >6
months
36
373,000
754,000
Centrelink
36
424,000
848,000
Centrelink
These are probably underestimates because the unemployment rate for all people aged under 25 is much higher than the overall rate;
savings per year for cohort of 607 additional children identified with unilateral or bilateral hearing impairment by 6 months of age or after;
c maximum rate of disability support pension received per fortnight – income and assets-tested.
a
b
Sensitivity analysis for government expenditure
The base case assumption was that all infants identified by 6 months of age would
develop normal language skills and would therefore not incur any extra education
expenditures over those generated by normal hearing children. By way of a sensitivity
analysis, this assumption has been modified to evaluate the effect on government
expenditure if a proportion of the children identified by 6 months of age do not achieve
a language skill level equal to that of their hearing peers (Figure 10). It should be noted
that any effects of reduced language skill acquisition on expenditure would occur in the
increased use of special education and in transfer payments.
Universal neonatal hearing screening
139
Figure 10 Predicted additional government expenditure/savings for the first 18 years
following the introduction of a UNHS program
40
Net
20
Outlays
$A (in millions
0
1
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18
-20
Net
-40
Net expenditure (100% norm al language)
-60
90% norm al language
Savings
80% norm al language
-80
50% norm al language
25% norm al language
-100
Year since introduction of UNHS program
% normal language = the percentage of a cohort identified by screening who are assumed to have attained normal for age language skills by 5
years of age
One-way sensitivity analysis illustrates that when 100 per cent, 90 per cent or 80 per cent of
infants diagnosed with PCHI attain the language skills of their hearing peers, the break-even
point (where net expenditures and net savings are equal) occurs between years 8 and 9 after
the inception of the program. However, when only 50 per cent of infants attain the language
skills of their hearing peers, this break-even point shifts down to year 12 of the program; and
when 25 per cent of infants attain the language skills of their hearing peers, the break-even
point will not be reached and a net expenditure of $2 million a year will be incurred.
140
Universal neonatal hearing screening
Summary: Predicted expenditures/savings for the first 18 years following the
introduction of a universal neonatal hearing screening program
Commonwealth and State/Territory Governments incur different elements of the overall
expenditure on hearing screening and diagnosis, treatment, special education and
rehabilitation, and carer and client pensions and benefits.
For the first 5 years that the universal neonatal hearing screening program is in operation,
government would be expected to incur a net expenditure. Thereafter, savings due to the
reduced need for specialist support and education when the cohorts enter mainstream
school quickly become realised. By the time the fourth cohort has entered mainstream
schooling (year 8 of the program), it is estimated that government would experience a net
saving that would increase progressively with each additional cohort. By the time the first
cohort has finished Year 12 at school (at age 18 years), government can be expected to
benefit from a net saving of $85 million in that year using a two-stage automated auditory
brainstem response (AABR) program and $98 million using a two-stage otoacoustic
emissions (OAE)–AABR program.
These projections rely on a set of strong assumptions whose validity may change over time.
The time horizon is 18 years. Once the cohort has entered adulthood, savings on disability
support pensions would also be anticipated.
Universal neonatal hearing screening
141
Discussion
Prevalence of permanent childhood hearing impairment
According to the international literature, approximately 0.7–1.8 infants per 1,000 have
bilateral permanent childhood hearing impairment (PCHI). The median bilateral
prevalence for PCHI of greater than 35 dB HL is 1.3 per 1,000 infants. Unilateral PCHI
occurs in 0.2–1.5 infants per 1,000, with a median unilateral prevalence for PCHI greater
than 35 dB HL of 0.6 per 1,000 infants.
The heterogeneous nature of prevalence rates in the literature is unremarkable given the
different case definitions of PCHI and the varying nature and locations of the sampled
populations. Between study heterogeneity may also be related to the different methods in
which prevalence was ascertained. Results obtained from universal neonatal hearing
screening programs may underestimate the prevalence of PCHI given that newborns are
often lost to follow-up between screening and diagnosis stages. In addition, the length of
follow-up in these studies is typically not sufficient for false negative cases to be
identified. Similarly, prevalence estimates obtained from retrospective examinations of
health and/or education records may be subject to bias through the incorrect or
incomplete reporting of PCHI cases or errors in abstracting records.
There are no population-based data on the prevalence of congenital PCHI in Australia.
Credible estimates, based on the median prevalence data from the international literature,
indicate that 325 Australian children are born annually with moderate to profound
bilateral PCHI. Unilateral PCHI of similar severity occurs in an additional 156 children
born each year. Overall, it is estimated that 481 Australian children are born annually
with either unilateral or bilateral moderate to profound PCHI.
Data available from Australian Hearing (2003) indicates that the above estimate is
reasonably accurate. Rates of hearing aid fittings for children born in the years 1986 to
1998 ranged from 1.66 to 4.12 per 1,000 (or between 415 and 1,002 children per year).
These data are, however, probably an over-estimation due to the inclusion of cases of
aided mild hearing loss, and acquired or progressive hearing losses that would not be
identified through a universal neonatal hearing screening (UNHS) program. The
predicted yield of unilateral and bilateral PCHI from an Australian UNHS program,
using two-stage AABR screening, was estimated in this report to be 607 infants. This is
higher than the median estimate from the literature but lies within the range of estimates
based on Australian Hearing data. It is also clear that early (<6 months of age)
identification and management of hearing loss has been steadily improving in Australia
over the last 16 years – a likely consequence of improvements in the identification of
children at risk of hearing impairment (ie targeted neonatal hearing screening).
Safety of universal neonatal hearing screening
Potential harms from UNHS include the harms from the screening process itself
(physical and/or psychosocial), harms from false positives, harms from false reassurance,
and harms that may arise from early diagnosis.
142
Universal neonatal hearing screening
No studies were available that reported on physical harm caused by UNHS. Local,
transient hypersensitivity reactions to electrode gels are possible but there were no
reported cases in the literature and the MSAC Advisory Panel was also not aware of any
cases.
The Western Australian neonatal hearing screening program has outlined protocols to
minimise cross-infection and possible electrophysical harms with the use of both the
automated auditory brainstem response (AABR) and transient evoked otoacoustic
emissions (TEOAE) screening equipment (see Appendix H). Other UNHS programs are
likely to have developed similar protocols, including the appropriate handling of babies in
neonatal intensive care.
A total of ten poor to average quality studies reported on the psychosocial effects of the
screening process and of obtaining screening results. Maternal anxiety about the
screening process was low across the studies. The screening process itself did not
significantly raise anxiety to a higher level than in mothers whose infants were not
screened, regardless of whether the infant received a positive or a negative result (level
III-2 interventional evidence).
Positive screening status was often associated with an increase in parental anxiety
compared to a negative screen (level III-2 interventional evidence), but the anxiety levels
reported in the studies were within the normal range, so the differences seen were not
deemed to be clinically important. Furthermore, higher levels of knowledge regarding the
meaning of a positive test were associated with lower levels of anxiety (level III-2
interventional evidence), so it is hypothesised that if appropriate information is given to
the parents of infants undergoing screening, few parents will report moderate to severe
anxiety. However, the use of a cross-sectional study designs means that any causal link is
unclear. An increase in knowledge regarding the meaning of the test result could decrease
anxiety. However, increased anxiety levels may also decrease the ability of the parents to
absorb the information about the meaning of test results (leading to lower knowledge).A
positive screening result had more effect on the parent if the screen was performed soon
after birth, as compared to being screened after 2 months (level III-2 interventional
evidence).
When UNHS was compared against a behavioural distraction test (occurring after 6
months of age), no clinically important differences between anxiety levels in mothers of
infants who screened positive (level III-2 interventional evidence) were found. However,
more satisfaction was expressed in the mothers whose babies were screened by UNHS
than the distraction test.
Two studies were inconsistent regarding whether screen status had an impact on the
parental attitude to the child, or on the quality of early interactions (level III-2
interventional evidence). However, it is unknown whether the observed reduced quality
of early mother-baby interactions was clinically important, as raw scores were not given.
The importance of this finding is therefore unclear.
No studies were identified that reported on the psychosocial effects of false reassurance.
While the rate of false negatives found in the literature are low (6.5%), the impact of false
reassurance should be considered. If a negative screening test delays the diagnosis of
hearing impairment, the infant would not received the potential benefits from early
intervention.
Universal neonatal hearing screening
143
The consequences of a ‘true positive’ diagnosis of PCHI were not within the scope of
this review, but also need to be evaluated. Although most cases are eventually identified,
it is possible that an early correct diagnosis of PCHI may lead to increases in maternal or
caregiver anxiety in the early postnatal period. This, in turn, may have an impact on the
parent–child bonding process and maternal postnatal depression. Possible delays in
receiving diagnostic confirmation of PCHI – as part of the screening, rescreening and
audiologic battery testing process – and concomitant delays in implementation of
rehabilitation or management interventions may also affect the psychological wellbeing
of parents.
Component
A
B
C
D
Excellent
Good
Satisfactory
Poor
level III studies with
low risk of bias, or
level I or II studies
with moderate risk
of bias
Evidence base
Consistency
most studies
consistent and
inconsistency
may be explained
Clinical impact
Generalisability
Applicability
slight or restricted
population/s
studied in the
body of evidence
are similar to the
target population
applicable to
Australian
healthcare
context with few
caveats
Effectiveness
Diagnostic accuracy of the screening tests
Test accuracy is crucial to the successful implementation of any screening program.
However, for this assessment there were very few studies that assessed the accuracy of
the screening tests at identifying permanent childhood hearing impairment (PCHI) in
predominantly healthy infants – the target population in a universal screening situation –
compared to an acceptable reference standard. The evidence that was available (level III1 and III-2 diagnostic evidence) was of average quality.
The accuracy of transient evoked otoacoustic emissions (TEOAE) testing – in a onestage screen and compared to conventional auditory brainstem response (ABR) testing –
appears to depend on the level of local ambient noise, as well as the condition of infant
ears at testing. Studies that use a ‘quiet’, although not sound-proofed, environment for
testing elicit sensitivity results of up to 100 per cent, although even under the best
conditions the rate of false positives can still be quite high (8%).
144
Universal neonatal hearing screening
In one study, testing in association with ‘real world’ ambient noise (ie within the nursery)
resulted in low test sensitivity – TEOAE could only accurately detect PCHI in half the
infants with the condition. Given the large proportion of infants with PCHI who may
not be identified (false negatives) under these ‘real world’ noise conditions, TEOAE
testing should only occur in environments that are quiet or even sound-proofed.
The ability of an initial TEOAE screen to positively predict PCHI is very low (1.5%) –
meaning that a failure on an initial TEOAE test would accurately predict PCHI for only
one or two infants out of 100 identified by the screen as having hearing impairment. This
is probably a consequence of the frequency of transient losses in newborns (ear
occlusion), as well as the low prevalence of PCHI in the general population.
In terms of identifying conductive hearing loss, TEOAE was found to have 100 per cent
sensitivity and specificity, as compared to tympanometry, in a study of infants who had
no cerumen occlusion of the ear.
One study of an early model automated auditory brainstem response (AABR) unit was
available to assess the ability of the test, in a one-stage screen, to accurately identify
PCHI in infants, compared to conventional ABR testing. On the basis of this study, it
would appear that the specificity of AABR is particularly good (95.6%) given that its
primary use is as a screening tool on a population of predominantly healthy infants.
However, the trade-off between sensitivity and specificity means that the test has good,
although not excellent, sensitivity at detecting PCHI and thus some false negatives may
result. The ability of an initial AABR screen to positively predict PCHI is very low
(2.2%), although marginally better than a TEOAE conducted under quiet conditions.
Expert opinion indicates that later models of the AABR may have improved diagnostic
accuracy although this has yet to be confirmed empirically.
False positives associated with either test could be reduced with the introduction of a
second-stage or third-stage screen of initial failures prior to diagnostic testing. This may,
however, result in unnecessary caregiver anxiety and added costs and delays in
rehabilitation. False negatives are not likely to be identified until the child is older and
this false reassurance may further lengthen the time until diagnostic assessment and thus
the child’s eventual rehabilitation.
Universal neonatal hearing screening
145
Component
A
B
C
D
Excellent
Good
Satisfactory
Poor
level III studies with
low risk of bias, or
level I or II studies
with moderate risk
of bias
Evidence base
most studies
consistent and
inconsistency
may be explained
Consistency
Clinical impact
Not Applicable
population/s
studied in the
body of evidence
are similar to the
target population
Generalisability
Applicability
directly applicable to
Australian healthcare
context
Effectiveness of universal neonatal hearing screening
Impact on referral, diagnosis and management
Altogether, five controlled studies assessed the impact of universal neonatal hearing
screening on the clinical management of infants with PCHI. This was assessed in terms of
the infants’ age at referral, age at PCHI diagnosis, and age at management. This
information was supplemented by 56 uncontrolled, descriptive studies that provided
information on screening protocol, coverage, failure rates (referrals), losses to follow-up,
false alarms and yield (diagnosis) for the universal screening arm alone.
The best evidence (III-1 screening evidence) available – a quasi-randomised controlled trial
in Wessex, UK – indicates that universal neonatal hearing screening has an impact on the
clinical management of PCHI. Referrals for diagnostic testing and the diagnosis and
management of PCHI commonly occur earlier and more frequently with, as opposed to
without, universal neonatal hearing screening.
Level III-1 screening evidence determined that infants who receive universal neonatal
hearing screening are nearly three times more likely [RR=2.9, 95%CI 1.4, 6.3] to be
referred for diagnostic testing within six months, than infants who are not screened
universally. This means that 1,619 infants would need to be universally screened for
hearing impairment, as compared to not screened, to ensure the referral of one infant
(aged under six months) for diagnostic testing.
Not surprisingly, early referral for diagnosis is believed to be associated with an increase in
earlier diagnosis. The evidence suggests that infants born during periods of universal
neonatal hearing screening are 2.3 times more likely [RR=2.3, 95%CI 1.1, 4.7] to receive a
diagnosis of PCHI, than infants born in periods without universal hearing screening.
However, the low prevalence of the condition leads to a small absolute increase in benefit 146
Universal neonatal hearing screening
with only an extra five children identified per 10,000. Level III-2 screening evidence of
good quality determined that children with PCHI were more likely to be diagnosed before
the age of six months when born during periods of, or in hospitals with, universal neonatal
hearing screening than children who were not exposed to screening programs. The larger
of the studies indicated that a diagnosis of bilateral PCHI before the age of six months was
5.1 times more likely to occur in children who were born in hospitals with universal
neonatal hearing screening than without screening [[RR=5.1, 95%CI 2.1, 12.4]. This means
that for every three children [95%CI 2, 5] with bilateral PCHI born in an Austrian hospital
with a screening program, one additional child would be diagnosed with PCHI before the
age of six months than if born in a hospital without a screening program.
Level III-1 screening evidence also reported that screening may increase the likelihood of
PCHI management before the age of ten months by 2.4 times [RR=2.4, 95%CI 1.0, 5.8].
The absolute increase in benefit over not screening, is early PCHI management (<10
months) for an approximate additional four children in 10,000. In real terms this translates
into screening 2,965 children [95%CI 1458, 86207] for hearing impairment, compared to
not screening, to ensure the early management (< 10 months) of one infant with PCHI.
Descriptive data indicate that the majority of universal neonatal hearing screening
programs manage to screen over 90 per cent of infants in their catchment area. These
programs are largely hospital-based with initial screening occurring prior to discharge.
Community-based studies obtain good coverage when screening is “piggy-backed” onto
other health or immunisation checks at the health clinic or when it occurs in the home.
The lowest coverage (45%) was seen in a study conducted in a private hospital in South
Africa where patients were not subsidised for the screening service for half of the follow
up period. This led to a sudden drop in the population being screened. Coverage, however,
does not appear to vary significantly according to the size of the source population being
screened, or according to the screening protocol utilised.
Losses to follow-up (LTFU) commonly occur when there is a long delay before rescreening or before diagnostic testing. In one study where there was an immediate rescreen, the 37 per cent LTFU was related to the very early discharge rates from the
hospital that was situated in an economically depressed inner-city area. Losses to follow-up
could be mitigated to some extent by instituting comprehensive reminder and educational
strategies for parents.
Given the low positive predictive value of the AABR and TEOAE screening tests and
protocols, the likelihood of having a hearing impairment after failing an initial screen or rescreen is very low. Uncontrolled studies of screening programs suggest that the number of
false alarms associated with TEOAE screen protocols is higher (up to approximately 10%)
than with AABR screen protocols (up to approximately 6%). It is, however, possible to
maintain low false alarm rates using either type of screening protocol.
Impact on adverse outcomes associated with PCHI
There is limited information available on the effect of universal neonatal hearing screening
on the adverse, patient-relevant outcomes associated with PCHI. Two good quality studies
(level III-2 screening evidence), one of which was small, assessed the impact of screening
on language acquisition and communication ability. While both studies were consistent in
finding receptive language to be better in children with bilateral PCHI born in hospitals
with universal neonatal hearing screening than children with PCHI born in hospitals
without screening, it is unclear whether expressive language is also better. Similarly,
Universal neonatal hearing screening
147
conflicting findings were observed for the communicative abilities of children. Based on
blinded assessment of the infant seen in one study, as opposed to subjective parental
assessment in the other, it appears that some children with bilateral PCHI born in a
screening hospital have superior speech intelligibility.
As universal neonatal hearing screening programs were predominantly introduced in the
mid- to late-1990s, it is unlikely that information on the longer term, but highly relevant,
outcomes (i.e. educational and employment status) will be reported in the peer-reviewed
literature for another decade or so.
Component
Evidence base
A
B
C
D
Excellent
Good
Satisfactory
Poor
one or two level II
studies with low
risk of bias or a
SR/multiple level
III studies with low
risk of bias
some
inconsistency
reflecting genuine
uncertainty around
clinical question
Consistency
Clinical impact
Generalisability
Applicability
substantial
population/s
studied in the
body of evidence
are similar to the
target population
applicable to
Australian
healthcare
context with few
caveats
Cost-effectiveness of universal neonatal hearing screening
The economic questions are whether the value to Australian society of implementing a
universal neonatal hearing screening (UNHS) program is likely to be greater than that of
the current situation, and how widespread the screening coverage should be.
The existing situation varied, and the design of a comprehensive screening system that
will cover all Australian infants remains to be completed.
Evidence from published economic evaluations
As the majority of identified published research examining the cost-effectiveness of
existing or modelled UNHS programs are from an American perspective, the results
obtained can only suggest what might occur under Australian conditions.
The published information up until 2003 on the cost-effectiveness of UNHS is limited
and at times contradictory. For instance, models of short-term cost-effectiveness are not
borne out by evaluations of actual screening programs in the field. Short-term cost148
Universal neonatal hearing screening
effectiveness models of two-stage UNHS protocols appear to be more cost-effective
than one-stage protocols, but this is not supported by published studies of existing
UNHS protocols.
From the literature it can be concluded that, in the short term, costs for the additional
cases identified and diagnosed by UNHS are greater per unit than those of targeted
screening. However, it is inappropriate to determine incremental cost-effectiveness
without considering long-term costs and cost savings.
While it was established that UNHS would be more expensive to operate than a targeted
screening program, taking a societal perspective over the long term suggests that
identifying a larger proportion of hearing-impaired infants at an early stage (ie ≤6 months
of age) would result in cost savings overall. These estimates of long-term cost savings are
based primarily on observational data and expert opinion.
Evidence from the economic model
When an experiment is either not ethical or not feasible, or has simply not yet been
carried out, decision analytic modelling can provide insight into the relationship between
the costs and the outcomes of the intervention. The base case of the model is simulated
with the available data thought best to approximate the true situation. Once its structure
has been developed, the model can then be run with the variables taking alternative
values in order to observe their influence on the decision. Modelling also serves to
indicate where the research needs to be targeted to obtain better information.
The model presented here is populated with an annual Australian birth cohort of 250,000
neonates. Using the highest quality and most representative literature available by 2003
and information provided by Australian stakeholders, transition probabilities were
identified for estimation of the yields and 2003 costs at the stages of screening,
rescreening and diagnostic assessment. These data were used to estimate the expected
final yields of bilateral and unilateral hearing-impaired neonates from three scenarios: a
UNHS program, a targeted program and no formal screening program (ie case finding).
In the short term this model predicted that implementing a two-stage automated auditory
brainstem response (AABR) UNHS program for a cohort of 250,000 newborns would
identify an extra 607 infants with unilateral or bilateral hearing impairment by the age of
6 months, compared to no formal screening program, at an incremental cost of $6–$11
million. Where a targeted screening program was already in place, expanding to a UNHS
program would identify 319 more infants at an incremental cost of $4–$8 million.
In this modelled UNHS program the cost per infant screened was within the range of
values presented in other studies. However, the cost per child identified derived from the
model was much lower than that reported in previous studies. This analysis used
transition probabilities that encompassed the testing and identification of both unilateral
and bilateral hearing impairment in infants and children at all levels of risk, resulting in a
higher yield of infants with permanent childhood hearing impairment (PCHI).
The long-term direct cost savings from the reduced need for special education and
rehabilitation, and the possibility of indirect savings from enhanced productivity in
adulthood, outweigh by an order of magnitude the costs of the actual screening and
diagnosis. Unfortunately, these potential savings are less well researched than the shortterm costs and so the estimates are more uncertain. Nevertheless, the long-term savings
Universal neonatal hearing screening
149
from implementing a UNHS program derived from this model are in agreement with
previous reports.
Regarding screening, increasing the yield of hearing-impaired infants likely to benefit
from early intervention will therefore be more efficient than choosing a screening
protocol merely because it is more cost-effective per infant diagnosed in the short term –
at least while the marginal cost of screening is in the neighbourhood of current figures.
Given the salience of the indirect cost savings in the eventual decision whether or not to
implement and to continue to support a national UNHS screening program, it is
important that more valid and accurate estimates of the indirect costs of hearing and
language skill impairment be obtained for Australian conditions.
These results are influenced by whether or not all infants diagnosed and treated before 6
months of age actually do achieve normal language skills. Cost savings are realised even
when only 25 per cent of children identified by 6 months of age attain a language level
equivalent to that of their hearing peers. Systematic follow-up of these infants is
therefore a high research priority.
The detection and long-term management of PCHI involves public expenditures from
both Commonwealth and State/Territory levels of government, and from both health
and non-health departments. Over the long term, the States/Territories stand to save on
special education and rehabilitation, and the Commonwealth on disability support
pensions. These savings would be realised even when less than 50 per cent of children
identified by 6 months of age attain a language level equivalent to that of their hearing
peers. But to obtain these substantial long-term savings will require expenditure now on
implementing a UNHS program and a decision as to how the different levels of
government will split the responsibility for the funding.
Overall assessment of universal neonatal hearing screening
Universal neonatal hearing screening has been assessed according to the criteria outlined
in Box 2 of the ‘Approach to Assessment’ section of this report. The overall assessment
of the viability, effectiveness and appropriateness of a universal neonatal screening
program is given in Box 9 below.
150
Universal neonatal hearing screening
Box 9
Criteria for appraising the viability, effectiveness and appropriateness of a screening program
(UK National Screening Committee 2000)
The condition
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.1 The condition should be an important health problem.
Yes
1.2 The epidemiology and natural history of the condition, including
development from latent to declared disease, should be adequately
understood and there should be a detectable risk factor, or disease
marker and a latent period or early symptomatic stage.
Yes
1.3 All the cost-effective primary prevention interventions should have been
implemented as far as practicable.
Not applicable
The test
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.4 There should be a simple, safe, precise and validated screening test.
Yes but low Positive Predictive
Value
1.5 The distribution of test values in the target population should be known
and a suitable cut-off level defined and agreed.
Yes
1.6 The test should be acceptable to the population.
Yes
1.7 There should be an agreed policy on the further diagnostic
investigation of individuals with a positive test result and on the choices
available to those individuals.
Yes
The treatment
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.8 There should be an effective treatment or intervention for patients
identified through early detection, with evidence of early treatment
leading to better outcomes than late treatment.
Limited evidence
1.9 There should be agreed evidence based policies covering which
individuals should be offered treatment and the appropriate treatment
to be offered.
No evidence-based data
1.10 Clinical management of the condition and patient outcomes should be
optimised by all health care providers prior to participation in a
screening program.
Yes
The screening program
Evidence-Based Decision
Yes/No/Not applicable/Comment
1.11 There must be evidence from high quality Randomised Controlled Trials
that the screening program is effective in reducing mortality or
morbidity.
No
1.12 Where screening is aimed solely at providing information to allow the
person being screened to make an “informed choice” (e.g. Down
syndrome, cystic fibrosis carrier screening), there must be evidence
from high quality trials that the test accurately measures risk. The
information that is provided about the test and its outcome must be of
value and readily understood by the individual being screened.
Not applicable
1.13 There should be evidence that the complete screening program (test,
diagnostic procedures, treatment/intervention) is clinically, socially and
ethically acceptable to health professionals and the public.
Yes
1.14 The benefit from the screening program should outweigh the physical
and psychological harm (caused by the test, diagnostic procedures and
treatment).
Insufficient evidence of benefit
but no evidence of lasting harm
Universal neonatal hearing screening
151
Box 9
Criteria for appraising the viability, effectiveness and appropriateness of a screening
program (UK National Screening Committee 2000)
The screening program (cont.)
Evidence-Based Decision
Yes/No/Not applicable/Comment
152
1.15 The opportunity cost of the screening program (including testing,
diagnosis, treatment, administration, training and quality assurance)
should be economically balanced in relation to expenditure on medical
care as a whole (i.e. value for money).
To be determined
1.16 There must be a plan for managing and monitoring the screening
program and an agreed set of quality assurance standards.
To be determined
1.17 Adequate staffing and facilities for testing, diagnosis, treatment and
program management should be made available prior to the
commencement of the screening program.
To be determined
1.18 All other options for managing the condition should have been
considered (e.g. improving treatment, providing other services), to
ensure that no more cost effective intervention could be introduced or
current interventions increased within the resources available.
Assuming evidence of benefit,
then considerable potential costsavings in the long term
1.19 Evidence-based information, explaining the consequence of testing,
investigation and treatment, should be made available to potential
participants to assist them in making an informed choice.
To be determined
1.20 Public pressure for widening the eligibility criteria, for reducing the
screening interval, and for increasing the sensitivity of the testing
process, should be anticipated. Decisions about these parameters
should be scientifically justifiable to the public.
Not applicable – all babies
screened at the earliest possible
time
Universal neonatal hearing screening
Conclusions
Clinical need
Median estimates from the literature suggest that the prevalence of bilateral and unilateral
PCHI >35 dB HL is 1.3 and 0.6 per 1,000 infants respectively. Using this estimate for
Australia it is calculated that 481 children may be born with bilateral or unilateral PCHI
each year. The predicted yield of unilateral and bilateral PCHI from an Australian UNHS
program, using two-stage AABR screening, was estimated in this report to be 607
infants.
Safety
There was no evidence available up until 2007 that reported physical harms from UNHS.
Limited poor to average quality studies suggested that there was a slight increase in
anxiety and depression in mothers if their infant received a negative screening test, but
these states remained within the normal range, and as such no clinically important
differences were identified. UNHS was found to cause no more psychosocial distress
than a behavioural test given after the infant is 6 months old, but was associated with
increased satisfaction levels. Based on the literature available it is therefore concluded
that UNHS does not cause psychosocial harm, although no data were found on the
harms caused by false reassurance.
Effectiveness
Diagnostic accuracy
Under quiet conditions, TEOAE possesses excellent sensitivity (up to 100%) and good
specificity (92%) for diagnosing PCHI. The positive predictive value of TEOAE is poor
however at 1.5 per cent. In comparison, AABR has excellent specificity (96%) and good
sensitivity (80%). The positive predictive value of AABR is very low (2.2%), although
marginally better than TEOAE conducted under quiet conditions.
Screening
While there is a lack of level I and II screening evidence, findings from one good quality
level III-1 study suggests that referral for definitive diagnostic testing, actual PCHI
diagnosis, and management of PCHI commonly occurs earlier and more frequently with
universal neonatal hearing screening (UNHS) than without it. This is supported by level
III-2 average and good quality studies.
The effect of universal neonatal hearing screening on primary or patient-relevant
outcomes is not readily apparent. Two cohort studies (Level III-2 screening evidence)
were available that measured linguistic and communicative abilities quite differently.
From the evidence, language acquisition improvements are seen, for receptive language
but unclear findings have been reported for expressive language. Similarly, the impact of
universal hearing screening on communicative abilities in children are as yet unclear as
the two available retrospective cohort studies reported contradictory results.
Universal neonatal hearing screening
153
Economic considerations
From the available literature it can be concluded that, in the short term, the costs for the
additional cases identified and diagnosed by UNHS are greater per unit than those of
targeted screening. However, taking a societal perspective over the long term suggests
that identifying a larger proportion of hearing-impaired infants at an early stage (ie ≤6
months of age) would result in a cost saving overall. The validity of these estimates of
long-term cost savings should be regarded with caution as they are based primarily on
observational data and expert opinion.
In the short term, the decision analytic model presented in this report predicts that
implementing a two-stage automated auditory brainstem response (AABR) universal
neonatal hearing screening (UNHS) program for a cohort of 250,000 newborns would
identify an extra 607 infants with unilateral or bilateral hearing impairment by the age of
6 months compared to no formal screening program, at an incremental cost of $6–$11
million. Where a targeted screening program is already in place, expanding to a universal
screening program would identify 319 more infants, at an incremental cost of $4–$8
million. These figures were reported in 2003 Australian dollars as only the literature
review for UNHS has been updated until 2007.
The long-term direct cost savings from the reduced need for special education and
rehabilitation and the possibility of indirect savings from enhanced productivity in
adulthood outweigh by an order of magnitude the costs of the actual screening and
diagnosis. Unfortunately, these potential savings are less well researched than the shortterm costs and so the estimates are more uncertain. Nevertheless, the long-term savings
from implementing a UNHS program derived from this model are in general agreement
with previous reports.
Over the long term, the States/Territories stand to save on special education and
rehabilitation, and the Federal Government to save on disability support pensions.
Research implications
Systematic reviews, by their nature, often identify areas in the evidence-base that are
deficient or incomplete and that require further research before a definite answer can be
ascertained. Areas that would benefit from further research are outlined below:
154
•
Primary research is needed to assess the diagnostic accuracy of screening tools in a
healthy ‘well baby’ population, particularly for the automated auditory brainstem
response (AABR) test, compared to established reference standards. The diagnostic
accuracy of newer models/versions of these screening tests should also be
determined.
•
Primary research on screen accuracy, especially a determination of the number of
false negatives from different types of screening protocols, is required. False
negatives should be framed in terms of congenital or perinatal hearing impairment
rather than acquired losses.
•
Evidence-based research is lacking regarding the identification of effective
interventions or methods to ensure adequate follow-up of parents and babies for
hearing rescreen or diagnostic appointments.
Universal neonatal hearing screening
•
Primary research (ie trials) would be of benefit to determine the effect of screening
on patient-relevant outcomes associated with moderate to severe bilateral permanent
childhood hearing impairment (PCHI). These include the further information on the
impact of UNHS on development of speech, language and communication, quality
of life and quality of family life including social, emotional and behavioural aspects.
These outcome measures should be objectively assessed using valid and reliable tools
and blinded to the screening or management status of the infant.
•
Primary research, as trials or cohort studies, is needed to assess the effect of
screening or early intervention (<6 months of age) on longer term outcomes such as
educational performance, employment status and quality of life.
•
Primary research (ie trials) would be of benefit to evaluate the effectiveness of
management strategies (ie amplification) compared to no management, and of early
versus late management, for mild or unilateral hearing losses.
•
Agreed evidence-based policies or clinical practice guidelines regarding which
individuals should be offered management, as well as the appropriate management to
be offered, are currently lacking.
•
Primary research would be useful to assess the psychological effects on the caregiver
and long-term consequences to the infant of false reassurance (regarding ‘normal’
infant hearing status) given through a UNHS program.
•
Primary research to evaluate the psychological effects on the caregiver of an infant
receiving a true-positive diagnosis, ascertained through UNHS, would also be of
value.
Implementation issues
The available evidence-base also provides insight into how screening interventions
should or should not be conducted and what issues need to be addressed. The issues
discussed below may assist with the development of universal neonatal hearing screening
programs:
•
Given the likely low sensitivity of transient evoked otoacoustic emissions (TEOAE)
testing under ‘real world’ noise conditions, and the concomitant large proportion of
infants with PCHI who may not be identified (false negatives) under these
conditions, TEOAE testing should only occur in environments that are quiet or even
sound-proofed. [Diagnostic level of evidence: III]
•
Given the low positive predictive value of both screening tests, parents of well babies
that fail the screen should be counselled in terms of a ‘refer’ for a rescreen or
diagnostic testing, rather than as a ‘failed’ test. It should be emphasised that it is
considerably more likely (>95% after an initial screen) for the positive test to be a
‘false’ positive rather than a true positive. [Diagnostic level of evidence: III]
•
As a consequence of the low positive predictive value of the screening tests, two- or
three-stage well baby screening programs should be introduced to improve the
predictive value of the result and lessen the burden on diagnostic services. If a one-
Universal neonatal hearing screening
155
stage screen is the only possible option, then the AABR has slightly better predictive
value. [Diagnostic level of evidence: III]
156
•
If two- or three-stage programs are introduced, delays between screening, rescreening
and diagnostic testing should be minimised wherever possible to reduce parental
anxiety and losses to follow-up, and to ensure early management of the condition
(assuming early management is more beneficial) and the smooth coordination of
services. [Intervention level of evidence: III and IV]
•
Outpatient or community-based programs should ensure that neonatal hearing
screening is ‘piggy-backed’ onto other health or immunisation checks, in order to
ensure good screen coverage. Good coverage can also be obtained if screening is
delivered in the home (eg mobile services). This may have particular relevance for
rural areas where accessibility and limited resources are often a problem.
[Intervention level of evidence: IV]
•
Community-based screening programs should ensure comprehensive follow-up of
re-screen and diagnostic referrals through reminder and parent education strategies.
Hospital-based screening programs should ensure similar strategies are in place for
follow-up of outpatient referrals. This is particularly important for programs
conducted in economically-depressed areas or with populations that are highly
mobile, indigent or non-English speaking. [Intervention level of evidence: IV]
Universal neonatal hearing screening
Recommendation
MSAC recommended that on the strength of evidence pertaining to <application name>
public funding <should/should not> be supported for this procedure.
- The Minister for Health and Ageing endorsed/did not endorse this recommendation on
<date>… OR
Since there is currently insufficient evidence pertaining to <application name>, MSAC
recommended that public funding should not be supported at this time for this
procedure.
- The Minister for Health and Ageing endorsed/did not endorse this recommendation on
<date> -
Universal neonatal hearing screening
157
Appendix A
MSAC terms of reference
and membership
MSAC's terms of reference are to:
•
advise the Minister for Health and Ageing on the strength of evidence pertaining
to new and emerging medical technologies and procedures in relation to their
safety, effectiveness and cost-effectiveness and under what circumstances public
funding should be supported;
•
advise the Minister for Health and Ageing on which new medical technologies
and procedures should be funded on an interim basis to allow data to be
assembled to determine their safety, effectiveness and cost-effectiveness;
•
advise the Minister for Health and Ageing on references related either to new
and/or existing medical technologies and procedures; and
•
undertake health technology assessment work referred by the Australian Health
Ministers’ Advisory Council (AHMAC) and report its findings to AHMAC.
The membership of MSAC comprises a mix of clinical expertise covering pathology,
nuclear medicine, surgery, specialist medicine and general practice, plus clinical
epidemiology and clinical trials, health economics, consumers, and health administration
and planning:
MSAC membership who evaluated the initial reference (2003)
158
Member
Expertise or affiliation
Dr Stephen Blamey (Chair)
general surgery
Associate Professor John Atherton
cardiology
Professor Bruce Barraclough
general surgery
Professor Syd Bell
pathology
Dr Michael Cleary
emergency medicine
Dr Paul Craft
clinical epidemiology and oncology
Dr Gerry FitzGerald
Australian Health Ministers’ Advisory Council representative
Dr Kwun Fong
thoracic surgery
Professor Jane Hall
health economics
Dr Terri Jackson
health economics
Professor Brendon Kearney
health administration and planning
Associate Professor Richard King
internal medicine
Dr Ray Kirk
health research
Dr Michael Kitchener
nuclear medicine
Dr Ewa Piejko
general practice
Ms Sheila Rimmer
consumer health issues
Universal neonatal hearing screening
Dr Jeffrey Robinson
obstetrics and gynaecology
Professor Bryant Stokes
neurological surgery
Professor Ken Thomson
radiology
Dr Douglas Travis
urology
Current MSAC membership (2007)
Member
Expertise or Affiliation
Dr Stephen Blamey (Chair)
general surgery
Associate Professor John Atherton
cardiology
Associate Professor Michael Cleary
emergency medicine
Associate Professor Paul Craft
clinical epidemiology and oncology
Professor Geoff Farrell
gastroenterology
Dr Kwun Fong
thoracic medicine
Professor Richard Fox
medical oncology
Dr David Gillespie
gastroenterology
Dr Bill Glasson
ophthalmologist
Professor Jane Hall
health economics
Professor John Horvath
Chief Medical Officer, Department of Health and Ageing
Associate Professor Terri Jackson
health economics
Professor Brendon Kearney
health administration and planning
Associate Professor Frederick Khafagi nuclear medicine
Dr Ray Kirk
health research
Dr Ewa Piejko
general practice
Dr Ian Prosser
haematology
Ms Sheila Rimmer
consumer health issues
Dr Judy Soper
radiology
Professor Ken Thomson
radiology
Dr Mary Turner
Australian Health Ministers’ Advisory Council
representative
Dr David Wood
orthopaedics
Mr Peter Woodley
Assistant Secretary, Medical Benefits Schedule (MBS) Policy
Development Branch, Department of Health and Ageing
Universal neonatal hearing screening
159
Appendix B
Advisory Panel, Evaluator
and Project Manager
Advisory panel for MSAC reference 17, Universal Neonatal Hearing
Screening (2003)
Chair
Professor Bryant Stokes AM
MBBS, FRACS, FRCS
Neurosurgeon
Perth, Western Australia
member of MSAC
Evaluators (AHTA)
Ms Tracy Merlin, Lead Researcher and Manager
Mr Brent Hodgkinson, Research Officer
Dr Petra Bywood, Research Officer
Ms Fiona Jenner, Research Assistant
Mr John Moss, Health Economist
Prof Janet Hiller, Director
Panel Members
A/Professor Harvey Coates
Co-opted member
MS (Otol), DABO, FRCS(C) (Otol), FRACS, FACS
Senior ENT Surgeon,
Princess Margaret Hospital for Children, Perth
Clinical A/Professor, School of Paediatrics and Child
Health, and School of Surgery and Pathology, University of
Western Australia
160
Dr Jill Duncan
PhD (Applied Linguistics), MEd (Management), MEd
(Hearing Impairment), BSci (Speech Pathology and
Audiology), CED, Cert AVT
Director, Cora Barclay Centre
Gilberton, South Australia
Co-opted deaf educator –
paediatrics and child health
Professor Jane Hall
BA, PhD
Director, Centre for Health Economics Research and
Evaluation
Professor of Health Economics, Faculty of Business,
University of Technology, Sydney
Honorary Professor, Faculty of Medicine, University of
Sydney
MSAC member
Universal neonatal hearing screening
Mrs Marion Maurer
B Nurs, PG Dip Aud, MAudSA (CCP)
Audiologist (ret. 9/2003), Mater Infant Hearing Program
Mater Mothers Hospital
Brisbane, Queensland
Co-opted audiologist
Mr Daniel McAullay
Research coordinator
Kulunga Research Network
Telethon Institute for Child Health Research
Subiaco, Western Australia
Co-opted epidemiologist
Dr Fiona Panizza
MBBS(Hons), FRACS,
VMO
Paediatric Otolaryngologist
Mater Children’s Hospital
Brisbane, Queensland
Co-opted otorhinolaryngologist
A/Professor Melissa Wake
MB ChB, FRACP, MD, Grad Dip Epi
Director, Research and Public Health
Centre for Community Child Health
Royal Children’s Hospital
Parkville, Victoria
Co-opted paediatrician
Nominee of the Consumer
Ms Diane Walsh
Health Forum
BA, Dip Ed
Member – Governing Committee, Consumer Health Forum
Public Member – Medical Board of the Northern Territory
Evaluators for update of MSAC reference 17, Universal Neonatal Hearing
Screening (2007)
Evaluators (AHTA)
Ms Tracy Merlin, Lead Researcher and Manager
Ms Hedyeh Hedayati, Research Officer
Mr Thomas Sullivan, Research Officer
Ms Skye Newton, Research Officer
Ms Liz Buckley, Research Officer
Prof Janet Hiller, Director
Universal neonatal hearing screening
161
Appendix C
Search strategies
Bibliographic databases used to identify literature
Electronic database
Time Period
AustHealth (Informit)
1997 – 08/2007
Australian Medical Index (Informit)
1996 – 08/2007
Australian Public Affairs Information Service (APAIS) – Health (Informit)
1990 – 08/2007
Cinahl (Silverplatter)
1977 – 08/2007
Cochrane Library – including, Cochrane Database of Systematic Reviews,
Database of Abstracts of Reviews of Effects, the Cochrane Central
Register of Controlled Trials (CENTRAL), the Health Technology
Assessment Database, the NHS Economic Evaluation Database
1966 – 08/2007
Current Contents Connect (ISI)
1993 – 08/2007
Embase (Embase.com)
1974 – 08/2007
Pre-Medline and Medline (PubMed)
1966 – 08/2007
ProceedingsFirst
1993 – 08/2007
PsycInfo (Silverplatter)
1983 – 08/2007
Web of Science – Science Citation Index Expanded (ISI)
1995 – 08/2007
EconLit
1969 – 08/2007
Other sources of evidence (1966 – 08/2007)
Source
Location
Internet
NHMRC – National Health and Medical Research Council (Australia)
http://www.health.gov.au/nhmrc/
Australian Department of Health and Ageing
http://www.health.gov.au/
US Department of Health and Human Services (reports and publications)
http://www.os.dhhs.gov/
New York Academy of Medicine Grey Literature Report
http://www.nyam.org/library/greylit/index.shtml
Scirus – for Scientific Information Only
http://www.scirus.com
Trip database
http://www.tripdatabase.com
Current Controlled Trials metaRegister
http://controlled-trials.com/
Health Technology Assessment International (HTAi)
http://www.htai.org/
International Network for Agencies for Health Technology Assessment
http://www.inahta.org/
National Library of Medicine Health Services/Technology Assessment
Text
http://text.nlm.nih.gov/
National Library of Medicine Locator Plus database
http://locatorplus.gov
U.K. National Research Register
http://www.update-software.com/National/
Websites of Health Technology Agencies
See Appendix D
Websites of Hearing Organisations
See Appendix D
Hand searching (journals 2006–2007)
162
American Journal of Otolaryngology
Library or electronic access
Archives of Otolaryngology – head and neck surgery
Library or electronic access
Clinical Otolaryngology
Library or electronic access
Current Opinion in Otolaryngology
Library or electronic access
Ear and Hearing
Library or electronic access
Hearing Research
Library or electronic access
International Journal of Pediatric Otorhinolaryngology
Library or electronic access
Universal neonatal hearing screening
Journal of Speech, Language and Hearing Research
Library or electronic access
Journal of the Association for Research in Otolaryngology
Library or electronic access
The Hearing Journal
Library or electronic access
The Journal of Otolaryngology
Library or electronic access
The Otolaryngologic Clinics of North America
Library or electronic access
Expert clinicians
Studies other than those found in regular searches
MSAC Advisory Panel
Pearling
All included articles had their reference lists searched for additional
relevant source material
Search terms utilised
All searches
MeSH:
Neonatology; Infant; Child; Hearing Disorders; Hearing Impaired Persons
Text words in title or abstract:
neonat*; infant*; baby; child*; hear*; deaf*
Prevalence
MeSH:
Morbidity; Epidemiologic Studies
Text words in title or abstract:
frequenc*; proportion*; prevalen*; inciden*; rate
Limits:
Human; not “heart”; English language; 1980 Diagnostic accuracy
MeSH:
Diagnostic Techniques, Otological; Sensitivity and Specificity; Diagnostic Errors; Mass Screening
Text words in title or abstract:
screen*; test*; universal; sensitiv*, specific*; false negative; false positive; predictive value*; accuracy; likelihood ratio*;
AABR; auditory brain stem response*; TEOAE; OAE; oto?acoustic emission*
Limits:
Human; not “heart”; 1980 Safety, effectiveness and cost-effectiveness of screening
MeSH:
Mass-Screening; Clinical Trials; Epidemiologic Studies
Text words in title or abstract:
screen*; test*; universal; AABR; auditory brain stem response; TEOAE; OAE; oto?acoustic emission*
Limits:
Human; not “heart”; 1980 Safety and effectiveness of early diagnosis
MeSH:
Cohort-studies; Incidence; Prognosis; Communication; Language Development Disorders
Text words in title or abstract:
inciden*; cohort*; registry; register; prognosis; language*; communicat*; speech; referral*; diagnos*; lip?read*; signing; sign
language
Limits:
Human; not “heart”
Safety and effectiveness of early management
MeSH:
Clinical Trials; Epidemiologic Studies; Hearing Aids; Otologic Surgical Procedures; Rehabilitation of Hearing Impaired;
Universal neonatal hearing screening
163
Language Development; Communication
Text words in title or abstract:
hearing aid*; cochlear implant*; rehabilitation; intervent*; program*; language*; communicat*; speech
Limits:
Human; not “heart”
164
Universal neonatal hearing screening
Appendix D Internet sites searched
Websites of health technology assessment groups
AUSTRALIA
•
Australian Safety and Efficacy Register of New Interventional Procedures –
Surgical (ASERNIP-S) http://www.surgeons.org/open/asernip-s.htm
•
Centre for Clinical Effectiveness, Monash University
http://www.med.monash.edu.au/healthservices/cce/evidence/
•
Health Economics Unit, Monash University http://chpe.buseco.monash.edu.au
AUSTRIA
•
Institute of Technology Assessment / HTA unit
http://www.oeaw.ac.at/ita/e1-3.htm
CANADA
•
Agence d’Evaluation des Technologies et des Modes d’Intervention en Santé
(AETMIS) http://www.aetmis.gouv.qc.ca/en/index.htm
•
Alberta Heritage Foundation for Medical Research (AHFMR)
http://www.ahfmr.ab.ca/publications.html
•
Canadian Coordinating Office for Health Technology Assessment (CCDTH)
http://www.cadth.ca/index.php/en/media-centre/2003/03/5/38
•
Canadian Health Economics Research Association (CHERA/ACRES) – Cabot
database http://www.mycabot.ca
•
Centre for Health Economics and Policy Analysis (CHEPA), McMaster
University http://www.chepa.org
•
Centre for Health Services and Policy Research (CHSPR), University of British
Columbi http://www.chspr.ubc.ca
•
Health Utilities Index (HUI) http://www.fhs.mcmaster.ca/hug/index.htm
•
Institute for Clinical and Evaluative Studies (ICES) http://www.ices.on.ca
DENMARK
•
Danish Institute for Health Technology Assessment (DIHTA)
http://www.dihta.dk/publikationer/index_uk.asp
FINLAND
•
FINOHTA http://www.stakes.fi/finohta/e/
Universal neonatal hearing screening
165
FRANCE
•
L’Agence Nationale d’Accréditation et d’Evaluation en Santé (ANAES)
http://www.anaes.fr/
GERMANY
•
German Institute for Medical Documentation and Information (DIMDI) / HTA
http://www.dahta.dimdi.de/
•
German Scientific Working Group of Technology Assessment
http://www.epi.mh-hannover.de/(eng)/hta.html
THE NETHERLANDS
•
Health Council of the Netherlands Gezondheidsraad
http://www.gr.nl/engels/welcome/frameset.htm
NEW ZEALAND
•
New Zealand Health Technology Assessment (NZHTA)
http://nzhta.chmeds.ac.nz/
NORWAY
•
Norwegian Centre for Health Technology Assessment (SMM)
http://www.oslo.sintef.no/smm/Publications/Engsmdrag/FramesetPublication
s.htm
SPAIN
•
Agencia de Evaluación de Tecnologias Sanitarias, Instituto de Salud “Carlos
III”I/Health Technology Assessment Agency (AETS)
http://www.isciii.es/aets/cdoc.htm
•
Catalan Agency for Health Technology Assessment (CAHTA)
http://www.aatm.es/cgi-bin/frame.pl/ang/pu.html
SWEDEN
•
Swedish Council on Technology Assessment in Health Care (SBU)
http://www.sbu.se/admin/index.asp
SWITZERLAND
•
Swiss Network on Health Technology Assessment (SNHTA)
http://www.snhta.ch/
UNITED KINGDOM
166
•
Health Technology Board for Scotland http://www.htbs.org.uk/
•
National Health Service Health Technology Assessment (UK) / National
Coordinating Centre for Health Technology Assessment (NCCHTA)
http://www.hta.nhsweb.nhs.uk/
•
University of York NHS Centre for Reviews and Dissemination (NHS CRD)
http://www.york.ac.uk/inst/crd/
Universal neonatal hearing screening
•
National Institute for Clinical Excellence (NICE)
http://www.nice.org.uk/index.htm
UNITED STATES
•
Agency for Healthcare Research and Quality (AHRQ)
http://www.ahrq.gov/clinic/techix.htm
•
Harvard Center for Risk Analysis – Cost-Utility Analysis Database Project
http://www.hcra.harvard.edu/tablesdata.html
•
U.S. Blue Cross / Blue Shield Association Technology Evaluation Center
(TEC) http://www.bcbs.com/consumertec/index.html
•
U.S. Dept. of Veterans Affairs Technology Assessment Program (VATAP)
http://www.va.gov/resdev/prt/pubs_individual.cfm?webpage=pubs_ta_reports.
htm
Universal neonatal hearing screening
167
Websites of relevant hearing organisations
Advanced Hearing Research Center http://ahrc.utdallas.edu/research_labs/pah.html
American Academy of Audiology http://www.audiology.org
American Society for Deaf Children http://www.deafchildren.org
American Society of Pediatric Otolaryngology http://www.aspo.us
American Speech-Language-Hearing Associates http://www.asha.org
Association for Research in Otolaryngology http://www.aro.org
Audiological Society of Australia http://www.audiology.asn.au
Audiology Foundation of America http://www.audfound.org
AudiologyNet http://www.audiologynet.com
Australian Hearing http://www.hearing.com.au
Canadian Hearing Society http://www.chs.ca
Centers for Disease Control and Prevention http://www.cdc.gov/ncbddd/ehdi
Child and Youth Health http://www.cyh.com
Deafness at Birth http://www.deafnessatbirth.org.uk
Defeating Deafness http://www.defeatingdeafness.org
HearingExchange http://www.hearingexchange.com
Hearing Concern http://www.hearingconcern.com
Hearing Review http://www.hearingreview.com
Hear It http://www.hear-it.org
Institute for Hearing Research http://www.ihr.mrc.ac.uk
International Federation of the Hard of Hearing http://www.ifhoh.org
Marion Downs National Center For Infant Hearing
http://www.colorado.edu/slhs/mdnc
National Center for Hearing Assessment and Management
http://www.infanthearing.org
National Newborn Screening and Genetics Resource Center
http://www.genes-r-us.uthscsa.edu
168
Universal neonatal hearing screening
New Zealand Audiological Society http://www.audiology.org.nz
Office of Hearing Services http://www.health.gov.au/hear
Pennsylvania Academy of Audiology http://www.paaudiology.org
Searchwave http://www.searchwave.com
Sound Beginnings http://www.kdhe.state.ks.us/sb
The British Society for Audiology http://www.b-s-a.demon.co.uk
Universal neonatal hearing screening
169
Appendix E
Critical appraisal checklists
Randomised controlled trial appraisal checklist
Source: (NHMRC 2000a)
1. Method of treatment assignment
a. Correct, blinded randomisation method described OR randomised, double-blind method
stated AND group similarity documented
b. Blinding and randomisation stated but method not described OR suspect technique (eg
allocation by drawing from an envelope)
c. Randomisation claimed but not described and investigator not blinded
d. Randomisation not mentioned
2. Control of selection bias after treatment assignment
a. Intention to treat analysis AND full follow-up
b. Intention to treat analysis AND <15% loss to follow-up
c. Analysis by treatment received only OR no mention of withdrawals
d. Analysis by treatment received AND no mention of withdrawals OR more than 15%
withdrawals/loss-to-follow-up/post-randomisation exclusions
3. Blinding
a. Blinding of outcome assessor AND patient and care giver
b. Blinding of outcome assessor OR patient and care giver
c. Blinding not done
4. Outcome assessment (if blinding was not possible)
a. All patients had standardised assessment
b. No standardised assessment OR not mentioned
Cohort study appraisal checklist
Source: (NHMRC 2000b)
1. How were subjects selected for the ‘new intervention’?
2. How were subjects selected for the comparison or control group?
3. Does the study adequately control for demographic characteristics, clinical features and other
potential confounding variables in the design or analysis?
4. Was the measurement of outcomes unbiased (ie blinded to treatment group and comparable
across groups)?
5. Was follow-up long enough for outcomes to occur?
6. Was follow-up complete and were there exclusions from the analysis?
Total
170
/6
Universal neonatal hearing screening
Case-control studies appraisal checklist
Source: (NHMRC 2000b)
1. How were subjects selected for the ‘new intervention’?
2. How were subjects selected for the comparison or control group?
3. Does the study adequately control for demographic characteristics, clinical features and other
potential confounding variables in the design or analysis?
4. Was the measurement of outcomes unbiased (ie blinded to treatment group and comparable
across groups)?
5. Was follow-up long enough for outcomes to occur?
6. Was follow-up complete and were there exclusions from the analysis?
Total
/6
Case series appraisal checklist
Source: (NHS Centre for Reviews and Dissemination 2001) Amended for this review (amendments
italicised)
1. Is the study based on a representative sample selected from a relevant population? (consecutive
and not significantly different than general population of neonates)
2. Are the criteria for inclusion explicit?
3. Did all individuals enter the survey at a similar point in their disease progression? (were the
neonates at a similar gestational age)
4. Was follow-up long enough for important events to occur and was follow-up adequate? (greater than
80%)
5. Were outcomes assessed using objective criteria or was blinding used?
Total
Universal neonatal hearing screening
/5
171
Diagnostic accuracy study appraisal checklist
Source: (Whiting P 2003)
Item
1. Was the spectrum of patients representative of the patients
who will receive the test in practice?
2. Were selection criteria clearly described?
Yes
()
No
()
Unclear
()
()
()
()
()
()
()
()
()
()
5. Did the whole sample or a random selection of the sample,
receive verification using a reference standard of diagnosis?
()
()
()
6. Did patients receive the same reference standard regardless
of the index test result?
7. Was the reference standard independent of the index test
(i.e. the index test did not form part of the reference
standard)?
8. Was the execution of the index test described in sufficient
detail to permit replication of the test?
()
()
()
()
()
()
()
()
()
9. Was the execution of the reference standard described in
sufficient detail to permit its replication?
()
()
()
10. Were the index test results interpreted without knowledge of
the results of the reference standard?
()
()
()
11. Were the reference standard results interpreted without
knowledge of the results of the index test?
()
()
()
12. Were the same clinical data available when test results were
interpreted as would be available when the test is used in
practice?
13. Were uninterpretable/ intermediate test results reported?
()
()
()
()
()
()
14. Were withdrawals from the study explained?
()
()
()
3. Is the reference standard likely to correctly classify the target
condition?
4. Is the time period between reference standard and index test
short enough to be reasonably sure that the target condition
did not change between the two tests?
Total
172
/14
Universal neonatal hearing screening
Checklist for appraising economic evaluation studies
Source: (NHMRC 2001)
Title of assessment:
Title of study:
Author(s):
Year:
Comparators:
Score :
/16
Appraisal items for internal validity
1. Was the study question well defined?
2. Were appropriate health care options chosen and clearly described?
3. Was an appropriate study type used?
4. Was the effectiveness of the health care options established?
5. Were the cost estimates related to baseline population risk?
6. Were all the relevant costs and consequences identified for each health care option?
7. Was differential timing considered?
8. Was an incremental analysis performed?
9. Was a sensitivity analysis performed?
10. Were modelling techniques used in a clear and reasonable way?
Criteria for assessing the generalisability of economic evaluation studies
11. Patient group
12. Health system setting
13. Health care option
14. Resource costs
15. Marginal versus average cost
16. Other specific issues
Universal neonatal hearing screening
173
Rank scoring for appraising the clinical importance of benefit/harm
Source: (NHMRC 2000b)
Title of review:
Title of study:
Author(s):
Year:
Comparators:
Clinically important effect:
Rank Score :
174
/4
Ranking
Clinical importance of benefit/harm
1
A clinically important benefit for the full range of plausible estimates.
The confidence limit closest to the measure of no effect (the ‘null’) rules out a clinically unimportant
effect of the intervention.
2
The point estimate of effect is clinically important BUT the confidence interval includes clinically
unimportant effects.
3
The confidence interval does not include any clinically important effects.
4
The range of estimates defined by the confidence interval includes clinically important effects BUT
the range of estimates defined by the confidence interval is also compatible with no effect, or a
harmful effect.
Universal neonatal hearing screening
Rank scoring for classifying the relevance of evidence
Source: (NHMRC 2000b)
Title of review:
Title of study:
Author(s):
Year:
Comparators:
Rank Score :
/5
Ranking
Relevance of the evidence
1
Evidence of an effect on patient-relevant clinical outcomes, including benefits and harms, and quality of
life and survival.
2
Evidence of an effect on a surrogate outcome that has been shown to be predictive of patient-relevant
outcomes for the same intervention.
3
Evidence of an effect on proven surrogate outcomes but for a different intervention.
4
Evidence of an effect on proven surrogate outcomes but for a different intervention and population.
5
Evidence confined to unproven surrogate outcomes.
Universal neonatal hearing screening
175
Appendix F
Studies included in the
review
Included studies on prevalence of permanent childhood
hearing impairment
Study
Location
Study design
Study population
Intervention
Prevalence
1,421/1,727 neonates
from the hospital
WBN b born between
October 1995 and
March 1997
1. Initial screen
with TEOAE c
2. Rescreen of
failures with
TEOAE
3. Re-test of
failures with
audiological
assessment
Sensorineural hearing
loss
>40 dB = 2/1,421
1.41/1,000
12,708 well and atrisk babies screened
in five hospitals from
a population of
13,214 born between
February 2000 and
June 2001
1. Initial screen
with TEOAE
2. Rescreen of
failures with
AABRd
3. Rescreen of
failures with
TEOAE ± AABR
4. Re-test of failures
with diagnostic
ABRe
Bilateral PCHI f
>35 dB = 9/12,708
0.71/1,000
17,602 well and atrisk babies who
received neonatal
care at the University
of Mississippi
between January
1997 and January
2002. 14,408 well
babies (81.9%) and
3,194 at-risk babies
(18.1%)
Well babies:
1. Initial screen
with AABR
2. Rescreen of
failures with AABR
3. Rescreen of
stage 2 failures
with repeated
AABR.
4. Re-test of
failures with
diagnostic ABR
At-risk babies:
1. Initial screen
with AABR
2. Rescreen of
failures with
repeated AABR
3. Re-test of
failures with
diagnostic ABR.
All other high risk
patients referred
for audiologic
follow-up
Sensorineural or
conductive hearing loss
≥35 dB = 78/17,602
4.43/1,000
Ascertainment via UNHS a – well babies only
(Aidan et al
1999)
France
Case series –
hospital based
Ascertainment via UNHS – well and at-risk babies
(Bailey et al
2002)
(Connolly et
al 2005)
176
Perth,
Australia
Mississippi,
USA
Multi-centre (5)
case series
Case series –
hospital based
Unilateral PCHI
>35 dB = 3/12,708
0.24/1,000
Universal neonatal hearing screening
Study
Location
Study design
Study population
Intervention
Prevalence
(Kennedy et
al 2005)
UK,
Wessex
trial
Multi-centre (4)
quasirandomised
controlled trial.
53,781 live births
between October
1993 and 1996 in 4
different hospitals
25,609 born during
periods of UNHS and
21,279 screened.
28172 born during
periods without
UNHS
During periods of
UNHS:
1. Initial screen with
TEOAE
2. Rescreen of
failures with AABR
3. Re-test of failures
with audiological
assessment
Group born during
period of UNHS:
Bilateral PCHI
>40 dB = 31/25,609
1.21/1,000
During periods
without UNHS:
1. Distraction test
2. Re-test of failures
with audiological
assessment
Group born during
period without UNHS:
Bilateral PCHI
>40 dB = 35/28,172
1.24/1,000
Entire sample:
Bilateral PCHI
>40 dB = 66/53,781
1.23/1,000
Details of PCHI
cases not initially
identified by UNHS
or following the
distraction test were
obtained from case
records and
outpatient lists from
local audiology
services
(Martines et
al 2007)
Sicily, Italy
Case series –
hospital based
1,191 well and at-risk
babies born in the
hospital of Sciacca
between 2003 and
2004. 1,068 babies
underwent screening.
942 well babies
(88.2%) and 126 atrisk babies (11.8%)
1. Initial screen with
TEOAE
2. Rescreen of
failures with TEOAE
at 7 months
3. Rescreen of
failures after a
further 2 weeks with
TEOAE
4. Re-test of failures
using TEOAE and
AABR
Sensorineural hearing
loss
> 56 dB = 2/1,068
1.87/1,000
(Neumann et
al 2006)
Hessen,
Germany
Multi-centre
(49) case series
17,439 well and atrisk babies born in
2005 and screened at
one of 46 maternity
clinics or 3 NICUs g
1. Initial screen with
AABR and/or
TEOAE
2. Re-test of failures
with diagnostic ABR
Congenital hearing loss
≥40 dB = 36/17,439
2.06/1,000
Bilateral = 30/17,439
1.72/1,000
Unilateral = 6/17,439
0.34/1,000
(Pastorino et
al 2005)
Milano,
Italy
Case series –
hospital based
19,777 well and atrisk babies born
between 1997 and
2001. 19,290 well
babies (97.5%) and
487 at-risk babies
(2.5%)
Well babies:
1. Initial screen with
TEOAE
2. Re-screen of
failures with TEOAE
3. Re-test of failures
with diagnostic ABR
At-risk babies:
1. Diagnostic ABR
Universal neonatal hearing screening
Congenital hearing loss
≥40 dB = 63/19,777
3.19/1,000
Bilateral = 33/19,777
1.67/1,000
Unilateral = 30/19,777
1.52/1,000
177
Study
Location
Study design
Study population
Intervention
Prevalence
(Stewart et al
2000)
USA
Multi-centre (5)
case series
11,711 neonates
enrolled from 5
clinical settings born
between December
1996 and December
1997. UNHS
performed on
neonates from WBN,
NICU and
intermediate care
nurseries (depending
on site)
1. Initial screen with
AABR
2. Rescreen of
failures prior to
discharge with AABR
(Algo 2)
Outpatient re-test of
missed failures also
with AABR (Algo 2)
3. Failures referred
for diagnostic ABR,
OAE or
tympanometry
Sensorineural hearing
loss
>35 dB = 32/11,711
2.73/1,000
169,487 well and atrisk babies born
before January 1
2004 who were
screened as part of
the national newborn
hearing screening
program in England
Well babies:
1. Initial screen with
TEOAE
2. Rescreen of
failures with AABR
3. Failures referred
for audiologic
assessment
Bilateral PCHI
≥40 dB = 169/169,487
1.00/1,000
(Uus &
Bamford
2006)
UK
Multi-centre
case series
At-risk babies:
1. Initial screen with
TEOAE and AABR
2. Those who failed
TEOAE or AABR
referred for
audiologic
assessment
(Watkin &
Baldwin
1999)
178
UK
Retrospective
cohort
25,199 neonates
born between
January 1992 and
1997
1. Initial screen with
TEOAE
2. Rescreen of
failures with TEOAE
or ABR
3. Re-test of failures
with ABR
4. Passed ABR <40
dB: referred for
behavioural
localisation tests at 8
months
5. Failed ABR >40
dB: re-tested with
ABR followed by
behavioural
observation
audiometry and
examined with
auroscopy and
tympanometry
Bilateral = 21/11,711
1.79/1,000
Unilateral = 11/11,711
0.94/1,000
Conductive
>35 dB = 8/11,711
0.68/1,000
Sensorineural, mixed or
conductive hearing loss
≥40 dB = 152/169,487
0.90/1,000
40-69 dB = 59/169,487
0.35/1,000
70-94dB = 40/169,487
0.24/1,000
≥95 dB = 53/169,487
0.31/1,000
Bilateral
>40 dB = 33/25,199
1.31/1,000
41–80 dB = 26/25,199
1.03/1,000
>80 dB = 7/25,199
0.28/1,000
Unilateral
>40 dB = 9/25,199
0.36/1,000
41–80 dB = 6/25,199
0.24/1,000
>80 dB = 3/25,199
0.12/1,000
Universal neonatal hearing screening
Study
Location
Study design
Study population
Intervention
Prevalence
(White et al
2005)
USA
Multi-centre (7)
case series
86,634 well and atrisk babies screened
in seven hospitals
born between May
2001 and January
2003. Only babies
from families whose
primary language
was English or
Spanish were
recruited
1. Initial screen with
OAE h
2.Rescreen of
failures with AABR
3.Re-test of OAE
failures with
tympanometry, OAE
or visual
reinforcement
audiometry
Sensorineural or
conductive hearing loss
> 40 dB = 133/86,634
1.54/1,000
All 29,317 children
born between
January 1983 and
December 1986 in
the Nottingham
health district to
mothers living in the
same district
(27,581 non-NICU)
Children’s Hearing
Assessment Centre
records examined for
all children (born
January 1983 to
December 1986)
fitted with hearing
aids
NICU graduate
status determined
through record
matching.
Determined family
history of congenital
hearing loss;
syndromes or
abnormalities at birth
Children with hearing
aids received either:
age appropriate
behavioural test, or
ABR or
tympanometry
Sensorineural/mixed
and congenital/
progressive hearing loss
>50 dB = 1.06/1,000
Children born
between 1985 and
1993 in the Trent
Health Region at the
time of data collection
Ascertained through
health and/or
education records.
Questionnaires were
also sent to some
parents
PCHI ≥40 dB in better
hearing ear (averaged
over 0.5, 1, 2, 4 Hz)
126,369 children born
between 1 January
1979 and 31
December 1990
Examination of
records/ databases
from the School for
the Deaf to identify
age of diagnosis
Questionnaire also
sent to parents to
validate data
Congenital hearing loss
>50 dB = 151/126,369 i
1.19/1,000
41-70 dB = 69/86,634
0.80/1,000
> 70 dB = 64/86,634
0.74/1,000
Ascertainment via other methods
(Davis &
Wood 1992)
(Fortnum &
Davis 1997)
(Hadjikakou
& Bamford
2000)
Nottingham
Health
District, UK
Trent region,
UK
Cyprus
Retrospective
cohort
Retrospective
cohort
Retrospective
cohort
51–64 dB = 0.32/1,000
65–79 dB = 0.26/1,000
80–94 dB = 0.19/1,000
>94 dB = 0.29/1,000
Congenital (1991–1993
n = 186,078)
≥40 dB = 147/186,078
0.79/1,000
51–69 dB = 0.30/1,000
70–94 dB = 0.42/1,000
>94 dB = 0.47/1,000
(Kubba et al
2004)
Glasgow, UK
Retrospective
cohort
Universal neonatal hearing screening
105,517 children born
in Greater Glasgow
between January
1985 and December
Ascertained through
hospital and
educational
audiology records
Bilateral PCHI
≥40 dB = 124/105,517
1.18/1,000
179
1994
(Mytton &
Mackenzie
2005)
Oldham, UK
Retrospective
cohort
54,448 children born
in Oldham between
January 1986 to
December 2002
Cases of PCHI
ascertained through
records from local
paediatric audiology
clinic and the Child
Health Database of
Oldham
Congenital hearing loss
Bilateral
≥40 dB = 98/54,448
1.80/1,000
(Neary et al
2003)
Warrington,
UK
Retrospective
cohort
Children born
between 1 January
1981 and 31
December 1992 and
living in Warrington
District in December
1998
Not stated
Unilateral PCHI
≥40 dB = 37 cases
1.15/1,000
(Nekahm et
al 2001b)
Tyrol, Austria
Retrospective
cohort
124,809 children born
between 1980 and
1994 in the Tyrol
Retrospective
examination of
medical records and
case histories from
the Childhood
Hearing Impairment
Register of the Tyrol
Congenital hearing loss
>40 dB = 158/124,809
1.27/1,000
Survey of all ENT j
practitioners in the
Tyrol to verify the
data
70–94 dB = 37/124,809
0.30/1,000
Retrospective
examination of the
Australian Hearing
database for all
children born in 1993
and fitted with
hearing aids for
hearing impairment
classified as
congenital (through
cross-matching with
VIHSP k database
and parent
questionnaires)
Congenital hearing loss
Bilateral
>40 dB = 1.12//1,000
>60 dB = 0.48/1,000
>90 dB = 0.17/1,000
(Russ et al
2003)
Victoria,
Australia
Retrospective
cohort
64,116 children born
in 1993 in the state of
Victoria and surviving
the neonatal period
41–69 dB = 84/124,809
0.67/1,000
>94 dB = 37/124,809
0.30/1,000
a Universal neonatal hearing screening; b well baby nursery; c transient evoked otoacoustic emissions testing; d automated auditory brainstem
response testing; e conventional auditory brainstem response testing; f permanent childhood hearing impairment; g neonatal intensive care unit;
h otoacoustic emissions testing; i 27 children with hearing loss were detected after 6 years of age. Age of detection for seven children was
unavailable. Exclusion of these children would underestimate the prevalence; j ear, nose and throat; k Victorian Infant Hearing Screening
Program.
180
Universal neonatal hearing screening
Included controlled and descriptive studies on screening safety
Study
Quality
Screen setting
Study population
Screening protocol
Study design
Method of data
ascertainment
Outcome(s)
assessed
(Clemens et al
2000)
N/A
Women’s Hospital of
Greensboro
49 parents of 76 well
babies with normal
hearing who screened
positive
2-stage
Stage 1: inpatient
AABR a (51% received repeat
AABR 12–24 hours after fail)
Stage 2: outpatient
AABR or dx ABR b
Diagnostic stage:
referral for further evaluation
Cross-sectional
survey
Questionnaire survey of
all false-positive results
after stage 2 screening
Parental anxiety
Cohort study
Questionnaire survey
sent to 99 mothers of
babies who received
HDVT, 35 mothers of
babies who screened
negative on first stage of
neonatal hearing
screening, and 30
mothers of babies who
screened positive on first
stage of neonatal hearing
screening
Spielberger State-Trait
Inventory and individual
item about ‘Worry about
baby’s hearing’
Parental anxiety
and worry
Cohort study
Questionnaire survey
sent 3 weeks after
completion of screening
to random sample of
mothers of infants with
different screening
results
Parental anxiety
and worry
North Carolina, USA
Study duration: 1 year
(Crockett et al
2005)
(Crockett et al
2006)
Level III-2
QS = 3.5/6
Level III-2
QS = 4/6
6 maternity
hospitals, health
visitor clinics and
general practice
surgeries
England
48 mothers of infants
who underwent health
visitor distraction test
(21 were referred, 27
passed)
42 mothers of
neonates who received
hearing screening (16
were referred, 26
passed)
2-stage health visitor
distraction test (HVDT)
Stage 1:
Hospitals
participating in
UNHS pilot program
England
344 mothers of infants
were screened
Group 1: clear
responses in both ears
from 1st or 2nd stage
Group 2: not clear
3-stage
Stage 1:
Universal neonatal hearing screening
HVDT
Stage 2:
HVDT
Diagnostic stage:
referral for further evaluation
3-stage neonatal hearing
screening
Stage 1:
OAE
Stage 2:
OAE
Stage 3:
AABR
Diagnostic stage:
referral for further evaluation
OAE
Stage 2:
OAE
Stage 3:
181
(Hergils & Hergils
2000)
N/A
University Hospital of
Linköping
responses in one or
both ears at 1st or 2nd
stage but clear on
AABR
Group 3: not clear
responses in one ear
at AABR and referred
for possible unilateral
hearing loss
Group 4: not clear
responses in either ear
at AABR and referred
for possible bilateral
hearing loss
AABR
Diagnostic stage:
referral for further evaluation
Parents of 83 well
babies who were
screened
2-stage
Stage 1:
Linköping, Sweden
Study duration: 1 year
(Kennedy 1999)
(Wessex Universal
Neonatal Screening
Trial Group)
Level III-2
QS = 2.5/5
4 maternity hospitals
Wessex, UK
Mothers of 150 lowrisk babies
75 screened positive
versus
75 screened negative
Study duration: 3 years
(Kolski et al 2007)
Level III-2
QS=1.5/6
Maternity hospital,
University Hospital of
Picardy
Amiens, France
182
143 mothers of well
babies
115 babies were
screened, 58 by 1st
strategy, 57 by 2nd
strategy
Study duration: 6
months for each
Spielberger State-Trait
Inventory and individual
item about ‘Worry about
baby’s hearing’
Cross-sectional
survey
Unstructured
questionnaire survey of
all screened in first year
of universal neonatal
screening program
Parental anxiety
2-stage
Stage 1: <48 hours (WBN) d;
<discharge (NICU) e
TEOAE
Stage 2: same day as fail
AABR
Diagnostic stage: 6–12 weeks
dx ABR
Nested casecontrol study
(III-2)
Questionnaire survey
sent to 100 mothers of
infants screening positive
and 100 mothers of
infants screening
negative
Spielberger State-Trait
Anxiety Inventory
Attitude towards the Baby
Scale
Negative attitude
to baby
Concern for the
baby
Strategy 1:
2-stage:
Stage 1: (day 3)
Cohort study
2 semi-structured
interviews comprising
MADRS scale for postpartum depression,
anxiety scale of EPDS
questionnaire, certain
items of the Kennerley
self-administered
questionnaire and the
Parental anxiety
Quality of early
interactions
TEOAE c
Stage 2:
TEOAE or AABR
Diagnostic stage:
dx ABR
OAE
Stage 2: (3-4 weeks later)
OAE
Diagnostic stage:
Universal neonatal hearing screening
strategy
not stated
interaction scale,
investigating visual,
physical, mental and
social dimensions of the
mother-infant relationship
1st interview held after
announcement of result
of 1st stage
2nd interview held just
before confirmation test
Strategy 2:
2-stage:
Stage 1: (2 months)
OAE
Stage 2: (3-4 weeks later)
OAE
Diagnostic stage:
referral for further evaluation
(Tatli et al 2007)
Level IV
QS = 4.5/5
Dokuz Eylul
University Hospital,
Izmir
Turkey
466 well and at-risk
babies
Study duration: 18
months
2-stage:
Stage 1: (last day of
discharge)
Cross-sectional
survey
466 mothers of screened
infants were interviewed
prior to testing and 28
mothers interviewed
again after a positive
result
Parental anxiety
Cross-sectional
survey
Questionnaire survey of
mothers of well babies at
(1) initial screen and prior
to receiving result (2)
rescreen, 2–8 weeks
after receiving initial fail
result
Parental anxiety
Prospective
cohort (III-2)
Speilberger State Trait
Anxiety Inventory
(modified to 10 items)
administered to mothers
of babies undergoing
rescreen appointment;
and posted to mothers of
babies (of the same age)
at another hospital who
do not have access to
Parental anxiety
TEOAE
Stage 2:
TEOAE
Diagnostic stage:
dx ABR
(Vohr et al 2001)
N/A
Women & Infants
Hospital
Rhode Island, USA
307 mothers of initial
screen well babies
40 mothers of rescreen
well babies
3-stage
Stage 1:
TEOAE
Stage 2:
TEOAE
Stage 3:
AABR
Diagnostic stage:
dx audiology
(Watkin et al 1998)
Level III-2
QS = 3.5/6
Whipps Cross
Hospital
London, UK
57 mothers of rescreen
babies
versus
61 mothers of
unscreened babies
2-stage
Stage 1: < discharge
TEOAE
Stage 2:
TEOAE
Diagnostic stage:
dx ABR
Universal neonatal hearing screening
183
screening
(Weichbold & Welzl
Mueller 2001)
N/A
University Hospital
Innsbruck, Austria
85 mothers of well
babies with normal
hearing who screened
positive
43 mothers of screen
positives referred for
diagnostic testing
Cross-sectional
survey
2-stage
Stage 1: <48 hours
TEOAE
Stage 2: < discharge
TEOAE
Diagnostic stage:
dx audiology
Questionnaire survey of
mothers of well babies at
(1) discharge for baby
who failed initial test but
passed re-test (2) at 1
month after baby failed
re-test and prior to
diagnostic test
Parental anxiety
Automated auditory brainstem response test; b diagnostic auditory brainstem response test; c transient evoked otoacoustic emissions test; d well baby nursery; e neonatal intensive care unit; MADRS= Montgomery Åsberg
Depression Rating Scale ; EPDS=Edinburgh post-natal depression score
a
184
Universal neonatal hearing screening
Included studies on diagnostic accuracy
All children were cross-classified on the test and reference standard.
Study
Location
Diagnostic
level of
evidence
Quality
score
Study
population
Testing
protocol
Diagnostic
test
Reference
standard
III-2
10/14
n = 119
babies (238
ears)
Tester:
audiologist
TEOAE –
ILO88
Otodynamic
analyser
Default
protocol, click
rate and
intensity level
used
Combined
AABR i and
conventional
ABR
TEOAE a versus ABR b
(Jacobson
& Jacobson
1994)
Norfolk,
Virginia, USA
56% at risk
M:F d =
43:24
44% well M:F
= 27:25
All in stable
physical
condition
Uncorrected
age: 33 to 41
weeks
(McNellis &
Klein 1997)
Charleston,
South
Carolina,
USA
III-2
10/14
n = 50 babies
All healthy,
low risk
babies
Risk factors
as defined by
Joint
Committee
on Infant
Hearing in
1994
Uncorrected
age: full-term
>37 weeks
Environment:
‘real world’ –
crib-side in
WBN e or
NICU f. Baby
asleep or
resting quietly
Time of tests:
immediately
prior to
discharge
Test order:
arbitrary but
ABR probably
done second
as all 14 ears
LTFU g didn’t
receive it.
Tester:
audiologist
Environment:
quiet, empty
room near
nursery. Baby
asleep or
resting quietly
Time of tests:
4–25 hours of
age for first
screen. Some
rescreened in
day 2. All
tested prior to
discharge
Test order: all
received ABR
first
Universal neonatal hearing screening
Screen fail:
unilateral
absence of
OAE h with
>50%
reproducibility
(consistent
with ≥30 dB
HL in midfrequency
range)
TEOAE -–
Otodynamics
ILO88 (vers.
4.20B)
Click stimulus
80–85 dB
SPL k
Screen fail:
unilateral
absence of
OAE ≥3 dB in
at least 3
frequency
bands (1.6,
2.4, 3.2, 4.0
kHz) or ≥4
dB at 3.2 and
4.0 kHz
Conventional
ABR –
Navigator,
Biologic
Systems Corp
Test fail:
absence of
identifiable
and replicable
wave V peak
at 35 dB
nHL j
AABR –
ALGO-1,
Natus Medical
Inc.
Test fail:
unilateral refer
at 35 dB HL
Conventional
ABR –
Intelligent
Hearing
Systems
SmartEP ABR
system in
screener
mode but did
not use
automatic
waveform
identification
option
Test fail:
absence of
repeatable
wave V
(interpeak
wave V
agreement of
0.2 msec) at
40 dB HL l
185
(Smyth et
al 1999)
Brisbane,
Queensland,
Australia
III-1
11/14
n = 37 babies
All normal,
well babies
without risk
factors for
hearing loss
and normal
prenatal,
perinatal,
postnatal and
maternal
history
Uncorrected
age: all fullterm (37 to
42 weeks)
Tester:
audiologist
Environment:
quiet room.
Baby asleep
or resting
quietly
Time of tests:
all tested
prior to
discharge at
gestational
age 37–42
weeks
Test order:
not stated
Data
available on
special care
neonates but
not relevant
for this
assessment
TEOAE –
ILO88
Otodynamic
analyser
(vers. 3.94)
(1) Default
protocol, click
rate and
intensity level
used, and
(2)
QuickScreen
program
option
Peak
stimulus 71–
83 dB SPL
Conventional
ABR –
Biologic
Traveller
System
Test fail:
absence of
any or all
ABR
waveforms IIII-V at 70 dB
nHL and/or if
the wave V
test-retest
threshold
poorer than
30 dB nHL
Screen fail:
unilateral
absence of
OAE ≥10 dB
plus
reproducibility
of ≥65%, for
the default
test format
TEOAE versus tympanometry
(Ho et al
2002)
Community
screening
clinics (29) in
Minnesota,
USA
III-2
10/14
n = 33 babies
Normal and
at-risk babies
in the
community,
excluding
those with
cerumen
occlusion of
the ear
Uncorrected
age: 0–6
months
Tester:
trained
graduate
audiology
students
Environment:
not stated
Time of tests:
<6 months.
Both tests
performed at
same
screening
TEOAE -–
Otodynamics
ILO88
Default
variables at
0.8, 1.6, 2.4,
3.2 and 4.0
kHz and
QuickScreen
program
option
Peak click
stimulus 80
dB SPL (±2
dB)
Data
available on
other ages
but not
relevant for
this
assessment
Test order:
not stated
Screen fail:
unilateral
absence of
OAE if
reproducibility
index ≤50%
at 0.8 and 1.6
kHz or ≤75%
at 2.4, 3.2
and 4.0 kHz
n = 50 babies
Tester: two
investigators
experienced
with the tests
AABR –
ALGO-1, Fa.
Nicolet
Screen fail:
unilateral
refer at 35 dB
HL
Tympanometry
– Welch-Allyn
MicroTymp II.
Handheld &
226 Hz probe
tone
Test fail: SA
<0.2 mmho /
TW >300
daPa
AABR versus ABR
(Schausei
l-Zipf &
Von
Wedel
1988)
186
Women and
children’s
clinic, Cologne,
Germany
III-2
9/14
25 healthy
newborns
and 25 at risk
for hearing
loss
Environment:
baby asleep
Conventional
ABR –
Pathfinder II
SignalAnalysis –
system of Fa.
Nicolet
Universal neonatal hearing screening
after feeding
Uncorrected
age: 38–41
weeks for the
well babies
Time of tests:
<discharge.
Both tests
performed at
same
screening
Test fail: not
stated
Test order:
AABR first
Transient evoked otoacoustic emissions test; b auditory brainstem response test; c assumption that selection was non-consecutive as
‘consecutive’ not mentioned in text; d male:female; e well baby nursery; f neonatal intensive care unit; g lost to follow-up; h otoacoustic
emissions; i automated auditory brainstem response test; j near hearing level; k sound pressure level; l hearing level.
a
Universal neonatal hearing screening
187
Included controlled studies on effectiveness of screening
Level of
evidence
Quality
Study
Setting
Study population
Screening protocol
Screen fail
criterion
Comparator
Outcome(s)
assessed
III-1
QS=3/6
(Kennedy et al
1998) (Wessex
Universal
Neonatal
Screening Trial
Group)
Princess Anne
Hospital,
Southampton
St Mary’s
Hospital,
Portsmouth
Royal United
Hospital, Bath
Princess
Margaret
Hospital, Swindon
53,781 well and atrisk babies born at
4 hospitals during
study period
2-stage
Stage 1: <48 hours
(WBN a); <discharge
(NICU b)
TEOAE c
Stage 2: same day as
fail
AABR d
Diagnostic stage: 6–12
weeks
dx ABR e
Bilateral fail if
TEOAE not
produced for
either
(1) ≥28 dB
(2) ≥98% whole
response
correlation
(3) 3 of 5
frequencies, with
3 dB SNR f at 0.8,
1.2 kHz and 6 dB
at 2.4, 3.6, 4.0
kHz
(protocol began
with a unilateral
fail but this
changed to a
bilateral fail
midway through
the study period)
Not universal
neonatal hearing
screening
Coverage
Failure rate/referral
False alarm rate
Absolute and
Incremental yield
Age at referral
Age at diagnosis
Age at management
False negative
Not universal
neonatal hearing
screening
Communication ability
Wessex, UK
25,609 babies born
during periods with
hearing screening
28,172 babies born
during periods
without hearing
screening
Study duration: 3
years
(Kennedy et al
2005)
Tester: trained testers
Environment: bedside
Follow-up: 8 years
AABR fail at ≥35
dB HL
III-2
188
QS=5/6
(Kennedy et al
2006)
Eight districts of
Southern
England:
Wessex
subgroup:
Princess Anne
Hospital,
Southampton
St Mary’s
Hospital,
Portsmouth
68,714 infants born
during periods
hearing screening
88,019 infants born
during periods
without universal
screening
168 PCHI children
identified
Wessex Subgroup: as
above
Greater London
Subgroup:
Hillingdon Hospital: 2stage auditory response
cradle (ARC)
Whipps Cross Hospital:
2-stage TEOAE
Diagnostic stage:
Wessex
Subgroup: as
above
Greater London
Subgroup:
Fail criteria not
stated
Language acquisition
ABR fail ≥40 dB
HL
Universal neonatal hearing screening
Royal United
Hospital, Bath
Princess
Margaret
Hospital, Swindon
Greater London
subgroup:
Whipps Cross
University
Hospital, London
One district
adjacent to
Whipps Cross
Hillingdon
Hospital
Postgraduate
Centre
East London
District of
Redbridge
III-2
QS=5.5/6
(Nekahm et al
2001a)
Tyrol, Austria
dx ABR
120 gave consent
to participate in
study
61 born during
periods with hearing
screening
Two other
hospitals formed
the ‘Not universal
neonatal hearing
screening’
59 born during
periods without
hearing screening
91 Tyrolean
children born
between 1990 and
1999 and registered
with PCHI g at the
only audiological
centre in the state
UNHS h = 1995–
1999
2-stage
Stage 1:
Not stated
TEOAE
Not universal
neonatal hearing
screening
Age at diagnosis
Stage 2:
TEOAE
Diagnostic stage:
dx audiology
Tester: not stated
Not UNHS = 1990–
1994
Environment: not stated
Study duration: 10
years, retrospective
Universal neonatal hearing screening
189
III-2
III-2
III-2
QS=5/6
QS=4.5/6
5.5/6
(Neumann et
al 2006)
(Weichbold et
al 2006)
(YoshinagaItano et al
2001)
Hessen
&Thuringia
Germany
Innsbruck, Austria
Colorado, USA
17, 349 well and atrisk babies in 2005
made up the UNHS
group
739 had 1 stage AABR
UNHS=
1995-2005
Not UNHS=
1990-2005
12,950 had 2-stage
Stage 1:
TEOAE
Stage 2:
AABR
321 hearing
impaired children
167 had been
screened
154 had not been
screened
Post-1995
UNHS criterion:
2-stage:
TEOAE-TEOAE
Diagnostic stage:
dx audiology
Study duration:
1990-2004
Retrospective study
Tester & Environment:
not stated
25 matched pairs of
children with
bilateral PCHI born
in hospital with and
without UNHS
Study duration: 5
years, retrospective
TEOAE ≥ 30dB
AABR ≥ 35dB
Not universal
hearing screening
Age at diagnosis
Not stated
Not universal
neonatal hearing
screening
Age at confirmation
Age at management
Bilateral fail ≥ 35
dB
Not universal
neonatal hearing
screening
Diagnosis <6 months
Bilateral fail ≥ 35
dB
Not universal
neonatal hearing
screening
Language acquisition
3,750 had 1 stage
TEOAE
1-stage
Stage 1:
AABR
Diagnostic stage:
dx audiology
Tester: not stated
Environment: not stated
III-2
5.5/6
(Yoshinaga
Itano et al
2000)
Colorado, USA
25 matched pairs of
children with
bilateral PCHI born
in hospital with and
without UNHS
1-stage
Stage 1:
AABR
Diagnostic stage:
dx audiology
Communication ability
Study duration: 5
190
Universal neonatal hearing screening
years, retrospective
Tester: not stated
Environment: not stated
a
g
Well baby nursery; b neonatal intensive care unit; c transient evoked otoacoustic emissions test; d automated auditory brainstem response test; e diagnostic auditory brainstem response test; f signal-to-noise-ratio;
permanent childhood hearing impairment; h universal neonatal hearing screening.
Universal neonatal hearing screening
191
Included descriptive studies on effectiveness of screening
Study
Quality & level
of evidence
Setting
Study population
Screening protocol
Screen fail criterion
Outcome(s) assessed
(Aidan et al 1999)
Level IV
QS=5/5
Hospital
1,727 well and at-risk
babies
2-stage
Bilateral or unilateral fail on at
least one of these criteria:
TEOAE response
1) >8 dB SPL c
2) broad spectrum (0.8–5 kHz)
3) correlation >60% between 2
traces of alternate buffers
Coverage
Failure rates
LTFU e
False alarm rates
Yield
Paris, France
Stage 1: day 2
Study duration: 18 months
TEOAE a
Stage 2: <30 days
TEOAE
Diagnostic stage:
dx audiology b
ABR fail: >40 dB HL d
conventional frequencies
Tester: clinical audiologist
Environment: quiet room
while babies slept
(Bailey et al 2002)
Level IV
QS=5/5
5 maternity
hospitals
13,214 well and at-risk
babies
Perth, Australia
Study duration: 18 months
2-stage (well-baby protocol)
Unilateral fail: ≥35 dB HL
(bilaterally in 40 babies)
Coverage
Failure rates
LTFU
False alarm rates
Yield
Fail not stated
Coverage
Failure rates
Yield
Stage 1: day 1
TEOAE
Stage 2: < discharge
AABR f
+ 1–2 weeks after discharge
TEOAE ± AABR
Diagnostic stage:
dx ABR g
Tester: trained hearing
screeners; variety of
backgrounds
Environment: not stated
(Bamford et al 2005)
Level IV
QS=4/5
23 areas of
England
About 120,000 births per
annum
2-stage
Stage 1:
TEOAE
192
Universal neonatal hearing screening
Stage 2:
AABR
Diagnostic stage:
dx ABR and middle ear
function test
Tester: trained screener
Environment: mother’s
bedside
(Bantock & Croxson
1998)
Level IV
QS=4/5
Hospital
1,492 well babies born in
hospital 1992–1995
London, UK
Study duration: 4 years
2-stage
Stage 1: first few days
TEOAE
Stage 2: 1 month later
TEOAE
Diagnostic stage:
dx ABR
Tester: not stated
Bilateral or unilateral fail on at
least one of these criteria:
TEOAE response
1) >5 dB SPL above noise
level
2) broad spectrum (0.8–4 kHz)
3) reproducibility >50%
Failure rate
False alarm rate
Yield
ABR fail: waves I, II and V not
present at >40 dB HL
Environment: hospital
incubator
Communitybased health
centre (~7% of
population)
London, UK
319 babies born June 1995
– May 1996
Study duration: 1 year
2-stage
Stage 1: 3–4 weeks
TEOAE
Stage 2: 1–2 weeks later
TEOAE
Diagnostic stage: 1–2 weeks
later
dx ABR
Tester: research nurse
Bilateral or unilateral fail on at
least one of these criteria:
TEOAE response
1) >5 dB SPL above noise
level
2) broad spectrum (0.8–4 kHz)
3) reproducibility >50%
Coverage
Failure rates
Yield
ABR fail: waves I, II and V not
present at >40 dB HL
Environment: quiet or soundproofed room – in car seat or
sling if asleep
Universal neonatal hearing screening
193
(Brennan 2004)
Level IV
QS=2.5/5
Instititution in
Illinois
Not stated
Level IV
QS=3.5/5
Hospital Israelita
Albert Einstein
Sao Paulo, Brazil
Fail not stated
Coverage
Failure rate
Fail if an absent TEOAE
response for 2 or more of 4
frequency bands – evaluated
by:
(1) ≥3 dB SPL above noise
level for 1.6 kHz or
reproducibility ≥50%
2) ≥6 dB SPL above noise
level for 2.4, 3.2 and 4.0 kHz
or reproducibility ≥70%
Coverage
Failure rates
LTFU
False alarm rates
Yield
OAE
Stage 2: (post-discharge)
AABR
Diagnostic stage:
Not stated
Tester: not stated
Environment: quite area of
the nursery
USA
(Chapchap & Segre
2001)
2-stage
Stage 1: (pre-discharge)
4,196 well babies and
those in the neonatal
intensive care unit (NICU)
born September 1996 to
August 1999
Study duration: 3 years
2-stage
Stage 1: day 2–3 – (well
babies), < discharge (NICU
babies)
TEOAE
Stage 2: <30 days
TEOAE
Diagnostic stage: 3 months
dx audiology (ABR, TEOAE,
behavioural, tympanometry)
Tester: not stated
Environment: not stated
(Chiong et al 2007)
Level IV
QS=2/5
Several
communities in a
rural area,
Bulacan province,
Philippines
724 neonates
Study duration: 2 years 10
months
1-stage: 565 infants
Stage 1: unclear
TEOAE+ABR
Fail not stated
Coverage
Yield
(Clarkson et al 1994)
Level IV
QS=2.5/5
Women and
Infants Hospital of
Rhode Island
1,850 well and at-risk
babies born 1990–1991
464 infants: 2-stage
Stage 1: day 1–4
TEOAE + ABR
Stage 2:
TEOAE + ABR
Fail: ≥60 dB HL referred for dx
ABR; <60 dB HL referred for
behavioural audiologic
evaluation
Fail: ≥25 dB HL for conductive
hearing loss
LTFU
False alarm rate
Yield
Some duplication with
(Maxon et al 1993)
Providence,
Rhode Island,
USA
194
Study duration: 6 months
1,386 infants: 3-stage
Stage 1: day 1–4
Universal neonatal hearing screening
TEOAE
Stage 2: day 1–4
ABR
Stage 3: 4–6 weeks later
TEOAE + ABR
Diagnostic stage: 12–16
weeks of age
dx ABR and behavioural
audiologic evaluation
Tester: paraprofessional
technician
Environment: quiet room
(Clemens et al 2000)
Level IV
QS=3/5
Women’s
Hospital of
Greensboro
5,034 well babies born in
hospital, July 1998 – June
1999
North Carolina,
USA
Study duration: 1 year
2-stage
Stage 1: inpatient
AABR (51% received repeat
AABR 12–24 hours after fail)
Stage 2: outpatient
AABR or dx ABR
Diagnostic stage:
referral for further evaluation
Fail: >35 dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Fail: >35 dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Tester: trained screening
technician for stage 1 and
audiologist for stage 2
Environment: not stated
(Clemens & Davis
2001)
Level IV
QS=5/5
Women’s
Hospital of
Greensboro
3,144 well babies born in
hospital, November 1999 –
May 2000
North Carolina,
USA
Study duration: 6 months
Universal neonatal hearing screening
3-stage
Stage 1: inpatient
AABR
Stage 2: inpatient (12–24
hours after fail)
AABR
Stage 3: outpatient
AABR or dx ABR
Diagnostic stage:
195
referral for further evaluation
Tester: trained screening
technician for stage 1 and
audiologist for stage 2
Environment: not stated
(Connolly et al 2005)
Level IV
QS=4.5/5
University of
Mississippi
17,602 well and at-risk
babies
Mississippi, USA
Study duration: 5 years
2-stage
Stage 1: (prior to discharge)
AABR
Stage 2: (outpatient)
AABR
Diagnostic stage:
Dx ABR and medical
evaluation
Fail >35 dBn HL
Coverage
Yield
Fail: not stated
(Data from 2000)
LTFU
False alarm rate
Yield
Tester: trained nurse
Environment: not stated
(Cox & Toro 2001)
Level IV
QS=5/5
Boston Medical
Centre
1,713 well and at-risk
babies
Massachusetts,
USA
Study duration: 1 year
2-stage (well babies)
Stage 1: day 1
DPOAE h
Stage 2: immediately after fail
AABR
Diagnostic stage: 2–4 weeks
dx ABR
All NICU babies received a 1stage AABR screen followed
by diagnostic referral
Tester: audiologist and
audiometric technicians
Environment: quiet room
196
Universal neonatal hearing screening
(Daemers et al 1996)
Level IV
QS=3/5
St Augustinus
Medical Institute
907 well babies born in
1993 and 1994
Antwerp, Belgium
Study duration: 2 years
2-stage
Stage 1: > day 3
TEOAE
Stage 2: ~3 weeks later
TEOAE
Diagnostic stage: 3 months
dx ABR
Bilateral or unilateral fail if
absent TEOAE response –
evaluated by:
(1) >3 dB SPL above noise
level for 1.6 kHz
2) >6 dB SPL above noise
level for 2.4, 3.2 and 4.0 kHz
(3) and reproducibility >50%
Failure rates
LTFU
False alarm rates
Yield
Fail if TEOAE reproducibility
<50% with n<3 frequencies
with intensity >3 dB SPL
Failure rates
Yield
Fail: ≥35 dB HL bilaterally/
unilaterally
Failure rate
LTFU
False alarm rate
Yield
Tester: clinical audiologist
Environment: soundproof or
quiet room
(De Capua et al 2003)
Level IV
QS=4.5/5
Hospital
532 well and at-risk infants
2 stage
Italy
Study duration: not stated
Stage 1: 4 days
TEOAE
Stage 2:15-30 days post 1st
test
TEOAE
Diagnostic stage: within 1
month of 2nd test
Click ABR
Tester: not stated
Environment: not stated
(Downs 1995)
Level IV
QS=2/5
17 hospitals in
Colorado
14,494 well and at-risk
babies born 1992 onwards
Colorado, USA
Study duration: not stated
1-stage
(unclear whether initial
screen was repeated)
Stage 1:
AABR
Diagnostic stage: 1 month
dx ABR
Tester: certified audiologist
Universal neonatal hearing screening
197
Environment: not stated
(Govaerts et al 2001)
Level IV
QS=4/5
St Augustinus
Hospital
2,012 well and at-risk
babies born in 1999
Antwerp, Belgium
Study duration: 1 year
2-stage
Stage 1: day 3–5
TEOAE
Stage 2: 3 weeks later in
hospital or community
TEOAE
Diagnostic stage: 3 months
dx ABR
Bilateral failure to produce
TEOAE – evaluated by:
1) Signal-to-noise ratio of 6 dB
for at least 3 of 4 frequencies
– 1.6, 2.4, 3.2 and 4.0 kHz
(2) and reproducibility >50%
Coverage
Failure rates
LTFU
False alarm rates
Yield
ABR bilateral fail >40 dB
Tester: clinical audiologist
Environment: not stated
(Habib & Abdelgaffar
2005)
Level IV
QS=4/5
Dr. Soliman
Fakeeh Hospital
11, 986 non-high-risk
neonates
2-stage:
Stage 1: (<48 hrs of life)
OAE
Jeddah, Saudi
Arabia
Stage 2: (*5th day of life)
Failure to produce more than
50% reproducibility and
response amplitude at least
1dB SPL per octave
Coverage
Failure rates
Yield
Unilateral fail if TEOAE did not
have ≥60% reproducibility and
≥80% stimulus stability
Failure rate
LTFU
False alarm rate
Yield
OAE
Diagnostic stage: (5 months)
ABR
Tester: audiology technician
Environment: not stated
(Hahn et al 1999)
198
Level IV
QS=2/5
Hospital
Study 1: 388 well babies
Study 1
1-stage: day 2–6
Münster,
Germany
Study duration: not stated
TEOAE
Diagnostic stage: 4 months
full pedaudiological
assessment
Universal neonatal hearing screening
Study 2: 55 well babies
Study duration: not stated
Study 2
1-stage: day 2–6
TEOAE + AABR
Diagnostic stage: 4 months
full pedaudiological
assessment
Unilateral fail if TEOAE did not
have ≥60% reproducibility and
≥80% stimulus stability
Unilateral AABR fail: >35 dB
HL
Failure rate
LTFU
False alarm rate
Yield
Fail to produce ≥ 70%
reproducibility, and a ≥6 dB
SNR (signal-to-noise ratio) for
at least 2 out of 5 frequency
bands
Coverage
Failure rates
LTFU
False alarm rates
Yield
Unilateral or bilateral fail on
AABR: ≥35 dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Tester: trained specialists
Environment: newborn ward,
after baby was fed
(Hatzopoulos et al
2007)
QS=3.5/5
Main maternity
hospital of Tirana
Tirana, Albania
1,561 well and at-risk
babies
(463 well & 1,098 NICU
babies)
Study duration: 1 year
2–stage:
Stage 1: (2-3 days)
TEOAE
Stage 2: (4 wks post birth)
TEOAE
Diagnostic stage
Click ABR
Tester: not stated
Environment: bedside testing
when possible
(Hunter et al 1994b)
Level IV
QS=5/5
Princess Anne
Hospital
217 well and at-risk babies
2-stage
Stage 1: day 1–2
Study duration: 4 weeks
Southampton, UK
TEOAE
Stage 2: < discharge
AABR
Diagnostic stage: 1 month
dx ABR
Tester: audiologist
Environment: quiet room
during sleep, or after feeding
Universal neonatal hearing screening
199
(Huynh et al 1996)
Level IV
QS=5/5
Air Force
community
hospital
Maryland, USA
639 well babies born
1994–1995
2-stage
Stage 1: 6–48 hours
Study duration: 6 months
Stage 2: 1–3 weeks
TEOAE
TEOAE
Diagnostic stage:
dx ABR
Bilateral fail at first screen and
unilateral fail at rescreen –
evaluated by:
(1) TEOAE without
reproducibility ≥ 80% at 2.4,
3.2 and 4.0 kHz
(2) overall reproducibility
<40%
Coverage
Failure rates
LTFU
False alarm rates
Yield
Unilateral AABR fail ≥35 dB
HL
Coverage
Failure rate
LTFU
False alarm rate
Yield
Tester: physician, audiologist,
or trained technician
Environment: newborn
nursery in open bassinet
(Iley & Addis 2000)
Level IV
QS=3.5/5
York district
hospital
North Yorkshire,
UK
48 babies born 2000
Study duration: 4 days
1-stage
Stage 1: day 1–7 days
(inpatients); 1–12 weeks
(outpatients)
AABR
Diagnostic fail ≥ –40 dB
Diagnostic stage:
dx ABR
Tester: audiologist
Environment: (inpatient) in
crib at mother’s bedside while
asleep; (outpatient) quiet
room in Dept of Audiology
(Isaacson 2000)
Level IV
QS=3.5/5
Temple University
Hospital, North
Philadelphia
Pennsylvania,
USA
200
2,137 well and at-risk
babies born 1998–1999 in
an economically
depressed, inner-city area.
Study duration: 1 year
2-stage
Stage 1: >16 hours
TEOAE
Stage 2: immediately after fail
TEOAE
Diagnostic stage: 4–6 weeks
after discharge
dx ABR
Bilateral or unilateral fail if
TEOAE absent at signal-tonoise ratio ≥3 dB in 3 of 4
frequency bands (1, 2, 3, 4
kHz)
Coverage
Failure rate
LTFU
Yield
Universal neonatal hearing screening
Tester: trained
paraprofessional screeners
Environment: quiet room
adjacent to nursery
(Iwasaki et al 2003)
Level IV
QS=4.5/5
SeireiHamamatsu
General Hospital
& SeireiMikatahara
General Hospital
Hamamatsu,
Japan
4,092 infants born between
January 2000 and
December 2001
2-stage
Stage 1: (2-3rd day of life)
AABR
Stage 2: (5-6th day of life)
AABR
Diagnostic stage
dx ABR
Tester: trained technicians
Environment: quiet room
adjacent to the nursery
Fail if likelihood ratio is less
than 160 after 15000 sweeps
Coverage
Failure rates
LTFU
False alarm rates
Yield
(Jakubikova et al 2003)
Level IV
QS=3.5/5
Two gynaecology
and neonatology
departments in
Bratislava
3,048 high-risk or well
babies
2-stage:
Stage 1: (4-12th day)
Fail not stated
Coverage
Failure rates
False alarm rates
Yield
Unilateral or bilateral failure to
produce emission spectrum of
significant gain across testing
frequency range, approx. >30
dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Study duration: not stated
Department of
Pathological
Newborn and
Intensive Care
Unit of Children’s
University,
Bratislava
TEOAE
Stage 2: (1 month later)
TEOAE
Diagnostic stage:
Tympanometry & ABR
Tester: two well trained
specialists
Environment: not stated
Slovak Republic
(Kanne et al 1999)
Level IV
QS=4.5/5
Madigan Army
Medical Centre,
Tacoma
Washington, USA
Universal neonatal hearing screening
2,537 well and at-risk
babies born 1995–1996
Study duration: 14 months
2-stage
Stage 1: 24–72 hours <
discharge (NICU) i; 2 weeks
(WBN) j
TEOAE
Stage 2:
TEOAE
Diagnostic stage:
dx ABR
201
Tester: audiologist
Environment: secluded room
(Khairi et al 2005)
Level IV
QS=4/5
Hospital Universiti
Sains
Malaysia
401 newborns
Study duration: Feb 2000March 2000 & Feb 2001
and May 2001
2-stage:
Stage 1: (24-48hrs post birth)
TEOAE
Stage 2: (pre-discharge)
TEOAE
Diagnostic stage
dx audiology
Fail not stated
Coverage
Failure rates
LTFU
False alarm rates
Yield
Not stated
Coverage
Failure rates
LTFU
False alarm rates
Yield
Bilateral fail
Coverage
Failure rates
LTFU
False alarm rates
Yield
Tester: audiology technicican
Environment: low noise ward
(Khandekar et al 2006)
Level IV
QS=4.5/5
Hospitals in
Oman, Turkey
32,080 live births born in
2003
Study duration: 1 year
2-stage:
Stage 1: (24-48hrs post birth)
TEOAE or DPOAE
Stage 2: (pre-discharge)
TEOAE or DPOAE
Diagnostic stage
dx audiology
Tester: trained health staff
Environment: not stated
(Kolski et al 2007)
Level IV
QS=3.5/5
Maternity hospital
Strategy 1:
3202 newborns
2-stage:
Strategy 1:
Stage 1: (day 3)
Strategy 2:
2588 2 month old babies
OAE
Stage 2: (3-4 weeks later)
OAE
Strategy 2:
Stage 1: (2 months)
OAE
France
Study duration: 6 months
for each strategy
202
Universal neonatal hearing screening
Stage 2: (3-4 weeks later)
OAE
Diagnostic stage: not stated
Tester: not stated
Environment: not stated
(Leveque et al 2007)
Level IV
QS=4.5/5
17 maternities,
private and
public, and one
neonatal
intensive care
unit (NICU) in
ChampagneArdenne
33, 873 well and at-risk
babies born from January
2004 to March 2006
2-stage
Stage 1:
TEOAE
Stage 2:
Bilateral fail ≥35 dB HL -40 db
HL depending on screening
device used, which can not be
modified
Coverage
Failure rate
False alarm rate
Yield
Fail to produce four pairs of
alternating positive & negative
peaks
Coverage
Failure rate
False alarm rate
Yield
TEOAE/AABR
Diagnostic stage:
dx audiology
France
Tester:
Stage 1: Paramedical staff
Stage 2 conducted by
physician
Environment: not stated
(Lin et al 2004)
Level IV
QS=4/5
2 hospitals & 4
obstetric clinics
Tainan, Taiwan
5938 neonates
Study duration: 2 years 9
months
2-stage
Stage 1: pre-discharge
TEOAE
Stage 2: 1 month later
TEOAE
Diagnostic stage:
dx audiology
Tester: trained staff member
or audiologist
Environment: not stated
(Lin et al 2005)
Level IV
QS=4/5
Mackay Memorial
Hospital
21,273 well babies born
1998-2004
Taipei, Taiwan
Study duration: 6 years
Strategy 1
3-stage
Stage 1: 48 hours
TEOAE
Stage 2: <72 hrs
Universal neonatal hearing screening
Diagnostic ABR:
Fail to produce wave V
latency within developmental
norms in response to 35dB
nHL clicks
Bilateral or unilateral failure to
produce TEOAE of (1) ≥5 dB
in 3 of 5 frequency bands or
(2) ≥3 dB in 4 of 5 frequency
bands
Failure rates
LTFU
False alarm rates
Yield
203
TEOAE
(up to 7 rescreens with
TEOAE)
Stage 3: 1 month later
TEOAE/Tympanogram
Diagnostic stage: >1 month
dx audiology, including dx
ABR
Strategy 2
3-stage
Stage 1: 48 hours
TEOAE
Stage 2: <72 hrs
AABR
(up to 7 rescreens with
TEOAE)
Stage 3: 1 month later
TEOAE/Tympanogram
Diagnostic stage: >1 month
dx audiology, including dx
ABR
Tester: trained screeners
(volunteers and student
nurses)
Environment: quiet room in
the nursery
(Low et al 2005)
Level IV
QS=4/5
National
University
Hospital (NUH),
Singapore
General Hospital
(SGH), KK
Women’s and
Children’s
Hospital
(KKWCH)
Singapore
204
36,093 well and at-risk
babies born April 2002 to
March 2004
2-stage
Stage 1: Before discharge
SGH
OAE
KKWCH
AABR
NUH
OAE
Stage 2: 4-6 weeks after fail
SGH
OAE
KKWCH
AABR
NUH
AABR
Dx ABR – failure to produce a
repeatable wave V at 35 dB
nHL unilaterally or bilaterally
Not stated
Coverage
Failure rate
Yield
Diagnostic stage <3 months
Universal neonatal hearing screening
Audiological tests and
medical evaluation
(Martines et al 2007)
Level IV
QS=3.5/5
Sciacca Hospital
Italy
(Mason & Herrmann
1998)
Level IV
QS=4/5
Kaiser
Permanente
Medical Center
Honolulu, Hawaii
1068 well and at-risk
babies born during 20032004
3-stage (well babies)
Stage 1: 3 weeks post birth
TEOAE
Stage 2: 2 weeks post initial
test
TEOAE
Stage 3: unclear
TEOAE
Diagnostic stage:
TEOAE & AABR
The first 2-stages for at-risk
babies
Tester: not stated
Environment: sleeping, wellfed neonate
Fail criterion not stated
Coverage
Failure rate
False alarm rate
Yield
10,773 well and at-risk
babies born 1992–1997
2-stage
Stage 1: 3–36 hours (WBN);
< discharge (NICU)
AABR
Stage 2: immediately after fail
AABR
Diagnostic stage: <1 month
dx ABR and audiological
tests
Bilateral AABR fail ≥35 dB HL
Coverage
Failure rate
Yield
Study duration: 5 years
Diagnostic fail >35 dB nHL
Tester: audiologists and
technicians
Environment: quiet room
adjacent to nursery or
mother’s bedside
(Maxon et al 1993)
Level IV
QS=3.5/5
Women and
Infants Hospital of
Rhode Island
Universal neonatal hearing screening
1,850 well and at-risk
babies born 1990–1991
Study duration: 6 months
464 infants: 2-stage
Stage 1: day 1–4
TEOAE + ABR
Fail: ≥60 dB HL referred for dx
ABR; <60 dB HL referred for
behavioural audiologic
evaluation
Failure rates
LTFU
False alarm rate
205
Providence,
Rhode Island,
USA
Stage 2:
TEOAE + ABR
Fail: ≥25 dB HL for conductive
hearing loss
Yield
Unilateral or bilateral fail if
TEOAE was not at least 3 dB
above the noise floor and at
least halfway across the test
frequency bands of 2–3 kHz
and 3–4 kHz
Coverage
Failure rates
LTFU
False alarm rates
Yield
1,386 infants: 3-stage
Stage 1: day 1–4
TEOAE
Stage 2: day 1–4
ABR
Stage 3: 4–6 weeks later
TEOAE + ABR
Diagnostic stage: 12–16
weeks of age
dx ABR and behavioural
audiologic evaluation
Tester: paraprofessional
technician
Environment: quiet room
adjacent to nursery; infant in
enclosed isolette
(McPherson et al 1998)
Level IV
QS=2/5
Eight community
health clinics,
Northern
Brisbane
Queensland,
Australia
1,305 children between 1.5
and 2.5 months of age
presenting for
immunisation at the
community health clinic
[Data available on other
age groups up to 5 years
but most not relevant or
incomplete]
Study duration: 30 months
2-stage
Stage 1: ~ 2 months
TEOAE
Stage 2: 2 weeks later
TEOAE
Diagnostic stage: 2 weeks –
1 month later
dx ABR and behavioural
audiologic evaluation
Tester: audiologist or
community health nurse
Environment: quietest room
available at each clinic
206
Universal neonatal hearing screening
(Messner et al 2001)
Level IV
QS=5/5
Lucile Packard
Children’s
Hospital, Stanford
California, USA
6,340 well babies born
1998–1999
3-stage
Stage 1: <24 hours
Study duration: 16 months
Stage 2:
Unilateral or bilateral failure
>35 dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Fail to produce at least 3 of
the 4 frequency bands centred
at 1600, 2400, 3200 and 4000
Hz.
Coverage
Failure rates
False alarm rates
LTFU
Yield
Fail not stated
Coverage
Failure rates
False alarm rates
LTFU
Yield
AABR
AABR
Stage 3:
TEOAE
Diagnostic stage:
dx ABR
Tester: trained volunteers
Environment: nursery or
mother’s bedside
(Molini et al 2004)
(Mukari et al 2006)
Level IV
QS=4/5
Level IV
QS=4/5
Hospital
2,425 full-term newborns
Italy
Study duration:
Study 1: 17 months
Study 2: 9 months
Hospital
University
Kebangsaan
Malaysia (HUKM)
4437 newborns
(315 NICU, 4122 nonNICU)
2-stage
Study 1:
Stage 1: within 4 days
TEOAE
Stage 2: within 3 months
ABR
Diagnostic stage:
Dx audiology
2-stage
Stage 1: within 24 hrs
DPOAE
Stage 2: 2 months
Study duration:11 months
DPOAE
Diagnostic stage:
dx ABR
Tester: trained nurse staff
Environment: bedside or
nursery room for well babies
Isolation room for NICU
Universal neonatal hearing screening
207
babies
(Neumann et al 2006)
Level IV
QS=4/5
46 maternity
clinics and 3
NICU’s
17,349 well and at-risk
babies in 2005
Level IV
QS=4/5
Tsan Yuk hospital
TEOAE ≥ 30dB
AABR ≥ 35dB
Coverage
Failure rates
False alarm rates
LTFU
Yield
Unilateral or bilateral failure
>40 dB HL
Coverage
Failure rates
False alarm rates
Yield
Unilateral or bilateral failure
>35 dB HL
Coverage
Failure rates
LTFU
False alarm rates
Yield
3,750 had 1 stage TEOAE
Hessen,
Germany
(Ng et al 2004)
739 had 1 stage AABR
12, 950 had 2-stage Stage 1:
TEOAE
Stage 2:
AABR
1064 infants born between
May 1999 to October 1999
Hong Kong
1-stage (3 re-screens):
Stage 1a: (day 1-4)
DPOAE
Stage 1b: (day 5-14)
DPOAE
Stage 1c: (day 21-30)
DPOAE
Diagnostic stage:
dx ABR
Tester: Trained nurse
Environment: maternity ward
or hospital room
(Oudesluys-Murphy &
Harlaar 1997)
Level IV
QS=5/5
Community well
baby clinic,
Barendrecht
Netherlands
288 well and at-risk babies
born in 1995
2-stage
Stage 1: 1–2 weeks
Study duration: 1 year
Stage 2: 2 weeks later
AABR
AABR
Diagnostic stage: 6 weeks
later
dx audiology
Tester: trained child care
district nurses
208
Universal neonatal hearing screening
Environment: in the home,
with infant sleeping
(Owen et al 2001)
Level IV
QS=3.5/5
Local health
centres and
homes in urban
and rural settings
West
Gloucestershire,
UK
683 well babies registered
at participating health
centres in 1999
Study duration: 1 year
2-stage
Stage 1: day 10–24
TEOAE
Stage 2: 2 weeks later
TEOAE
Diagnostic stage: 2 weeks
later
dx audiology
Bilateral or unilateral failure if
TEOAE response <28 dB, or
response correlation <98%, or
SNR k did not reach target
level for 3 wavebands (3 dB at
0.8, 1.2 kHz; 6 dB at 2.4, 3.6,
4.0 kHz).
Coverage
Failure rates
LTFU
False alarm rates
Yield
Screening:
Fail to produce ≥ 70% total
reproducibility
Fail to produce ≥50% in the
1.6-kHz band and 70% in 2,4-,
3.2- and 4-kHz bands
Diagnostic ABR ≥40 dB HL
Failure rates
LTFU
False alarm rates
Yield
Not stated
Coverage
Yield
Tester: health visitors with
screening training
Environment: home or health
clinic for screens
(Pastorino et al 2005)
Level IV
QS=4/5
Instituti Clinici di
Perfezionamento
19,777 well and at-risk
babies
Milan, Italy
Study duration: not stated
3-stage (Well babies)
Stage 1: pre-discharge
TEOAE
Stage 2: 15-30 days after
discharge
TEOAE
Stage 3:
AABR
Diagnostic stage:
dx audiology
1-stage for at-risk babies
Tester: not stated
Environment: not stated
(Rao et al 2002)
Level IV
QS=2.5/5
5 small rural
hospitals, central
Minnesota
USA
Universal neonatal hearing screening
217 well babies born in
1999 and 2000
2-stage
Stage 1: 1 month
Study duration: 1 year
Stage 2: immediately
AABR
209
AABR
Diagnostic stage:
dx audiology
Tester: registered nurse with
screening training
Environment: not stated
(Rouev et al 2004)
Level IV
QS=4/5
Maternity Hospital
Stara Zagora
1750 well and at-risk
babies
Bulgaria
Study duration: 329 days
1-stage
Stage 1: (well babies 6-72 hrs
after birth; NICU babies 3-60
days
AABR
Stage 2: (3-4wks):
AABR
Diagnostic stage:
dx ABR
Bilateral fail >30 dB HL
Failure rates
LTFU
False alarm rates
Yield
Fail not stated
Failure rates
Yield
Tester: not stated
Environment: quiet room
adjoining the nursery
(Shoup et al 2005)
Level IV
QS=4.5/5
Large public
hospital
48,211 well and at-risk
babies
Dallas, Texas
Study duration: 3 years
4-stage (well babies)
Stage 1: newborn
AABR
Stage 2: 24 hrs post birth
AABR
Stage 3: pre-discharge by
audiologist
AABR
Stage 4: Outpatient
AABR
Diagnostic stage:
dx audiology
Tester: trained technicians
210
Universal neonatal hearing screening
Environment: not stated
(Swanepoel et al 2007)
Level IV
QS=4/5
Private hospital in
urban Gauteng
6,241 well and at-risk
babies
South Africa
Study duration: 4 years
2-stage
Stage 1: <discharge
TEOAE
Stage 2: 6 weeks later
TEOAE
Diagnostic stage:
dx OAE & ABR
Bilateral and unilateral
sensorineural hearing loss of
≥ 35dB
A 70% reproducibility and a 6dB signal-to-noise ratio at 2,3
and 4 kHz was set as pass
criteria
Coverage
Failure rates
False alarm rates
LTFU
Yield
Unilateral fail
Failure rates
LTFU
False alarm rates
Yield
Fail ≥ 35dB
Coverage
Failure rates
LTFU
False alarm rates
Yield
Tester: two qualified
audiologists
Environment: room with
acceptably low noise levels
(Tatli et al 2007)
Level IV
QS=4.5/5
Dokuz Eylul
University
Hospital, Izmir
711 well and at-risk babies
Study duration: 18 months
2-stage:
Stage 1: (last day of
discharge)
TEOAE
Turkey
Stage 2:
TEOAE
Diagnostic stage:
dx ABR
Tester: trained staff
Environment: quite room
adjacent to nursery
(Tsuchiya et al 2006)
Level IV
QS=4/5
Kumamoto
University
Hospital
Kumamoto Japan
8,979 well and at-risk
babies born during 19992004
Study duration: 5 years
2-stage:
Stage 1: day 4 post-partum
TEOAE
Stage 2: 1 month
TEOAE
Diagnostic stage:
dx audiology
Tester: not stated
Universal neonatal hearing screening
211
Environment: not stated
(Vohr et al 1998)
Level IV
QS=5/5
8 maternity
hospitals
53,121 well and at-risk
babies born 1993–1996
Rhode Island,
USA
Study duration: 4 years
2-stage
Stage 1: 6–52 hours (WBN);
< discharge (NICU)
NICU babies that failed the
screen went straight to
diagnostic evaluation
TEOAE
Stage 2: 2–6 weeks
TEOAE
Diagnostic stage:
dx ABR
Failure if absence of TEOAE
response between 2 and 4
kHz with 75% reproducibility
ABR fail if wave V not present
>30 dB nHL
Coverage
Failure rates
LTFU
False alarm rates
Yield
Tester: trained technicians,
nurses and nurses aides
Environment: nursery, parent
room or adjacent room
(WBN); NICU or adjacent
area (NICU)
(Watkin & Baldwin
1999)
Level IV
QS=4/5
Whipps Cross
hospital
28,890 babies born 1992–
1997
2-stage
Stage 1: < discharge
London, UK
Study duration: 6 years
Stage 2:
TEOAE
TEOAE
Diagnostic stage:
dx ABR
Bilateral fail on initial screen
and r-escreen, although
unilateral fail allowed if
obvious parental anxiety
Coverage
LTFU
False alarm rate
Yield
ABR fail >40 dB nHL in better
hearing ear
Tester: not stated
Environment: not stated
(Yee-Arellano et al
2006)
212
Level IV
QS=4/5
Private hospital in
San Pedro Garza
Garcia
3066 well and at-risk
babies
2-stage
Stage 1: < discharge
Study duration: 2 years
Stage 2: 3 weeks
ABR unilateral or bilateral fail
>35-40 dB nHL
AABR
Coverage
Failure rates
LTFU
False alarm rates
Universal neonatal hearing screening
Mexico
AABR
Diagnostic stage: not stated
Yield
Tester: Fully trained
technician
Environment: not stated
(Zaputovic et al 2005)
Level IV
QS=3.5/5
Rijeka University
Hospital
6,019 neonates
Study duration: 26 months
Croatia
2-stage
Stage 1: < discharge
CE-OAE
Stage 2: 3 weeks
CE-OAE
Diagnostic stage:
dx AABR
Fail not stated
Coverage
Failure rates
False alarm rates
Yield
Tester: not stated
Environment: not stated
Transient evoked otoacoustic emissions test; b diagnostic audiology; c sound pressure level; d hearing level; e loss to follow-up; f automated auditory brainstem response test; g diagnostic auditory brainstem response
test; h distortion product otoacoustic emissions test; I neonatal intensive care unit; j well baby nursery; k signal-to-noise-ratio.
a
Universal neonatal hearing screening
213
Included economic studies
Study
Location
Population
Score
Protocols – Comparisons
Outcomes
(Keren et al
2002)
Boston
Hypothetical
Massachusetts, cohort of
USA
80,000
infants
15/16
Modelled program
A. No screening
B. Targeted screening
Only babies with identified risk
factors as defined by the Joint
Committee on Infant Hearing
(2000)
2-stage screening
1st stage: AABRa then if fail:
2nd stage AABR
all prior to discharge
if fail: diagnostic evaluation by
an otorhinolaryngologist and
an audiologist
C. Universal screening
All newborns
2-stage screening
1st stage: TEOAEb then if fail:
2nd stage AABR
all prior to discharge
if fail: diagnostic evaluation by
an otorhinolaryngologist and
an audiologist
Costs considered
Short-term costs of screening
(equipment, consumables and
staff wages)
Follow-up diagnostic
assessment
Long term societal costs; loss
of productivity, requirement for
special education, vocational
rehabilitation and use of
assistive devices and medical
services
A.
Modelled program
A. Targeted screening
Only babies with identified risk
factors (time in neonatal
intensive care unit, family
history, presence of
craniofacial abnormality)
2-stage screening
1st stage: TEOAE then if fail:
2nd stage AABR
all prior to discharge
if fail: diagnostic evaluation by
an otorhinolaryngologist and
an audiologist
B. Universal screening
All neonates:
2-stage screening
1st stage: TEOAE then if fail:
2nd stage AABR
all prior to discharge
if fail: diagnostic evaluation by
A.
(Kemper &
Downs
2000)
214
Chapel Hill
North
Carolina,
USA
Hypothetical
cohort of
100,000
infants
13/16
B.
B.
Incremental cost
per infant
diagnosed by 6
months with
bilateral hearing
impairment ($US)
Incremental cost
reduction per child
with bilateral
hearing impairment
possessing normal
language ($US)
Cost per case of
bilateral hearing
impairment
detected ($US)
Cost per extra
case of bilateral
hearing impairment
detected ($US)
Universal neonatal hearing screening
an otorhinolaryngologist and
an audiologist
Costs considered
Not clearly specified
(Kezirian et
al 2001)
Seattle,
Washington
and Logan,
Utah, USA
Hypothetical
cohort of
2,000 births
in one
hospital per
year
13/16
(Gorga et
al 2001)
Omaha,
Nebraska,
USA
Hypothetical
cohort of
4,000 babies
(Boshuizen
et al 2001)
Netherlands
All newborn
children not
admitted to
neonatal
intensive care
units in the
Netherlands
Modelled program
4 protocols examined:
1) 1-stage AABR followed
by diagnostic evaluation if
infant failed screen
2) 2-stage AABR followed
by diagnostic evaluation if
infant failed both
screening tests
3)
1-stage OAEc followed
by diagnostic evaluation if
infant failed screen
4) 2-stage OAE followed by
diagnostic evaluation if
infant failed both
screening tests
Costs considered
Costs directly absorbed by the
hospital: equipment, personnel,
overhead and clerical
Follow-up diagnostic
evaluation
A.
12/16
Modelled program
Three screening protocols
examined:
1) 1-stage AABR with
follow-up assessment if
baby fails screen
2) 1-stage OAE with followup assessment if baby
fails screen
3) 2-stage starting with OAE
then AABR if baby fails
first screen. Follow-up
assessment if baby fails
AABR screen
Costs considered
Capital equipment (screening
tools)
Disposables
Salaries and benefits
Cost per baby screened
10.5/16
Modelled program
Community-based screening
program
2- and 3-stage screening
protocols
1) 2-stage OAE screening
1st stage: OAE then if fail: 2nd
stage also OAE
performed at different times at
community health centre
if fail: diagnostic evaluation
2) 2-stage AABR screening
1st stage: AABR then if fail:
2nd stage also AABR
performed at different times at
Cost per child screened
Cost per child detected
with a hearing loss of
≥40 dB or more in the
better ear
Universal neonatal hearing screening
B.
Cost per infant
screened
Cost per infant
identified with
unilateral or
bilateral hearing
loss
215
community health centre
if fail: diagnostic evaluation
3) 3-stage OAE screening
1st stage: OAE then if fail: 2nd
stage also OAE, third test of
OAE if fail results
performed at different times at
community health centre
if fail: diagnostic evaluation
4) detection of bilateral
losses only
2-stage OAE as above but
rescreened and evaluated only
if fail in both ears
5) home visits only
2-stage OAE as above but
screening performed on
special visits to baby’s home
Costs considered
Equipment depreciation
Personnel
Consumables
Travel costs of screeners
Screener training
Administration
Follow-up diagnostic testing
(Vohr et al
2001)
5 US
hospitals
(sites
undisclosed)
12,081
sequentially
tested, well
newborns
13.5/16
Existing program
3 UNHSd programs
1) 1-stage TEOAE
initial testing with TEOAE and
if fail tested again with TEOAE.
Fail referred for diagnostic
testing
2) 1-stage AABR
initial testing with AABR and if
fail tested again with AABR.
Fail referred for diagnostic
testing
3) 2-stage TEOAE then
AABR
initial testing with TEOAE and
if fail tested again with TEOAE.
If fail again then screened with
AABR. Finally a fail referred for
diagnostic testing
Note: screening was performed
by nurses but an audiologist
was required to interpret
TEOAE results
Costs considered
Staff: audiologist,
administration and clerical
assistance, audiologist training,
screener training
Equipment: TEOAE, AABR
Screening supplies
Cost per child born
Cost per child identified
with bilateral PCHIe
(Lemons et
al 2002)
Cleveland,
Ohio &
Indianapolis
Indiana, USA
1,530
newborns
tested using
TEOAE;
12/16
Existing program
2 UNHS programs
1) 2-stage TEOAE
unilateral fail one or more
Cost per infant
screened: with and
without follow-up
216
Universal neonatal hearing screening
1,412
newborns
tested using
AABR
times using the same
technology. Screen performed
by 2 masters student
audiologists and 5 other
screeners
2) 2-stage AABR
unilateral fail one or more
times using the same
technology. Screen performed
by 26 nursing staff
Costs considered
Personnel: audiologist, nurse
or audiology technician
screeners, administration and
clerical assistance
Equipment: TEOAE and
AABR, cart, printer and
personal computer
Other screening materials:
probes, tips, electrodes
Paper supplies
(Gorga et
al 2001)
Omaha,
Nebraska,
USA
4,460
newborns at
local birthing
hospital
(97.5%
coverage)
13.5/16
Existing program
For both well and NICUe
babies: 2-stage DPOAEf
followed by second DPOAE if
returned for fail in either ear
diagnostic ABRg if fail
All tests performed by clinical
audiologists
Costs considered
Equipment
Disposables
Salary and benefits
Follow-up diagnostic
assessment
Cost per baby screened
(Driscoll et
al 2000)
Brisbane, Qld
1,305 infants
of mean age
2 months
(1.5–2.5
months);
voluntary
enrolment
11/16
Existing program
1-stage TEOAE screening with
refer for diagnostic assessment
at auditory clinic if fail recorded
in either ear
Screening performed by
audiologist
Costs considered
Audiologist’s salary
Equipment
Maintenance of equipment
Replacement costs of TEOAE
probe (yearly)
Consumables
Costs averaged out over a
period of 6 years of screening
(using estimated 6-year life
span of equipment)
Cost per child entering
program
Cost of screening
protocol/ child
Cost of diagnostic
assessment protocol/
child
Cost per child identified
with sensorineural/
mixed hearing
impairment
Cost per child identified
with sensorineural/
mixed/conductive
hearing impairment
Total annual program
cost
(Stone et al
2000)
Rapid City,
South
Dakota, USA
1,002 well
newborns 6–
72 hours old
at first screen
11/16
Existing program
1-stage TEOAE performed in
hospital. Fail referred for
diagnostic assessment at 8
weeks of age using DPOAE
and tympanometry
A fail of both tympanometry
and DPOAE indicated middle
Cost per screen
Cost per infant
diagnosed with hearing
loss
Universal neonatal hearing screening
217
ear disease and treatment
A fail of only DPOAE resulted
in testing with AABR
Costs considered
DPOAE screening equipment
Full-time registered nurse
Administrator for 1 hour/week
Audiology interpretation of
results
Supplies
Follow-up assessment
(Maxon et
al 1995)
Rhode Island,
USA
4,253 infants
born during
6-month
period in
1993
595 special
care infants
3,658 well
infants
11/16
Existing program
2-stage TEOAE program
1st stage performed prior to
hospital discharge. If fail
recorded, then 2nd stage
rescreen 4–6 weeks later. ABR
performed by hospital
screeners if second fail
recorded. Fail of ABR referred
for diagnostic audiologic
evaluation
Costs considered
Personnel including trained
screening technicians,
audiologist, coordinator,
clerical assistance
Fringe benefits @ 28% of
salaries
Overhead @ 29% of salaries
Supplies
Cost of 3 TEOAE and 1 ABR
machine, 4 computers, 2
printers (all amortised over 5
years)
Cost per child screened
Cost per child identified
(Weirather
et al 1997)
Logan, Utah,
USA
380 well and
NICU
newborns
(98.4%
coverage)
11.5/16
Existing program
2-stage TEOAE
Newborns screened prior to
discharge. If fail recorded,
infants brought back to hospital
for stage 2 re-test with TEOAE.
If fail recorded then infant
referred to hospital audiology
department for diagnostic
assessment
Costs considered
Personnel: screening,
rescreening, screening
management, program
management, patient
management, scoring
Fringe benefits (30% of
salaries)
Supplies
Equipment (not screening
equipment – assumed already
in place)
Overhead (20% of costs)
Cost per baby screened
Automated auditory brainstem response test; b transient evoked otoacoustic emissions test; c otoacoustic emissions test; d universal
neonatal hearing screening; e permanent childhood hearing impairment; f neonatal intensive care unit; g distortion product
otoacoustic emissions test; h (conventional) auditory brainstem response test.
a
218
Universal neonatal hearing screening
Appendix G Excluded studies
Studies that met the inclusion criteria listed in the assessment protocol but on closer
inspection were excluded from review are listed below. Studies are grouped by the reason
for exclusion. ‘Data not useable’ indicates that the data on relevant outcomes presented
in the paper may have been in graphical format or summarised in such a way that raw
figures (means, standard deviations, numerators, denominators) could not be extracted.
‘Duplicated data’ means that the results were presented in some form or other in another
paper included in this assessment. ‘Not available’ means that the paper was requested via
inter-library loan services and could not be retrieved before the submission deadline of
this report. It is possible that these studies may not have been relevant (ie met the
inclusion criteria) in any event.
Prevalence
Data not useable
Cone Wesson, B., Vohr, B.R. et al (2000). ‘Identification of neonatal hearing impairment:
infants with hearing loss’. Ear and Hearing, 21 (5), 488–507.
Das, V K. (1990). ‘Prevalence of otitis media with effusion in children with bilateral
sensorineural hearing loss’. Archives of Disease in Childhood, 65 (7), 757–759.
Das, V.K. (1996). ‘Aetiology of bilateral sensorineural hearing impairment in children: A
10 year study’. Archives of Disease in Childhood, 74 (1), 8–12.
Davis, A., Wood, S. et al (1995). ‘Risk factors for hearing disorders: epidemiologic
evidence of change over time in the UK’. Journal of the American Academy of Audiology, 6 (5),
365–370.
Finitzo, T.A.K. & O’Neal J. (1998). ‘The newborn with hearing loss: detection in the
nursery’. Pediatrics, 102 (6), 1452–1460.
Fortnum, H.M., Marshall, D.H. & Summerfield, A.Q. (2002). ‘Epidemiology of the UK
population of hearing-impaired children, including characteristics of those with and
without cochlear implants – audiology, aetiology, comorbidity and affluence’. International
Journal of Audiology, 41 (3), 170–179.
Kanne, T.J., Schaefer, L. & Perkins, J.A. (1999). ‘Potential pitfalls of initiating a newborn
hearing screening program’. Archives of Otolaryngology Head and Neck Surgery, 125 (1), 28–32.
Kok, M.R., van Zanten, G.A. et al (1993). ‘Click-evoked oto-acoustic emissions in 1036
ears of healthy newborns’. Audiology, 32 (4), 213–224.
Kvaerner, K.J. & Arnesen, A.R. (1994). ‘Hearing impairment in Oslo born children 198991. Incidence, etiology and diagnostic delay’. Scandinavian Audiology, 23 (4), 233–239.
Maki Torkko, E.M., Lindholm, P.K. et al (1998). ‘Epidemiology of moderate to
profound childhood hearing impairments in northern Finland. Any changes in ten
years?’. Scandinavian Audiology, 27 (2), 95–103.
Universal neonatal hearing screening
219
Norton, S.J., Gorga, M.P. et al (2000). ‘Identification of neonatal hearing impairment:
summary and recommendations’. Ear and Hearing, 21 (5), 529–535.
Stewart, D.L. & Pearlman, A. (1994). ‘Newborn hearing screening’. Journal of the Kentucky
Medical Association, 92 (11), 444–449.
Uus, K. & Davis, A.C. (2000). ‘Epidemiology of permanent childhood hearing
impairment in Estonia, 1985-1990’. Audiology, 39 (4), 192–197.
Duplicated data
Johnson, J. L., White, K. R. et al (2005). 'A multicenter evaluation of how many infants
with permanent hearing loss pass a two-stage otoacoustic emissions/automated auditory
brainstem response newborn hearing screening protocol', Pediatrics, 116 (3), 663-672.
Kennedy, C. R., Kimm, L. et al (1998). 'Controlled trial of universal neonatal screening
for early identification of permanent childhood hearing impairment', Lancet, 352 (9145),
1957-1964.
Kennedy, C.R. (1999). ‘Controlled trial of universal neonatal screening for early
identification of permanent childhood hearing impairment: coverage, positive predictive
value, effect on mothers and incremental yield. Wessex Universal Neonatal Screening
Trial Group’. Acta Paediatrica, Supplementum, 88 (432), 73–75.
MacAndie, C., Kubba, H. & McFarlane, M. (2003). 'Epidemiology of permanent
childhood hearing loss in Glasgow, 1985-1994', Scott Med J, 48 (4), 117-119.
Watkin, P. (1996). ‘Neonatal otoacoustic emission screening and the identification of
deafness’. Archives of Disease in Childhood, 74 (1), F16–25.
Not available
Author(s) unknown (2000). ‘Identification of neonatal hearing impairment’. Ear and
Hearing, 21 (5), 345–535.
Diagnostic accuracy
Data not useable
Harrison, W.A., Dunnell, J.J. et al (2000). ‘Identification of neonatal hearing impairment:
experimental protocol and database management’. Ear and Hearing, 21 (5), 357–372.
Kennedy, C.R., Kimm, L. et al (1991). ‘Otoacoustic emissions and auditory brainstem
responses in the newborn’. Archives of Disease in Childhood, 66 (10 Spec No), 1124–1129.
Kennedy, C.R., Kimm, L. et al (1998). ‘Controlled trial of universal neonatal screening
for early identification of permanent childhood hearing impairment’. Lancet, 352 (9145),
1957–1964.
Stevens, J.C., Webb, H.D. et al (1987). ‘A comparison of oto-acoustic emissions and
brain stem electric response audiometry in the normal newborn and babies admitted to a
special care baby unit’. Clinical Physics and Physiological Measurement, 8 (2), 95–104.
Tognola, G., Ravazzani, P. et al (2001). ‘ “Linear” and “derived” otoacoustic emissions in
newborns: a comparative study’. Ear and Hearing, 22 (3), 182–190.
220
Universal neonatal hearing screening
Duplicated data
Kennedy, C.R. (1999). ‘Controlled trial of universal neonatal screening for early
identification of permanent childhood hearing impairment: coverage, positive predictive
value, effect on mothers and incremental yield. Wessex Universal Neonatal Screening
Trial Group’. Acta Paediatrica, Supplementum, 88 (432), 73–75.
Screening safety and effectiveness
Data not useable
De Ceulaer, G., Daemers, K. et al (1999). ‘Neonatal hearing screening with transient
evoked otoacoustic emissions: a learning curve’. Audiology, 38 (6), 296–302.
De Ceulaer, G., Daemers, K. et al (2001). ‘Neonatal hearing screening with transient
evoked otoacoustic emissions – retrospective analysis on performance parameters’.
Scandinavian Audiology, Supplementum, (52), 109–111.
Gabbard, S.A., Northern, J.L. & Yoshinaga-Itano, C. (1999). ‘Hearing screening in
newborns under 24 hours of age’. Seminars in Hearing, 20 (4), 291–305.
Gorga, M.P., Preissler, K. et al (2001). ‘Some issues relevant to establishing a universal
newborn hearing screening program’. Journal of the American Academy of Audiology, 12 (2),
101–112.
Korres, S., Balatsouras, D. G. et al, (2005). 'Overcoming difficulties in implementing a
universal newborn hearing screening program', Turk J Pediatr, 47 (3), 203-212.
Korres, S. G., Balatsouras, D. G. et al, (2006). 'Making universal newborn hearing
screening a success', Int J Pediatr Otorhinolaryngol, 70 (2), 241-246.
Lim, G. & Fortaleza, K. (2000). ‘Overcoming challenges in newborn hearing screening’.
Journal of Perinatology, 20 (8 Pt 2), S138–142.
Mehl, A.L. & Thomson, V. (1998). ‘Newborn hearing screening: the great omission’.
Pediatrics, 101 (1), E4.
Mehl, A.L. & Thomson, V. (2002). ‘The Colorado newborn hearing screening project,
1992-1999: on the threshold of effective population-based universal newborn hearing
screening’. Pediatrics, 109 (1), E7.
Mathur, N. N. & Dhawan, R., (2007). 'An alternative strategy for universal infant hearing
screening in tertiary hospitals with a high delivery rate, within a developing country, using
transient evoked oto-acoustic emissions and brainstem evoked response audiometry',
Journal Of Laryngology And Otology, 121 (7), 639-643.
Owen, M., Webb, M. & Evans, K. (2001). ‘Community based universal neonatal hearing
screening by health visitors using otoacoustic emissions’. Archives of Disease in Childhood,
84 (3), F157–162.
Parving, A. & Salomon, G. (1996). ‘The effect of neonatal universal hearing screening in
a health surveillance perspective – a controlled study of two health authority districts’.
Audiology, 35 (3), 158–168.
Universal neonatal hearing screening
221
Stuart, A., Moretz, M. & Yang, E.Y. (2000). ‘An investigation of maternal stress after
neonatal hearing screening’. American Journal of Audiology, 9 (2), 135–141.
Swedish Council on Technology Assessment in Health, C. (2004). 'Universal newborn
hearing screening - early assessment briefs (Alert) (Structured abstract)'.
Tait, K. (1997). ‘A model for success: creating the model for rural success for newborn
hearing screening’. Neonatal Intensive Care, 10 (5), 34–35.
Wroblewska-Seniuk, K., Chojnacka, K. et al, (2005). 'The results of newborn hearing
screening by means of transient evoked otoacoustic emissions', Int J Pediatr
Otorhinolaryngol, 69 (10), 1351-1357.
Duplicated data
Barry, H. (2000). ‘Is screening for newborn hearing loss cost-effective?’ Evidence Based
Practice, 3 (8), 8, insert 2p NLI: 100894030.
Corabian, P., Eng, K. et al (2007). IHE Report: Screening Newborns for Hearing,
Insititute of Health Economics, Alberta, Canada.
Lin, H., Shu, M. et al (2003). 'Establishing a newborn hearing screening programme in
Taiwan', Asia Pacific Journal of Speech, Language & Hearing, 8 (3), 174.
Lin, H. C., Shu, M. T. et al (2002). 'A universal newborn hearing screening program in
Taiwan', International Journal of Pediatric Otorhinolaryngology, 63 (3), 209-218.
Prieve, B.A. & Stevens, F. (2000). ‘The New York State universal newborn hearing
screening demonstration project: introduction and overview’. Ear and Hearing, 21 (2), 85–
91.
Slawson, D. (2002). ‘Can universal newborn hearing screening improve language
outcomes in children identified with profound bilateral permanent hearing loss (PHL)?’
Evidence Based Practice, 5 (1), 8, insert 2p NLI: 100894030.
Spivak, L., Dalzell, L. et al (2000). ‘New York State universal newborn hearing screening
demonstration project: inpatient outcome measures’. Ear and Hearing, 21 (2), 92–103.
White, K.R., Vohr, B.R. & Behrens, T.R. (1993). ‘Universal newborn hearing screening
using transient evoked otoacoustic emissions: Results of the Rhode Island hearing
assessment project’. Seminars in Hearing, 14 (1), 18–29.
White, K.R., Vohr, B.R. et al (1994). ‘Screening all newborns for hearing loss using
transient evoked otoacoustic emissions’. International Journal of Pediatric Otorhinolaryngology,
29 (3), 203–217.
Not available
De Capua, B., Tozzi, A. et al (1999). ‘Neonatal audiological screening with transient
evocated otoacoustic emission (TEOAE): Results and comments about one years of
experience’. Otorinolaringologia Pediatrica, 10 (1–2), 5–7.
222
Universal neonatal hearing screening
El-Naggar, M. & Hashlamoun, M., (2005). 'Paediatric hearing assessment and screening
clinic at Fujairah: Analysis of the results of the first 6 months of clinic practice', Emirates
Medical Journal, 23 (1), 15-20.
Jacobson, C.A. & Jacobson, J.T. (1990). ‘Follow-up services in newborn hearing
screening programs’. Journal of the American Academy of Audiology, 1 (4), 181–186.
Joseph, R., Tan, H. K. et al, (2003). 'Mass newborn screening for hearing impairment',
Southeast Asian J Trop Med Public Health, 34 Suppl 3, 229-230.
Mahulja-Stamenkovic, V., Prpic, I. et al, (2005). 'Incidence of hearing loss assessed by the
universal newborn hearing screening in the region of Rijeka', Paediatria Croatica, 49 (4),
219-222.
Molteni, G., (2006). 'Early detection of newborn hearing impairment by transiently
evoked otoacoustic emissions and auditory evoked potentials. Personal experience in
10,454 children', Otorinolaringologia, 56 (2), 93-96.
Paul, A. K., (2003). 'Hearing loss in children: Need for early detection and intervention',
Indian Journal of Practical Pediatrics, 5 (4), 353-355.
Quintos, M. R., Isleta, P. F. et al, (2003). 'Newborn hearing screening using the evoked
otoacoustic emission: The Philippine General Hospital experience', Southeast Asian J
Trop Med Public Health, 34 Suppl 3, 231-233.
Universal neonatal hearing screening
223
Appendix H
Guidelines for using screening
devices
Extract from (Bailey 2003).
Use of the Natus Algo Portable AABR Screener
This protocol describes the use and care of the Natus Algo Portable AABR Screener in
the Western Australian Newborn Hearing Screening Programme. Further information is
found in the user manual. An instructional video is available.
Contents
1.
Description of the Natus Algo Portable AABR Screener
2.
How does it work?
3.
Associated equipment
4.
How to screen a baby
5.
Infection control and how to clean the machine after use
6.
Battery charging
7.
Routine checks
1. Description of the Natus Algo Portable AABR Screener
The Natus Algo Portable AABR Screening equipment consists of:
•
the screening unit
•
attached printer
•
attached earphone cable and sensor cable.
Each unit has a user manual; storage bag; charging unit; Acoustic Check kit; and small
screwdriver.
2. How does it work?
The machine measures auditory brainstem responses. Earphones attached to the baby’s
ear deliver clicking noises set at 35dB nHL. Sensors that are attached to the baby’s
forehead, nape and shoulder record the brainwave response to the stimuli. The screener
compares the baby’s responses to the stimuli to a template derived from other babies
(AABR). It measures the probability that an adequate ABR signal is present. The
machine indicates whether a Pass response has been obtained, otherwise further
224
Universal neonatal hearing screening
screening or diagnostic testing is indicated. The machine is not a diagnostic machine and
cannot be used to indicate the degree of hearing loss.
A Pass response indicates normal middle ear, cochlear and retro-cochlear function.
If a Pass response is not obtained, further screening is required. Reasons that a Pass
response will not be obtained include:
•
the ear cups are not attached to the baby’s ear or are compressing the ear canal.
•
there is too much noise (ambient light on)
•
the baby is too awake or active (myogenic light on)
•
the sensors are not in the correct places or are not well adhered to the skin (high
impedance)
•
there is other interference such as other equipment in the area or the sensor or its
wires are in contact with other equipment
•
there is a hearing loss.
The screen will generally run rapidly if the baby is settled, and the impedance is low. The
screening device is designed for use in babies between 34 weeks gestational age and 6
months only.
3. Associated equipment
It should be used with disposable Algo Paks (set of ear cups and 3 jelly sensors) supplied
by Natus only.
For skin preparation:
•
mild baby liquid soap
•
gauze squares
For cleaning:
•
alcohol wipes/other cleaning agent
4. How to screen a baby
It should be done in a quiet room when the baby is asleep or in a quiet state. Babies can
be screened while feeding although this will prolong the screen. This should only be
attempted when feeding has been well established and the mother and baby are
comfortable with the feeding process. At the beginning of a feed, babies may make too
much noise but generally they quieten down as the feed progresses. The screen will
progress very slowly if the baby is wide awake and looking around even if the baby is
quiet.
Universal neonatal hearing screening
225
•
Always tell parents the reasons why a result may not be obtained prior to
screening.
•
Use the term ‘stickers’ rather than electrodes when explaining the screening
process to parents.
•
Switch the machine on prior to attaching sensors to the baby and remove the
sensors before switching off the machine.
•
In many cases, it is easier to leave the machine in the carry bag while the screen is
being conducted.
1.
See the user manual for full instructions how to perform the screen and
for sensor placement.
2.
The better the sensor placement and contact with the skin, the faster the
screen will run.
3.
If the baby needs skin preparation, use only mild liquid baby soap on a
gauze square moistened with warm water. Be very gentle as babies have
delicate skin. Moistening the sensor and holding it in place for a few
seconds may improve contact and lower impedance.
4.
Place the sensors as follows
5.
1.
white - nape
2.
green - shoulder (common)
3.
black - high on forehead
Attach the earphones to the acoustic cables when still adhered to the
paper card prior to attaching to the baby:
1.
red - right ear
2.
blue - left ear
When attached to baby, the cables should emerge towards the top of the
baby’s head.
226
6.
Commence the screen as instructed in user manual.
7.
During the screen, the display shows the number of good sweeps
completed (SWP) and the LR (Likelihood ratio). The screen is complete
when the LR reaches at least 160. The minimum number of sweeps is
1,000 and the maximum is 15,000. The LR is updated every 500 sweeps.
8.
If the LR is not increasing after about 2000 sweeps and the baby is quiet,
stop the screen and check the ear cup is correctly positioned and the
acoustic cable still attached.
Universal neonatal hearing screening
9.
If the myogenic light is lit, try and settle baby. Check also if any sensor
wires are in contact with each other.
10.
If impedance is a problem, try re-applying the sensors. A new set of
sensors may be required.
11.
Sometimes other medical equipment may interfere with the screen.
Consult with nursing staff to see if this can be corrected.
12.
Babies can be screened while in incubators but the screen may progress
very slowly.
13.
When the screen is completed, use gauze moistened with warm water to
help remove the sensors and ear cups from the baby.
5. Infection control and how to clean the machine after use
•
Wash hands thoroughly before and after handling the baby.
•
Use a new set of disposable sensors and ear couplers for each baby.
•
At the completion of the test, discard the sensors and ear couplers. Wipe the sensor
cables and acoustic transducers with an alcohol swab and allow them to dry.
•
Carefully roll up the cables and store inside the carry case.
6. Battery charging
When fully charged the battery will last for approximately 10 hours of continuous
screening. See the user manual for instructions on how to charge the battery. The
machine should be charged overnight on a regular basis. To prolong the battery life, the
battery should have a deep discharge prior to recharging. As this takes a lengthy period
of time, it is best not to do it until there are 2 or less of the 4 battery charge level lights lit
on the machine. At smaller sites, it may be sufficient to charge the battery about every 46 weeks.
7. Routine checks and maintenance
•
Always inspect the acoustic and sensor cables prior to use: Attach the sensor clips to
the Acoustic Check kits, start the machine and listen to the acoustic cables to hear if
clicks are present.
•
Perform the equipment checks listed in the user manual weekly and record that it has
been done in the record book. If the machine passes these checks, it can be used. If
not, do not use it and contact the coordinator.
•
See the user manual for instructions on how to load a new printer roll. Ensure that
the machine is switched off after loading the paper in order to preserve the battery.
•
The Algo portable screener and charging unit should have a hospital safety check
annually and before being used after being repaired.
Universal neonatal hearing screening
227
•
The acoustic cables require yearly calibration by Natus. Each set of cables has a
calibration certificate with the date when the calibration is due. Contact Scanmedics
to arrange the loan of replacement cables at least 2 months before this is due.
Extract from (Bailey 2003).
Use of the ECHOCHECK hand-held ILO OAE Screener
This protocol describes the use and care of the Echocheck in the Western Australian
Newborn Hearing Screening Programme. Further information is found in the user
manual.
Contents
1.
Description of the Echocheck
2.
How does it work?
3.
Associated equipment
4.
How to screen a baby
5.
Infection control and how to clean the machine after use
6.
Battery charging
7.
Routine checks
1. Description of the Echocheck
The Echocheck consists of a hand held, battery operated unit; and a detachable cable and
probe which uses disposable ear tips for each use. The probe uses replaceable coupler
tubes (changed when contaminated). Replacement probe/coupler covers are available.
Each unit has a user manual; storage bag, charging unit; test cavity and test plug.
2. How does it work?
The machine measures Transient Evoked OtoAcoustic Emissions (TEOAEs or OAEs
for short) and it concentrates on the main speech frequency band of 1.6-3.6Hz. The
Echocheck is a screening device so no interpretation is required. The green light
indicates a pass response, otherwise further screening is indicated. The machine is not a
diagnostic machine and cannot be used to indicate the degree of hearing loss.
TEOAEs are low intensity sounds originating from the outer hair cells of the cochlear in
response to stimuli. A probe that contains a very small microphone is placed in the
baby’s ear. When switched on a continuous clicking noise is emitted from the probe (the
stimulus). The microphone then detects the response. By various processes, the
machine sorts out TEOAEs from other noises. If the TEOAEs are at least 6dB louder
than other noises, a green pass light is obtained. If they are between 3-6dB louder, an
228
Universal neonatal hearing screening
orange light is obtained and the screen should be repeated later, preferably in quieter
conditions.
A pass response indicates normal middle ear and cochlear function.
If a pass response is not obtained, further screening is required. A pass response will not
be obtained if :
•
the probe does not fit in the ear canal properly
•
there is too much noise (either from the baby, such as noisy breathing, or the general
environment)
•
there is debris in the ear canal such as vernix or other birth products or wax
•
there is middle ear fluid
•
there is a hearing loss.
The younger the baby is, the more likely they are to have transient middle ear fluid and
debris in the ear canal. We have found that babies have less than a 60% chance of
passing if less than 24 hours old. By the time the baby is about 3 days old or more, they
have about a 90% or more chance of passing.
3. Associated equipment
•
Three sizes of disposable ear tips and spare coupler tubes should be available at all
times for hospital-based newborn screening:
Tip sizes -
•
1.
T4.5C most common used
2.
T5.5B for larger/older babies
3.
T3E for very small babies
4.
Community based programmes may require larger size tips.
Alcohol wipes/other cleaning agent
These should be stored in a suitable storage container such as a small, divided plastic craft
box.
4. How to screen a baby
To optimise the chances of a baby passing, the screen should be done as late as possible
during the hospital stay. The exact timing will depend on the length of stay in hospital at
individual hospitals. Women who have had a caesarean section generally have a longer
stay than those who have a vaginal delivery so the screening can be delayed.
It should be done in a quiet room when the baby is asleep or in a quiet state. Babies can
be screened while feeding. This should only be attempted when feeding has been well
Universal neonatal hearing screening
229
established and the mother and baby are comfortable with the feeding process. At the
beginning of a feed, babies may make too much noise but generally they quieten down as
the feed progresses.
•
Always tell parents the reasons why a result may not be obtained (eg noise, debris
etc) prior to screening.
•
See the user manual for instructions how to perform the screen.
•
If the Stimulus OK light is not illuminated, check if the cable is properly attached to
the machine and check if there is debris blocking the ear tip.
•
Debris in the ear tip can be removed by taking the ear tip off the probe and
cleaning/squeezing it with an alcohol swab. If necessary, a new ear tip can be used.
•
Check that the coupler tubes are clear and change them if necessary.
•
If there is debris in the ear, sometimes this can be removed by reinserting the cleaned
ear tip/probe in the ear, rotating the tip/probe in the ear canal and repeating the
cleaning process.
5. Infection control and how to clean the machine after use
•
Wash hands thoroughly before and after handling the baby.
•
Use a new ear tip for each baby.
•
After completion of the test, remove used ear tip from the probe casing and discard
it.
•
Wipe the probe casing, cable and control panel with an alcohol swab and allow them
to dry. If necessary, dry with a tissue.
•
Check that the coupler tubes are free of debris before and after use. If debris is
present, discard the contaminated coupler tubes and insert new ones (see user
manual).
6. Battery charging
The Echocheck has an internal, rechargeable battery. When fully charged, the battery
will last for about 6 hours of continuous screening. The machine should be charged
overnight on a regular basis. To prolong the battery life, the battery should be allowed to
‘run down’ before being recharged. At smaller sites, it may be sufficient to charge the
battery about once a fortnight. Instructions on how to charge the battery are in the user
manual.
7. Routine checks and maintenance
Always inspect the Echocheck, probe and its cable before use:
230
Universal neonatal hearing screening
•
check that the coupler tubes are free of debris before and after use. If debris is
present, discard the contaminated coupler tubes and insert new ones (see user
manual).
•
if the back of the probe casing is missing, replace the whole casing.
•
if the cable has white wiring visible near the probe, obtain a new probe. Careful
handling of the probe and its cable will prolong its life.
•
if the front panel of the machine is cracked, contact the coordinator to arrange to
have this replaced.
Routine checks should be carried out weekly or after the coupler tubes or probe is
changed. Record the results of the tests in the Echocheck record book. See the user
manual for details of the probe checks. If the machine passes these checks it can be
used. If not, do not use it and contact the coordinator.
The Echocheck and charging unit should have a hospital safety check annually and
before being used after being repaired.
Universal neonatal hearing screening
231
Appendix I
Analysis of published
economic evaluations (up
to 2003)
Comparison of universal and targeted screening in the short term
Five studies presenting economic models of universal neonatal hearing screening
(UNHS) were identified and have been critically appraised.
Two of these studies compared targeted and universal screening (Kemper & Downs
2000; Keren et al 2002). A summary of the costs included in the evaluation of UNHS
and targeted screening protocols in both studies is provided in Table A.
Table A
Resource items considered in studies comparing modelled universal neonatal and targeted
hearing screening programs
Resource items
Targeted screening
2-stage AABR
(Keren et al 2002)
Capital equipment
Universal screening
2-stage TEOAE–
AABR (Kemper &
Downs 2000)
2-stage TEOAE–
AABR
(Keren et al 2002)
A
A
OAE instrument
n/a
9
AABR instrument
9
9
Computer
8
8
Recurrent items
A
A
OAE probes
n/a
9
ABR electrodes
9
9
OAE probe tips
n/a
9
Ear couplers/muffins
9
9
Cables
9
9
Machine calibration
9
9
Other consumables (eg paper)
8
8
Screener training
8
8
Overheads
8
8
Personnel
2-stage TEOAE–
AABR (Kemper &
Downs 2000)
A
A
Nurse or other trained screener
9
9
Audiologist
9
9
Administrative support
9
9
Follow-up
9
9
9
9
Patient and family
9
8
9
8
Other sectors
9
8
9
8
A = assumed, indicated as included but not described; AABR = automated auditory brainstem response test; TEOAE = transient evoked
otoacoustic emissions test; OAE = otoacoustic emissions; 9 = included; 8 = not included; n/a = not applicable.
In the higher quality study (Keren et al 2002) cost-effectiveness was evaluated from a
societal perspective for selective (targeted) and universal screening in a hypothetical
cohort of 80,000 newborns from one US state. This cohort consisted of 10,400 high-risk
and 69,600 low-risk infants. Under targeted screening only infants with identified risk
232
Universal neonatal hearing screening
factors for congenital deafness were screened (Box 10). The decision tree for this model
is shown in Figure 11.
Box 10
Risk factors for congenital deafness (Joint Committee on Infant Hearing
2000)
Residence in neonatal intensive care unit / special care baby unit for ≥48 hours
Prolonged usage of aminoglycosides
Family history of permanent childhood deafness
Craniofacial abnormality noticeable at birth
Perinatal infection (either suspected or confirmed)
Birthweight <1.5 kilograms
Birth asphyxia
Chromosomal abnormality, including Down syndrome
Exchange transfusion or intrauterine transfusion
Figure 11 Decision tree for universal and targeted neonatal hearing screening model
(Keren et al 2002)
Universal neonatal hearing screening
233
Based on literature from published universal neonatal hearing screening (UNHS)
programs reporting an estimate of 1.6 cases of permanent childhood hearing impairment
(PCHI) per 1,000 births, the cohort was assumed to contain 128 infants with permanent
hearing impairment (Aidan et al 1999; Clemens et al 2000; Kennedy et al 1998; Mason &
Herrmann 1998). The point estimates were based on an 80 per cent coverage of high-risk
infants through targeted screening and 100 per cent coverage for all infants using UNHS.
A referral rate for diagnostic evaluation of 0.18 per cent was assumed for targeted
screening and 1.6 per cent for UNHS. For both programs it was assumed that 77 per
cent of babies referred for diagnostic assessment were followed up (Prieve et al 2000).
Yield from the two screening programs was estimated at 48 and 91 per cent for targeted
and universal screening, respectively. Resources considered in the analysis were those of
screening (equipment, consumables and staff wages); follow-up diagnostic assessment;
and the societal costs of loss of productivity, requirement for special education,
vocational rehabilitation and use of assistive devices and medical services (Table A)
(Vohr et al 1998).
The incremental cost of diagnosing an infant with bilateral PCHI by 6 months of age was
found to be higher with UNHS (Table B). Sensitivity analysis identified two variables
with the greatest effects on the incremental cost. Reducing the coverage of high-risk
infants in targeted screening from 100 per cent to 40 per cent reduced the incremental
cost-effectiveness ratio (ICER) of universal screening compared to targeted screening
from $76,000 to $45,000. Increasing the diagnostic follow-up in both programs from 50
per cent to 100 per cent increased the ICER from $52,000 to $77,000 per child identified
by 6 months of age using UNHS compared to targeted screening. The authors state that
the higher incremental cost of UNHS is due largely to the greater number of patients
referred for diagnostic evaluation. In targeted screening the number of patients referred
was estimated to be 145 of 80,000 babies (0.2%), while use of UNHS would lead to
referral of nearly 10 times that number at 1,314 (1.6%). Despite the short-term costs of
UNHS being substantially higher than those for the targeted or no screening options,
estimates of long-term costs resulted in savings using UNHS over the lifetime of a child
who developed normal language. The inclusion of long-term societal costs resulted in an
incremental saving of $2 million per child with normal language using targeted screening,
and a further saving of $1.2 million per child using UNHS.
234
Universal neonatal hearing screening
Table B
Study
(Keren et al
2002)
Published economic models of universal vs targeted neonatal hearing screening
Population
Hypothetical
cohort of
80,000
infants from
one US State
10,400 highrisk, 69,600
low-risk
infants
Quality
Ratio
Incremental cost per
infant diagnosed by 6
months with bilateral
hearing impairment
Hypothetical
cohort of
100,000
infants
0
Targeted
screening
Universal
screening
2-stage AABR
2-stage OAE
and AABR
22,820
61,600b
Varying coverage of
high-risk infants
from 100% to 40%
45,800–76,400
Varying follow-up
diagnostic
evaluation from
50% to 100%
52,800–77,800
Incremental cost
reduction per child
with bilateral hearing
impairment
possessing normal
language [see also
Table J]
(Kemper &
Downs 2000)
Protocol / Results ($AUS)
No screening
15/16
13/16
Cost per case of
bilateral hearing
impairment detected
Cost per extra case of
bilateral hearing
impairment detected
0
2,030,000a
1,212,000b
Targeted screening
Universal screening
2-stage TEOAE and AABR
2-stage TEOAE and AABR
4,640
17,350
35,650
a ICER targeted vs no screening; b ICER UNHS vs targeted (assuming 80% of high-risk patients screened in targeted screening scenario and
77% of infants followed up with diagnostic assessment after positive screen in both scenarios). AABR = automated auditory brainstem
response test; OAE = otoacoustic emissions test; TEOAE = transient evoked otoacoustic emissions test.
The other study of the cost-effectiveness of targeted versus universal screening was
modelled from a health care perspective in a population of 100,000 hypothetical
newborns (Kemper & Downs 2000). The decision tree for this model is presented in
Figure 12.
Based on the results of other evaluations of neonatal hearing screening (Davis et al 1997;
Maxon et al 1995) and a population-based surveillance program (Van Naarden et al
2000), the prevalence of bilateral hearing loss was estimated to be 0.11 per cent or 1.1 per
1,000 newborns (range 0.10–0.59%). From the model, targeted screening resulted in 67
referrals for diagnostic assessment from the total cohort of 100,000 (0.07%). UNHS
resulted in 406 referrals or 0.4 per cent of the total cohort. Cost analysis was performed
from a health care perspective. If the cost of risk screening was assumed to be $1 per
child screened (range $0.5–$15), the authors determined that for UNHS the use of
automated transient evoked otoacoustic emissions (TEOAE) testing would cost $7.42
(range $5–$15) and automated auditory brainstem response (AABR) testing $25 (range
$15–$40) per child screened. Diagnostic auditory brainstem response (ABR) testing
would cost $150 (range $100–$200) per child sent for assessment. Only the process
measures of cost per case identified and cost per extra case identified were reported.
Short- or long-term outcomes were not considered. In agreement with the previous
study (Keren et al 2002), the authors determined that the incremental cost per case of
Universal neonatal hearing screening
235
identifying bilateral PCHI was more expensive with UNHS than for targeted screening
(Table B). However, it was predicted that UNHS would detect 40 per cent more PCHI
cases compared to targeted screening. The resulting cost of $35,650 per extra case
detected would be due largely to the increase in the number of false positives (from 16 to
320) that would be referred for follow-up diagnostic evaluation.
Figure 12 Decision tree for universal and targeted neonatal hearing screening model
(Kemper & Downs 2000)
Summary of cost-effectiveness of modelled universal and targeted neonatal hearing
screening programs, contrasting the short term and the long term
While both studies determined that universal neonatal hearing screening would, in the short
term, be more expensive and less cost-effective to operate than a targeted screening
program, modelling of long-term costs by Keren and colleagues (2002) suggested that
identifying a larger proportion of hearing-impaired infants at an early stage (ie ≤6 months of
age) would result in savings in other sectors (such as education and social welfare) that
would far outweigh the initial expense. Disparity in the incremental cost-effectiveness ratio
between these two studies, for incremental cost per case identified, would probably stem
from the different referral rates used to determine the costs of follow-up diagnostic
assessment. In the study by Kemper & Downs (2000), the assumed difference in referral
rates from universal, as opposed to targeted, hearing screening programs was much smaller
than that assumed by Keren and colleagues (2002). With similar assumed differences in
referral rates, the overall cost of diagnostic follow-up for each of the programs presented by
Kemper & Downs (2000) would be closer, reducing the incremental cost.
236
Universal neonatal hearing screening
Comparison of modelled universal neonatal hearing screening programs in
the short term
Three identified studies compared the costs and outcomes of models of different
protocols for universal neonatal hearing screening (UNHS) (ie automated auditory
brainstem response (AABR) vs otoacoustic emissions (OAE)) (Boshuizen et al 2001;
Gorga et al 2001; Kezirian et al 2001). All studies took a health care sector perspective. A
summary of the costs considered in all screening protocols is presented in Table C.
The highest quality study examined the cost effectiveness of four different screening
methods: one-stage AABR and OAE–AABR, and two-stage AABR and OAE on a
hypothetical cohort of 2,000 babies born in one hospital in 1 year (Kezirian et al 2001).
The one-stage OAE–AABR technique involved performing an AABR directly after a
positive (fail) result on the initial OAE test. The decision tree for this model is presented
in Figure 13.
Table C
Resource items considered in studies comparing different modelled universal neonatal
hearing screening programs
Resource item
Capital equipment
(Kezirian et al 2001)
1-stage AABR, 1 stage
OAE–AABRa, 2-stage OAE,
2-stage AABR
(Gorga et al 2001)
1-stage AABR, 1-stage OAE,
2-stage OAE–AABR
Ab
A
(Boshuizen et al 2001)
2-stage OAE, 2-stage AABR,
3-stage OAE
OAE instrument
9
AABR instrument
9
Computer
8
Recurrent items
A
A
A
OAE probes
ABR electrodes
OAE probe tips
Ear couplers/muffins
Cables
Machine calibration
Other consumables
(eg paper)
Screener training
Overheads
Personnel
9
9
A
A
A
Follow-up
9
9
9
Patient and family
8
8
8
Other sectors
8
8
8
Nurse or other trained
screener
Audiologist
Administrative support
9
Protocol considered 1-stage but AABR performed immediately after OAE fail result; b A = assumed (indicated as included but not described).
AABR = automated auditory brainstem response test; ABR = auditory brainstem response test; OAE = otoacoustic emissions; 9 = included;
8 = not included.
a
Universal neonatal hearing screening
237
Figure 13 Decision tree for modelled universal neonatal hearing screening protocols
(Kezirian et al 2001)
Based on previous studies (Watkin et al 1990; White 1997), the prevalence of combined
bilateral or unilateral hearing loss was estimated to be 3.5 per 1,000 infants. Referral rates
of each program were not provided but were used in cost calculations. Costs included
were those directly absorbed by the hospital such as equipment, personnel, overheads
and clerical, and the cost of follow-up diagnostic evaluation (Table C). Costing
information was obtained from a report compiled by the National Centre for Hearing
Assessment and Management (NCHAM) at Utah State University (NCHAM 2003). The
method with the least cost per infant screened was found to be two-stage OAE (Table
D), which was also the most cost-effective in terms of the cost per infant identified with
unilateral or bilateral permanent childhood hearing impairment (PCHI). The least costeffective method was a one-stage AABR. These results were consistent when best case –
worst case scenarios were examined. Although the one-stage AABR was determined as
identifying the greatest number of infants with hearing loss, analysis revealed that more
than $126,000 of extra funding would be required under this protocol in order to identify
one more case than with two-stage OAE (sic). However, exclusion of infants lost to
follow-up from effectiveness calculations, and the use of sensitivity and specificity values
from studies examining targeted (ie ‘at-risk’) infants means that these results are likely to
be conservative and should be regarded with caution.
In a hypothetical cohort of 4,000 babies, the cost-effectiveness of one-stage AABR, onestage OAE or two-stage OAE followed by AABR was modelled (Gorga et al 2001). The
decision tree for this model is presented in Figure 14.
238
Universal neonatal hearing screening
Table D
Published economic models of various universal neonatal hearing screening protocols
Study
Population
Kezirian et
al 2001
Hypothetic
al cohort of
2,000 births
in one
hospital in
1 year
Quality
Ratio
2-stage
AABR
1-stage
OAE–AABRa
2-stage OAE
35
28
28
18
Best case –
worse caseb
26–70
25–46
20–45
14–28
Cost per
infant
identified with
hearing loss
13,350
11,430
11,270
7,200
9,680–28,020
9,370–20,290
7,610– 20,230
5,240–12,560
Cost per
infant
screened
Best case –
worse caseb
Gorga et al
2001
Boshuizen
et al 2001
Hypothetic
al cohort of
4,000
babies born
in 1 year
All newborn
children not
admitted to
NICUe
Protocol / Results ($AUS)
1-stage
AABR
13/16
1-stage AABR
12/16
1-stage OAE
2-stage OAE–
AABR
Cost per
baby
screenedc
32
33
24
Sensitivity
analysisd
27–218
31–104
16–280
10.5/16
2-stage
OAE
(CHCf)
2-stage
AABR
(CHC)
3-stage
OAE
(CHC)
2-stage
OAE
(CHC)g
2-stage
OAE
(home)
2-stage
OAE
(CHC +
home)
Cost per child
screened
29
41
25
23
33
28
Sensitivity
analysish
±3
±4
±2
±2
±4
±2
Cost per child
detected with
hearing loss
≥40 dB in the
better ear
40,880
57,380
35,240
33,130
45,250
38,630
Sensitivity
± 9020
± 13250
± 7610
± 8310
± 11560
± 8450
analysish
a Considered 1-stage in hospital: AABR test performed if baby fails initial OAE; b worst case scenario when costs maximised and sensitivity/
specificity of the protocol is minimised; in the best case scenario the values are the reverse; c includes costs of follow-up for 1-stage AABR and 2stage OAE–AABR of 2% and for 1-stage OAE of 8%; d based on number of births per year ranging from 8,000 to 25; e NICU = neonatal intensive
care unit; f CHC = child health clinic; g detection of bilateral losses only; h sensitivity expressed as the standard error determined from 500
simulations using Monte-Carlo simulation model. AABR = automated auditory brainstem response test; OAE = otoacoustic emissions test.
Universal neonatal hearing screening
239
Figure 14 Decision tree for modelled universal neonatal hearing screening protocols
(Gorga et al 2001)
Coverage was assumed to be 100 per cent and the referral rate for diagnostic assessment
was assumed to be 2 per cent for AABR and 8 per cent for OAE testing (no source
given). Two-stage OAE with AABR was assumed to have a final referral rate that was
the same as one-stage AABR, namely 2 per cent. Costs included in this study were for
capital equipment (screening tools), disposables, salaries and employee benefits (Table
C). Follow-up costs of babies returning a positive screening test for bilateral PCHI were
estimated at $280 per baby. Based only on the cost per baby screened, the two-stage
OAE followed by AABR was found to be the most cost-effective (Table D). Although
the follow-up rate was assumed to be the same for both one-stage AABR and two-stage
OAE–AABR (2%), the six-fold greater expense of disposables used in screening 4,000
newborns initially with AABR ($40,000 vs $7,200 for the 2-stage) makes this method less
cost-effective than an initial screen with OAE. Sensitivity analysis varied the number of
babies born in the hospital in a 5-year period (the assumed useful life of the screening
equipment) from between 25 and 8,000 babies per year. As expected, using any of the
programs became more expensive if fewer babies were born to the hospital. This is also
illustrated when comparing similar programs, as modelled by Kezirian et al (2001) using a
birth rate of 2,000 babies per year.
In the only non-US study, Boshuizen and colleagues (2001) modelled the costeffectiveness of six different existing programs aimed at identifying both unilateral and
bilateral hearing impairment in an unspecified cohort of newborn infants not admitted to
a neonatal intensive care unit. Variation in the programs consisted of two- or three-stage
screening, testing at a child health clinic (CHC) or at home, identification of unilateral
and/or bilateral hearing impairment, and use of automated OAE or AABR tests (Figure
15).
240
Universal neonatal hearing screening
Figure 15 Decision tree for modelled universal neonatal hearing screening protocols
(Boshuizen et al 2001)
There was no mention of the coverage obtained by any of the programs. The percentage
of infants referred for diagnostic testing for the six programs was determined to range
from 1.8 ± 0.6 per cent for three-stage OAE to 3.6 ± 1.0 per cent for two-stage OAE.
The perspective was that of the medical system in the Netherlands. Costs included
equipment depreciation, personnel, consumables, travel costs (of screeners), training,
administration and follow-up diagnostic testing (Table C). The study reported the costs
per child screened and per child detected with PCHI. The most cost-effective program
was found to be a two-stage screen with OAE performed at a CHC, with babies who
tested positive for bilateral hearing impairment being referred for diagnostic assessment
(Table D). However, if the criteria for a diagnostic referral included infants testing
positive for either unilateral or bilateral hearing impairment, then a three-stage screen
using OAE at the CHC was found to be the most cost-effective program. These
conclusions were mirrored in analysis of the cost per child detected with a hearing loss of
≥40 dB in the better ear.
Summary of modelled universal neonatal hearing screening protocols in the short
term
In three modelled evaluations of universal neonatal hearing screening protocols, two- or
three-stage otoacoustic emissions (OAE) screening, along with two-stage OAE and
automated auditory brainstem response screening, were recognised as the most costeffective methods in the short term for identifying permanent childhood hearing impairment.
While costs per child screened are similar, costs per child identified are substantially different
in the two studies (Kezirian et al 2001; Boshuizen et al 2001). This could only be due to
differences in referral rates and/or costs assumed for diagnostic testing, neither of which can
be compared from the information published.
Universal neonatal hearing screening
241
Comparison of existing universal neonatal hearing screening programs in
the short term
Seven studies were included that performed an economic analysis on programs that
already existed within their health system (Driscoll et al 2000; Gorga et al 2001; Lemons
et al 2002; Maxon et al 1995; Stone et al 2000; Vohr et al 2001; Weirather et al 1997).
Again, all included studies took a health care sector perspective.
Table E
Resource items considered in studies comparing actual UNHS programs
(Vohr et al
2001)
(Gorga et
al 2001)
(Lemons
et al 2002)
(Weirather
et al 1997)
1-stage
AABR or
TEOAE,
2-stage
TEOAE–
AABR
2-stage
DPOAE
2-stage
TEOAE or
AABR
2-stage
TEOAE
OAE instrument
9
9
9
8
9
9
9
AABR instrument
9
9
9
8
n/a
9
n/a
Computer
8
8
9
8
8
9
8
A
A
A
A
A
A
(Driscoll et
al 2000)
1-stage
TEOAE
(Maxon et
al 1995)
(Stone et
al 2000)
2-stage
TEOAE
1-stage
DPOAE
Capital equipment
Recurrent items
OAE probes
9
ABR electrodes
9
OAE probe tips
9
Ear
couplers/muffins
9
Cables
8
Machine calibration
8
Other consumables
(eg paper)
9
9
Screener training
9
9
Overheads
9
8
9
9
Personnel
Nurse or other
trained screener
9
n/a
9
9
n/a
9
9
Audiologist
9
9
9
9
9
9
9
Administrative
support
9
8
9
9
n/a
9
9
Follow-up
9
9
9
8
9
9
9
Patient and family
8
8
8
8
8
8
8
Other sectors
8
8
8
8
8
8
8
A = assumed (indicated as included but not described); AABR = automated auditory brainstem response test; OAE = otoacoustic emissions
test; TEOAE = transient evoked otoacoustic emissions test; DPOAE = distortion product otoacoustic emissions test; 9 = included; 8 = not
included; n/a = not applicable.
In a high quality retrospective study, the cost-effectiveness of three universal neonatal
hearing screening (UNHS) programs was compared on a total of 12,081 sequentiallytested well newborns at five different American hospitals (Vohr et al 2001). Each
hospital followed one of three protocols: one-stage transient evoked otoacoustic
emissions (TEOAE); one-stage automated auditory brainstem response (AABR); or twostage TEOAE followed by AABR (Figure 16). In reality, the first stage of each of the
242
Universal neonatal hearing screening
protocols consisted of testing with the designated screening method prior to hospital
discharge, and if the infant failed the test in both ears then the same test was
administered again. For the one-stage protocols, infants failing the (second) test were
referred for diagnostic testing. In the two-stage protocol, infants failing TEOAE (twice)
were then tested with AABR post-discharge. Infants failing the AABR were then referred
on for diagnostic assessment. In all cases nurses performed the screenings, although an
audiologist interpreted the TEOAE results.
Figure 16 Screening protocols for existing 1- and 2-stage universal neonatal hearing
screening programs (Vohr et al 2001)
Costs considered in this cost-effectiveness analysis were the wages of the audiologist, any
administration and clerical support, and the time for training of the audiologist and the
nurse screeners. The costs of purchasing the TEOAE and AABR equipment along with
all screening supplies and overheads were also considered. Failure and referral rates were
determined using the retrospective cohort data, while the costing was performed on a
hypothetical population of 1,500 newborns. The authors assumed that for every 1,500
infants, two new pieces of equipment were required for first-stage testing and one new
piece of equipment was needed for second-stage testing. The cost of follow-up
diagnostic assessment was also included in each protocol. Outcomes of interest were the
cost per child tested and cost per child identified with bilateral permanent childhood
hearing impairment (PCHI).
Results of all three protocols are similar (Table F). While the TEOAE technology was
the cheapest to purchase, it had the highest referral rate to diagnostic assessment (6.5%),
resulting in the highest post-discharge costs. Conversely, one-stage AABR was the most
expensive technology to purchase but resulted in the lowest referral rate (3.2%). It must
be remembered that these one-stage protocols actually consist of multiple tests using the
same technology in the event of an initial failed screening test. This would result in lower
referral rates than might be expected if only one test was performed prior to discharge.
The two-stage program yielded an intermediate referral rate (4.7%), but the overall higher
cost per child tested and per child identified with bilateral PCHI is probably due to (1)
the need to purchase two different technologies, thereby increasing costs, and (2) the
acknowledgement that a significant proportion of newborns at one two-stage screening
hospital (9.9%) were bypassed for TEOAE and tested directly with AABR, resulting in
increased overall referral rates for this protocol.
Universal neonatal hearing screening
243
Table F
Study
Vohr, Oh et al
2001
Short-term costs and cost-effectiveness of existing UNHS programs
Population
12,081
sequentially
tested well
newborns
Quality
Ratio
13.5/16
Protocol / Results ($AUS)
2-stage TEOAE
and AABR
1-stage
TEOAE
1-stage
AABR
Cost per child
born
46
40
46
Scenario 1a
52
57
41
Scenario
2b
Cost per child
identified with
bilateral PCHI
Gorga et al
2001
Lemons et al
2002
Driscoll et al
2000
4,460 newborns
at local hospital
(97.5%
coverage)
1,530 newborns
screened with
TEOAE; 1,412
newborns
screened with
AABR
1,305 infants of
mean age 2
months (1.5–
2.5 months);
voluntary
enrolment
43
40
41
23,300
20,220
23,120
13.5/16
2-stage DPOAE
Cost per baby
screened
Year 1
Year 5 c
Without followup costs
32
19
With follow-up
costs
37
25
2-stage TEOAE
2-stage AABR
Without followup costs
44
46
With follow-up
costs
80
63
12/16
Cost/baby
screened
1-stage TEOAE
11/16
Cost per child entering program
12.99
Cost of screening protocol per child
6.91
Cost of diagnostic assessment protocol per
child
6.08
Cost per child identified with sensorineural /
mixed hearing impairment
2,333
Cost per child identified with sensorineural /
mixed / conductive hearing impairment
318
Total annual program cost
Maxon et al
1995
Stone et al
2000
Weirather et al
1997
4,253 infants
born during 6month period in
1993
595 special
care; 3,658 well
11/16
1,002 well
newborns 6 to
72 hours old at
screen
11/16
380 well and
NICU newborns
11.5/16
84,019
2-stage TEOAE
Cost per child screened
40
Cost per child identified
6780
1-stage DPOAE
Cost per child screened
Cost per child diagnosed with PCHI
29
32,950
2-stage TEOAE
Cost per baby screened
11
Use of non-dedicated screeners (ie no cost); b performed by dedicated screeners ($17.00/hr); c average over 5-year life-span of the DPOAE
equipment. AABR = automated auditory brainstem response test. TEOAE = transient evoked otoacoustic emissions test. DPOAE = distortion
product otoacoustic emissions test. NICU = neonatal intensive care unit.
a
244
Universal neonatal hearing screening
Sensitivity analysis was used to demonstrate the effect of changing the type of personnel
used in screening. This was based on the author’s observation that non-dedicated
screeners have higher referral rates than dedicated screeners. Costs of using nondedicated screeners were assumed unrealistically to be $0, pushing overall costs down,
but because the referral rates of non-dedicated screeners were higher, two-stage and onestage TEOAE screening became more expensive per baby screened (Table F). The
referral rate for one-stage AABR did not change (as non-dedicated screeners were
already in use in this program), and was found to be the most cost-effective option.
Using dedicated screeners ($17 per hour) lowered the referral rate for the two-stage and
one-stage AABR protocols. The TEOAE referral rate did not change as dedicated
screeners were already being used at the hospital site using this program. In this analysis
one-stage TEOAE became the most cost-effective option.
This study highlights a number of issues. The choice of program is dependent upon
those goals that are more important to the health care system contemplating a screening
program. Ultimately, each program identifies the same number of infants with PCHI
within the newborn population. If cost minimisation was a priority, TEOAE may be
selected as the method of choice. However, if the goal was to alleviate the burden on
follow-up screening and reduce unnecessary parental anxiety due to false positive results,
then a two-stage or one-stage AABR program may be preferable.
A study of the cost-effectiveness of a single program (2-stage distortion product
otoacoustic emissions (DPOAE)) on 4,460 newborns from both well and neonatal
intensive care unit (NICU) populations was evaluated by Gorga and colleagues (2001)
(Figure 17). Pre-discharge babies were screened with DPOAE and rescreened with the
same technology if they failed the test in either ear. Following a second fail result,
diagnostic auditory brainstem response (ABR) was performed. A clinical audiologist
performed all tests.
Figure 17 Protocol for universal screening of well and neonatal intensive care unit
babies (Gorga et al 2001)
Costs considered in the evaluation were those of equipment, disposables, salary and
benefits of the audiologists as well as the cost of follow-up diagnostic assessment. Costs
were estimated over the period of 5 years (the estimated lifetime of the screening
equipment), assuming a birth cohort of 2,200 babies per year and a referral rate of 2 per
cent (as determined from results of the local program). Analysis attempted to determine
the cost per baby screened, but did not amortise the purchase cost of the equipment over
its useful clinical life. First-year costs per baby, with and without follow-up costs
included, were $32 and $37 respectively (Table F). However, since the cost of the
purchase of screening equipment was only included in this first year, over the lifetime (5
years) of the equipment the average cost per baby screened, with and without follow-up
costs, fell to $19 and $25 respectively.
Universal neonatal hearing screening
245
Cost-effectiveness of start-up of either a two-stage TEOAE or a two-stage AABR
screening program at two separate hospitals was compared in another recent study
(Lemons et al 2002). All newborn infants were targeted for screening at both hospitals,
resulting in 99 per cent coverage using TEOAE (1,530 newborns) and 96 per cent
coverage using AABR (1,412 newborns). A fail in either ear resulted in a second test
using the same technology. If another fail resulted, then the infant was referred for
audiologic assessment (Figure 18).
Figure 18 Protocol for 2-stage TEOAE or AABR universal neonatal hearing screening
(Lemons et al 2002)
Audiologists always performed TEOAE screening, while nursing staff performed AABR
screening. Resources included in the costing were personnel (audiologist and nurse time,
and administration and clerical assistance), screening materials (probes, tips, electrodes),
screening equipment (TEOAE, AABR, cart, printer, computer) and paper supplies.
Eighty-one per cent of infants passed the initial TEOAE screen, while 77 per cent passed
the initial AABR test. However, rates for follow-up assessment were reported to be 15
per cent for TEOAE and averaged 7 per cent for AABR during the study period. The
cost per infant screened, with and without follow-up, is presented in Table F. Due to
higher numbers of infants requiring follow-up assessment in the TEOAE program, the
costs per child screened were higher than for AABR despite the initial higher cost of
AABR equipment (total equipment start-up costs estimated to be $25,000 for AABR and
$13,000 for TEOAE). Supplies such as electrodes and probes were also determined to be
more costly for AABR than for TEOAE ($20,600 for AABR and $5,100 for TEOAE)
over the duration of the study. Conversely, the cost of personnel was more expensive for
the TEOAE program due to the requirement of using audiologists to interpret the
results, as opposed to using nurses for the AABR program where no such interpretation
was required ($19,600 for AABR and $50,000 for TEOAE).
In a study of 386 well and NICU infants born over a 2-month period, the costeffectiveness of a hospital universal neonatal screening program using two-stage TEOAE
was evaluated (Weirather et al 1997). First-stage screening was performed prior to
hospital discharge. A fail on the screening test meant the infant was scheduled to return
to hospital for a second TEOAE at 1–3 weeks of age. The infant was referred to the
hospital audiology department for diagnostic assessment if they failed the second
TEOAE test (Figure 19).
246
Universal neonatal hearing screening
Figure 19 Protocol for 2-stage TEOAE universal neonatal hearing screening (Weirather
et al 1997)
With this program the authors reported 98 per cent coverage (380/386) prior to
discharge, and 100 per cent coverage by 1 week of age. Eighty-nine per cent passed the
first-stage screen and 11 per cent returned for a second screening test. No information
was provided on the number of infants referred for diagnostic assessment. Costs
considered in the analysis included: personnel for screening, rescreening, screening
management, program management, patient management and scoring; fringe benefits
(30% of salaries); supplies; equipment (not screening equipment, which was assumed
already in place); and overhead costs (20% of costs). This paper argued that the resulting
cost per baby screened of only $11 dispelled any claim that the cost of a UNHS program
could be prohibitive (Table F). However, the study did not include in its analysis the
costs of purchasing screening equipment such as TEOAE devices. Further, large costs
incurred for diagnostic assessment of infants who had been referred were not taken into
account for costing of the program, nor were the referral rates reported in the study.
In an Australian study the cost-effectiveness of a one-stage TEOAE screening program
operated within community child health clinics (CHC) was evaluated on a voluntary
enrolment of 1,305 infants (Driscoll et al 2000). The mean age at time of testing was 2
months (range 1.5–2.5 months). The test was performed by an audiologist, and if the
infant failed the test in either ear they were referred to an auditory clinic for diagnostic
assessment. The referral rate for diagnostic evaluation was determined to be 11 per cent.
Costs considered in the evaluation were the audiologist’s salary, screening equipment and
maintenance, replacement cost of TEOAE probes on a yearly basis, consumables and
follow-up costs. Costs were averaged over an assumed lifetime of the TEOAE apparatus
of 6 years. The cost per child screened (entering program) was determined to be $13 and
the cost per child identified with sensorineural or mixed hearing loss was $2,333 (Table
F). However, due to the voluntary nature of the enrolment, it is possible that parents
who were concerned about the hearing of their infant were more likely to attend a CHC.
This may have increased the level of risk in the sample and reduced the cost per child
identified. Further, identifying infants with either sensorineural or mixed hearing
impairment may also have increased the prevalence of affected infants in this sample,
lowering the cost per infant identified.
A two-stage TEOAE program in Rhode Island (USA) was evaluated for the costeffectiveness of screening 4,253 infants (595 special care and 3,658 well infants) born
during a 6-month period in 1993 (Maxon et al 1995). First-stage screening was
performed prior to hospital discharge. If a failure was recorded the infant was scheduled
for a rescreen 4–6 weeks later. If a second failure was recorded, ABR was performed by
the hospital screening staff, and infants who failed the ABR test were referred for
diagnostic audiologic evaluation (Figure 20).
Universal neonatal hearing screening
247
Figure 20 Protocol for 2-stage TEOAE universal neonatal hearing screening (Maxon et
al 1995)
Analysis of effectiveness was based on a 93 per cent pass rate at the first-stage screen and
a referral rate of 1 per cent. The final yield of unilateral and bilateral permanent
childhood hearing impairment (PCHI) for this population of infants was determined to
be 6.0/1,000. Annual costs considered for the program were: personnel (including
trained screening technicians, audiologists, a coordinator, clerical assistance), fringe
benefits (28% of salaries), overhead costs (29% of salaries), supplies, cost of diagnostic
ABR evaluation, and equipment (three TEOAE screening tools, one ABR machine, four
computers and two printers – all amortised over 5 years). In this study two-stage
TEOAE was determined to cost $40 per child screened and $6,780 per child identified
with unilateral or bilateral PCHI (Table F).
Cost-effectiveness of a one-stage TEOAE UNHS program was performed on 1,002 well
newborns 6–72 hours old (Stone et al 2000). An infant who failed the first-stage TEOAE
test was referred for diagnostic assessment at 8 weeks of age using DPOAE and
tympanometry. A fail on both tympanometry and DPOAE indicated middle ear disease
and treatment. A fail on the DPOAE test alone resulted in further testing with AABR.
Costs included in the analysis were for a full-time registered nurse, administrator time for
1 hour/week, interpretation of results by an audiologist, DPOAE screening equipment,
supplies and follow-up assessment. Effectiveness of the program was based on a failure
rate of 11 per cent of infants tested with DPOAE in the first-stage screen (hence referral
rate to diagnostic assessment and a yield of 0.2 %). Outcomes of interest were the cost
per child screened and the cost per child diagnosed with PCHI. It is unclear whether the
protocol was interested in identifying unilateral and bilateral PCHI or only bilateral
PCHI. The cost per child screened was slightly lower than that reported by Vohr, Oh
and colleagues (2001) in a one-stage TEOAE ($29 vs $40 respectively) but the resulting
cost per child identified was found to be roughly 60 per cent higher than that of Vohr,
Oh and colleagues (2001) (Table F). Again, this is probably due to higher referral rates in
this study compared to the previous one (11% versus 6%).
248
Universal neonatal hearing screening
Summary of existing universal neonatal hearing screening protocols in the short term
From studies of existing universal neonatal hearing screening protocols, the cost per child
screened, including follow-up, was quite variable. Two-stage protocols ranged from $25 to
$80 per child screened while one-stage protocols ranged from $29 to $46 per child screened.
Estimates of the cost per child identified were similarly varied, ranging from $6,700 to $23,000
for two-stage screening programs, and from $20,000 to $32,000 for one-stage protocols.
In the only study comparing two-stage transient evoked otoacoustic emissions (TEOAE) –
automated auditory brainstem response (AABR) and one-stage TEOAE or one-stage AABR
hearing screening, the cost per child screened and cost per child identified with permanent
childhood hearing impairment (PCHI) were roughly equivalent (Vohr et al 2001). In a single
study comparing two-stage otoacoustic emissions testing with two-stage AABR, the latter was
considered more cost-effective because referral rates were lower, this despite the initial
higher cost outlay for AABR equipment (Lemons et al 2002).
Two included studies presented cost per child screened estimates that were substantially
lower than for comparable studies examined in this evaluation. In the one-stage TEOAE study
by Driscoll (Driscoll et al 2000) the screened sample was drawn from volunteer walk-in
subjects to a child health clinic (CHC). Cost per child screened here was noticeably lower
($13), which was probably due to the higher prevalence of hearing impairment in this
population, since concerned parents are more likely to attend the CHC with their child than
parents of seemingly healthy infants. Similarly, the cost of using a two-stage TEOAE protocol
for screening was determined to be only $11 per child screened before the costs of diagnostic
assessment (Weirather et al 1997). In this study, however, the initial purchase cost of
screening equipment was not included.
On the basis of this evidence, it is difficult to recommend any one protocol for UNHS in the
short term.
Universal neonatal hearing screening
249
Overall summary of published economic evaluation studies of neonatal
hearing screening
With the exception of two studies, all identified published research examining the costeffectiveness of existing or modelled universal neonatal hearing screening (UNHS)
programs are from an American perspective. In the only study from Australia, the study
design did not reflect the optimal conditions for a universal screening program, as the
participation was voluntary. Therefore, due to differences in costs and the structure of
the US medical system, the results obtained from the majority of the literature can be
only suggestive of what might occur under Australian conditions.
Short-term costs and benefits
Based on existing evidence from Keren et al (2002) and Kemper and Downs (2000), it is
suggested that UNHS is less cost-effective in the short term (incremental cost per child
identified with permanent childhood hearing impairment (PCHI)) than targeted
screening (Table B). The difference in the incremental cost-effectiveness ratio between
the two studies (Table G) may be due in part to costs incurred by the larger number of
infants referred from UNHS compared to targeted screening in the Keren et al study
(UNHS/targeted = 1.6%/0.18%, or 8-fold higher), compared to the Kemper and
Downs study (UNHS/targeted = 0.4%/0.07%, or 6-fold higher).
Table G
Short-term incremental cost-effectiveness ratio, cost per case identified, in modelled studies
comparing UNHS and targeted screening
Protocol
ICER (universal/targeted)
$AUS
Study
UNHS
Targeted
TEOAE/AABR
AABR/AABR
61,600a
(Keren et al 2002)
TEOAE/AABR
TEOAE/AABR
35,650b
(Kemper & Downs 2000)
ICER = incremental cost-effectiveness ratio. a all relevant costs included with the exception of a computer, consumables, screener training and
overheads; b assumed capital, consumable and personnel costs included but not described
Amongst studies of both modelled and existing UNHS protocols, the cost per child
screened including follow-up was quite variable (Table H). Two-stage protocols ranged
from $18 to $80 per child screened, while one-stage protocols ranged from $29 to $46
per child screened. Modelled programs were consistently less expensive per child
screened than similar protocols for existing programs (eg 2-stage transient evoked
otoacoustic emissions (TEOAE) – automated auditory brainstem response (AABR)
model vs existing program, $24 vs $46 per child screened). This suggests that some costs
incurred by the existing programs have not been accounted for in the modelled
programs. Variations in cost per child screened appear to depend on rates of referral to
diagnostic assessment. An exception to this would be the study by Driscoll and
colleagues (2000), in which the referral rate is high.
Within studies comparing different protocols (Table H – studies designated by G, V, K,
B, and L), modelled programs have shown that two-stage protocols are less costly per
child screened than one-stage protocols, probably due to the lower referral rate of twostage protocols (Table H – studies G and K). Two-stage protocols incorporating
otoacoustic emissions (OAE) testing are less costly than two-stage AABR protocols due
to the increased cost of consumables in AABR testing for the first stage of screening
(Table H – studies K and B). However, existing UNHS programs tell a somewhat
different story, with one study finding no difference in cost per child screened between
250
Universal neonatal hearing screening
two-stage TEOAE with AABR and one-stage TEOAE or one-stage AABR (Table H –
study V). Another existing program found the opposite result of that of a modelled study
(Table H – study K), with a two-stage AABR being less costly than a two-stage TEOAE
(Table H – study L).
Table H
Technique
Short-term cost per child screened for reported universal neonatal hearing screening
programs (Lemons et al 2002)
Type of
assessment
Referral
rate (%)
2003 cost/child screened ($AUS)
Without followup costs
Study
With followup costs
2-stage
TEOAE–AABR
Modelc
2
Existingc
4.7
46
V
(Vohr et al 2001)
Modelc
?
28
K
(Kezirian et al 2001)
Modelc
2–4
41
B
(Boshuizen et al 2001)
Existingb
6.5
46
63
L
(Lemons et al 2002)
DPOAE
Existingc
2
19
25
TEOAE
Existingb
15
44
80
Existingc
1
Existingd
?
Modelc
?
18
K
(Kezirian et al 2001)
Modelc
2–4
29
B
(Boshuizen et al 2001)
Existingc
6.5
40
V
(Vohr et al 2001)
Existingc
10.8
13
(Driscoll et al 2000)
DPOAE
Existingc
0.2
29
(Stone et al 2000)
AABR
Modelc
AABR
OAE
24
G
(Gorga et al 2001)
(Gorga et al 2001)
L
40
(Lemons et al 2002)
(Maxon et al 1995)
11a
(Weirather et al 1997)
1-stage
TEOAE
OAE
?
35
K
(Kezirian et al 2001)
Modelc
2
32
A
(Gorga et al 2001)
Existingc
3.2
46
V
(Vohr et al 2001)
Modelc
2
33
A
(Gorga et al 2001)
AABR = automated auditory brainstem response test; TEOAE = transient evoked otoacoustic emissions test; DPOAE = distortion product
otoacoustic emissions test; OAE = otoacoustic emissions test; a Costs do not include purchase of screening equipment; b most resources
included in costing; c assumed some or most resources included in costing but not described; d important resources absent from costing.
The outcome of ‘cost per child identified’ showed similar variation amongst studies,
ranging from $6,700 to $23,000 for two-stage screening programs and from $20,000 to
$32,000 for one-stage protocols (Table I). Modelled UNHS programs once more
identified a lower cost (per child identified) than existing programs using similar
screening protocols. This discrepancy is again probably due to the omission of cost items
in the modelled program that were included in the existing programs. Within studies
comparing different protocols (studies designated by V, K and B in Table I), modelled
programs have shown that two-stage OAE protocols are less costly per child identified
than one- or two-stage AABR protocols (Table I – studies K and B). However, the one
existing program comparing cost per child identified for different UNHS protocols
found no difference in cost per child screened between two-stage TEOAE with AABR
and one-stage TEOAE or one-stage AABR (Table I – study V).
Universal neonatal hearing screening
251
Table I
Technique
Short-term cost per child identified in reported universal neonatal hearing screening
programs
Type of assessment
Cost/child identified ($AUS)
Study
Modela
17,350
Existinga
23,300
TEOAE
Existinga
6,780
OAE
Modela
7200
K
(Kezirian et al 2001)
Modela
40,880
B
(Boshuizen et al 2001)
Modela
11,430
K
(Kezirian et al 2001)
Modela
57,380
B
(Boshuizen et al 2001)
Existinga
20,220
V
Existinga
2,333
(Driscoll et al 2000)
DPOAE
Existinga
32,950
(Stone et al 2000)
AABR
Modela
13,350
K
(Kezirian et al 2001)
Existinga
23,120
V
(Vohr et al 2001)
2-stage
TEOAE–AABR
AABR
(Kemper & Downs 2000)
V
(Vohr et al 2001)
(Maxon et al 1995)
1-stage
TEOAE
(Vohr et al 2001)
AABR = automated auditory brainstem response test; TEOAE = transient evoked otoacoustic emissions test; DPOAE = distortion product
otoacoustic emissions test; OAE = otoacoustic emissions test; a assumed some or most resources included in costing but not described.
Summary
From the literature it can be concluded that, in the short term, the costs for the additional
cases identified and diagnosed by universal neonatal hearing screening (UNHS) are greater
per unit than those of targeted screening. However, it is inappropriate to determine
incremental cost-effectiveness without considering long-term costs and cost savings. Further,
modelled two-stage UNHS protocols appear to be more cost-effective than modelled onestage protocols, but this observation is not supported by reports of existing UNHS programs
using either one- or two-stage screening protocols. Therefore, based on the available
evidence, no determination of the most cost-effective protocol for UNHS in the short term can
be made.
252
Universal neonatal hearing screening
Long-term costs and benefits
Only one study attempted to model the cost-effectiveness of universal neonatal hearing
screening (UNHS) over the long term and from a societal perspective (Keren et al 2002).
A hypothetical cohort of 80,000 newborns from one US state was used and consisted of
10,400 high-risk and 69,600 low-risk infants. Under targeted screening, only infants with
identified risk factors for congenital deafness were screened (Box 10). The decision tree
for this model is shown in Figure 11. Based on literature estimates of 1.6 cases of
permanent childhood hearing impairment (PCHI) per 1,000 births, the cohort was
assumed to contain 128 infants with PCHI. The point estimates presented were based on
80 per cent coverage of high-risk infants through targeted screening and 100 per cent
coverage for all infants using UNHS. Yield from the two screening programs was
estimated at 48 per cent (61 infants) for targeted screening and 91 per cent (116 infants)
for UNHS.
Lifetime societal costs of deafness that were considered for this model were lost
productivity, special education, vocational rehabilitation, additional medical costs and
assistive devices. Analysis resulted in an incremental saving of $2 million per child with
normal language using targeted screening compared to having no screening program, and
a further saving of $1.2 million per child using UNHS.
Table J
Study
Keren et al
2002
Societal benefit of hearing screening protocols
Population
Hypothetical
cohort of
80,000 infants
from one US
state
10,400 highrisk, 69,600
low-risk
infants
Quality
Ratio
Protocol /Results ($AUS)
No
screening
15/16
Incremental cost
reduction per child with
bilateral hearing
impairment possessing
normal language
0
Targeted
screening
Universal
screening
2-stage AABR
2-stage OAE and
AABR
2,030,000a
1,212,000b
a ICER for targeted vs no screening; b ICER for UNHS vs targeted (assuming 80% of high-risk patients screened in targeted screening scenario
and 77% of infants followed up with diagnostic assessment after positive screen in both scenarios). AABR = automated auditory brainstem
response test; OAE = otoacoustic emissions test; TEOAE = transient evoked otoacoustic emissions test.
Summary
While it was established earlier in this report that universal neonatal hearing screening would
be more expensive to operate than a targeted screening program, taking a societal
perspective over the long term suggests that identifying a larger proportion of hearingimpaired infants at an early stage (ie ≤6 months of age) would result in an overall costeffective program. However, these societal costs are based primarily on observation and
expert opinion and have not been generated from properly designed studies.
Universal neonatal hearing screening
253
Glossary
Absolute SpPin
A diagnostic finding where the Specificity is so high that a
Positive result rules-in the diagnosis.
Absolute SnNout
A diagnostic finding where the Sensitivity is so high that a
Negative result rules-out the diagnosis.
Absolute yield or yield
The number of cases of permanent childhood hearing
impairment ultimately identified in the screened population.
Aminoglycosides
A group of antibiotics used to treat certain bacterial
infections.
Confounding
A variable that is related to or causative of the outcome of
interest and is differentially distributed between the groups
with and without the intervention/condition.
Cytomegalovirus
Any of a group of herpes viruses that induces birth defects
and affects humans with impaired immunological systems.
External validity
The degree to which the results of a study can be applied to
the wider population.
Failure rate
The number of infants who failed a screen or rescreen
divided by the number of individuals who were intended to
be screened.
False alarm rate
Complement of the positive predictive value of a test. In
other words, the number of infants falsely identified with
hearing impairment divided by the total number of infants
testing positive for hearing impairment.
False negative rate
Complement of the sensitivity of a test. In other words, the
number of infants incorrectly identified with normal hearing
(false negative) divided by the total number of infants with
hearing impairment.
False positive rate
Complement of the specificity of a test. In other words, the
number of infants falsely identified with hearing impairment
(false positive) divided by the total number of infants without
hearing impairment.
False reassurance rate
Complement of the negative predictive value of a test. In
other words, the number of infants incorrectly identified as
having normal hearing divided by the total number of infants
testing negative for hearing impairment.
Hyperbilirubinaemia
In the newborn, a disorder usually caused by immaturity of
the liver that usually subsides spontaneously.
Incremental yield
The number of additional cases of permanent childhood
hearing impairment ultimately identified in the screened
population, compared to usual targeted screening or casefinding methods.
254
Universal neonatal hearing screening
Positive predictive value
The number of infants correctly identified with hearing
impairment divided by the total number of infants testing
positive for hearing impairment.
Power
The ability of a study to demonstrate an association, given
that the association exists.
Sensitivity
The number of infants correctly identified by a test as having
hearing impairment divided by the total number of infants
with hearing impairment.
Specificity
The number of infants correctly identified by a test as having
no hearing impairment divided by the total number of infants
without hearing impairment.
Toxoplasmosis
A disease caused by infection with the parasite Toxoplasma
gondii. In newborns, congenital toxoplasmosis is an infection
that results from the transplacental passage of parasites from
the infected mother to the fetus (Braunwald et al 2001)
True positive
The number of cases correctly identified by a test as having
the condition.
Universal neonatal hearing screening
255
Abbreviations
AABR
automated auditory brainstem response (test)
ABR
auditory brainstem response (test)
AHMAC
Australian Health Ministers’ Advisory Council
AHMC
Australian Health Ministers’ Conference
ANCOVA
analysis of covariance
ARTG
Australian Register of Therapeutic Goods
AR-DRG
Australian Refined Diagnosis Related Groups
CDI
Communicative Development Inventory
CDR
clinical decision rule
CEA
cost-effectiveness analysis
CHC
child health clinic
CHIP
Colorado Home Intervention Program
CI
confidence interval
Clin I
clinical importance
CPI
consumer price index
CPG
clinical prediction guide
CQ
cognitive quotient
dB
decibel
DECS
South Australian Department of Education & Children’s
Services
DPOAE
distortion product otoacoustic emissions (test)
DQ
development quotient
dx
diagnostic
ECA
economic cost analysis
ENT
ear, nose and throat
FA
false alarm
FN
false negative
FP
false positive
HL
hearing level
HVDT
health visitor distraction test
ICER
incremental cost-effectiveness ratio
kHz
kilohertz
LQ
language quotient
LR
likelihood ratio
256
Universal neonatal hearing screening
LTFU
loss/lost to follow up
MBS
Medicare Benefits Schedule
MCDI
Minnesota Child Development Inventory
MeSH
medical subject heading
MSAC
Medical Services Advisory Committee
n
number
n/a
not applicable
nHL
near hearing level
NHMRC
National Health and Medical Research Council
NICU
neonatal intensive care unit
NNTB
number needed to treat to benefit
NS
not significant
OAE
otoacoustic emissions (test)
OECD
Organisation for Economic Cooperation and Development
OR
odds ratio
PAQ
play assessment questionnaire
PCHI
permanent childhood hearing impairment
PPV
positive predictive value
QS
quality score
R
relevance
RCT
randomised controlled trial
RR
relative risk / rate ratios
SCBU
special care baby unit
SD
standard deviation
SNHL
sensorineural hearing loss
SP
statistical precision
SPL
sound pressure level
SR
systematic review
SSEP
steady state evoked potentials (test)
SWP
sweeps
TEOAE
transient evoked otoacoustic emissions (test)
TN
true negative
TP
true positive
UMDNS
Universal Medical Device Nomenclature System
UNHS
universal neonatal hearing screening
WBN
well baby nursery
Universal neonatal hearing screening
257
References
ABS (2003). Labour force Australia [Internet]. Australian Bureau of Statistics. Available
from: www.abs.gov.au/Ausstats/aba106 [Accessed 12th November].
ABS (2004). 6401.0 Consumer Price Index [Internet]. Australian Bureau of Statistics.
Available from: http://www.abs.gov.au/Ausstats/abs%40.nsf/e8ae5488b598839cca
[Accessed 28/09].
ABS (2005). 3301.0 Births, Australian Bureau of Statistics, Australia.
Aidan, D., Avan, P. & Bonfils, P. (1999). 'Auditory screening in neonates by means of
transient evoked otoacoustic emissions: a report of 2,842 recordings', Ann Otol Rhinol
Laryngol, 108 (6), 525-531.
Armitage, P., Berry, G. & Matthews, J. N. S. (2002). Statistical methods in medical research.
Fourth Edition Blackwell Science, Oxford.
Australian Hearing (2001). Universal neonatal hearing screening (UNHS)- Implications for
Australian Hearing, Australian Hearing.
Australian Hearing (2003). Causes of hearing loss [Internet]. Australian Hearing. Available
from: www.hearing.com.au [Accessed July 24 2003].
Australian Hearing (2007). 'Demographic details of persons under the age of 21 years
with a hearing impairment who have been fitted with a hearing aid - 2007'. Audiology
Circular, 2007/6.
Australian Institute of Health and Welfare (2002). The Knowledgebase - national health data
dictionary [Internet]. version 11. Available from:
http://www.aihw.gov.au/knowledgebase/indexkbhealth.html [Accessed Accessed April
2 2003].
Bailey, H. D., Bower, C. et al (2002). 'Newborn hearing screening in Western Australia',
Medical Journal of Australia, 177 (4), 180-185.
Bamford, J., Uus, K. & Davis, A. (2005). 'Screening for hearing loss in childhood: Issues,
evidence and current approaches in the UK', Journal of Medical Screening, 12 (3), 119-124.
Bantock, H. M. & Croxson, S. (1998). 'Universal hearing screening using transient
otoacoustic emissions in a community health clinic', Archives of Disease in Childhood, 78 (3),
249-252.
Barsky Firkser, L. & Sun, S. (1997). 'Universal newborn hearing screenings: a three-year
experience', Pediatrics, 99 (6), E4.
Boshuizen, H. C., van der Lem, G. J. et al (2001). 'Costs of different strategies for
neonatal hearing screening: a modelling approach', Archives of disease in childhood. Fetal and
neonatal edition, 85 (3), F177-181.
258
Universal neonatal hearing screening
Braunwald, E., Fauci, A. S. et al (2001). Harrison's (15th Edition) principles of internal medicine,
Vol. 1. 15 McGraw-Hill Medical Publishing Division, New York.
Brennan, R. A. (2004). 'A nurse-managed universal newborn hearing screen program',
MCN: The American Journal of Maternal Child Nursing, 29 (5), 320.
Chapchap, M. J. & Segre, C. M. (2001). 'Universal newborn hearing screening and
transient evoked otoacoustic emission: new concepts in Brazil. Proceedings from the 3rd
International Congress of Paediatric Audiology, Copenhagen, Denmark, May 9-12, 2000',
Scandinavian Audiology Supplementum, 30, 33-36.
Child and Youth Health (2001). Universal neonatal hearing screening in Australia: a national
forum for consensus and implementation [Internet]. Child and Youth Health. Available from:
http://www.cyh.com/cyh/contents/cms_content.html [Accessed 13 November 2002].
Chiong, C., Ostrea Jr, E. et al (2007). 'Correlation of hearing screening with
developmental outcomes in infants over a 2-year period', Acta Oto-Laryngologica, 127 (4),
384-388.
Clarkson, R. L., Vohr, B. R. et al (1994). 'Universal infant hearing screening and
intervention: the Rhode Island program', Infants and Young Children, 6 (3), 65-74.
Clemens, C. J. & Davis, S. A. (2001). 'Minimizing false-positives in universal newborn
hearing screening: a simple solution', Pediatrics, 107 (3), 1 - 3.
Clemens, C. J., Davis, S. A. & Bailey, A. R. (2000). 'The false-positive in universal
newborn hearing screening', Pediatrics, 106 (1), E7.
Columbia University College of Physicians and Surgeons (2002). Hearing loss [Internet].
Columbia Presbyterian Medical Center. Available from:
http://www.entcolombia.org/hearloss.htm [Accessed 14/05/2003].
Connolly, J. L., Carron, J. D. & Roark, S. D. (2005). 'Universal newborn hearing
screening: are we achieving the Joint Committee on Infant Hearing (JCIH) objectives?'
Laryngoscope, 115 (2), 232-236.
Cox, L. C. & Toro, M. R. (2001). 'Evolution of a universal infant hearing screening
program in an inner city hospital', International Journal of Pediatric Otorhinolaryngology, 59 (2),
99-104.
Crockett, R., Baker, H. et al (2005). 'Maternal anxiety and satisfaction following infant
hearing screening: A comparison of the health visitor distraction test and newborn
hearing screening', Journal of Medical Screening, 12 (2), 78-82.
Crockett, R., Wright, A. J. et al (2006). 'Maternal anxiety following newborn hearing
screening: The moderating role of knowledge', Journal of Medical Screening, 13 (1), 20-25.
Daemers, K., Drickx, J. D. et al (1996). 'Neonatal hearing screening with otoacoustic
emissions: an evaluation', Acta Oto-Rhino-Laryngologica Belgica, 50 (3), 203-209.
Universal neonatal hearing screening
259
Davies, S. C., Cronin, E. et al (2000). 'Screening for sickle cell disease and thalassaemia: a
systematic review with supplementary research.' Health Technology Assessment, 4 (3), i-v, 1 99.
Davis, A., Bamford, J. et al (1997). 'A critical review of the role of neonatal hearing
screening in the detection of congenital hearing impairment', Health Technology Assessment,
1 (10), i-iv, 1-176.
Davis, A. & Wood, S. (1992). 'The epidemiology of childhood hearing impairment: factor
relevant to planning of services', British Journal of Audiology, 26 (2), 77-90.
De Capua, B., De Felice, C. et al (2003). 'Newborn hearing screening by transient evoked
otoacoustic emissions: analysis of response as a function of risk factors', Acta
Otorhinolaryngol Ital, 23 (1), 16-20.
Deeks, J. J. (2001). 'Systematic reviews of evaluations of diagnostic and screening tests',
British Medical Journal, 323 (21 July), 157-162. Department of Medical Oncology (1998).
On-Line Medical Dictionary [Internet]. University of Newcastle upon Tyne. Available from:
http://cancerweb.ncl.ac.uk/cgi-bin/omd?query=hearing+loss [Accessed 17/07/2003].
Downs, M. P. (1994). 'The case for detection and intervention at birth', Seminars in
Hearing, 15 (2), 76-84.
Downs, M. P. (1995). 'Universal newborn hearing screening--the Colorado story',
International Journal of Pediatric Otorhinolaryngology, 32 (3), 257-259.
Driscoll, C., Kei, J. & McPherson, B. (2000). 'Cost comparison of hearing screening
programs: Transient evoked otoacoustic emissions in infant screening and screening at
school entry', Australian and New Zealand Journal of Audiology, 22 (2), 77-84.
Drummond, M. F., O'Brien, B. J. et al (1997). Methods for the economic evaluation of health care
programmes. 2nd Edition, Oxford University Press, Oxford.
Egger, M., Juni, P. et al (2003). 'How important are comprehensive literature searches
and the assessment of trial quality in systematic reviews? Empirical study.' Health
Technology Assessment, 7 (1), 76.
Fortnum, H. & Davis, A. (1997). 'Epidemiology of permanent childhood hearing
impairment in Trent Region, 1985-1993', British Journal of Audiology, 31 (6), 409-446.
Gallaudet Research Institute (2001). Regional and national summary report of data from the
1999-2000 annual survey of deaf and hard of hearing children and youth, Gallaudet University,
Washington, D.C.
Gorga, M. P., Preissler, K. et al (2001). 'Some issues relevant to establishing a universal
newborn hearing screening program', Journal of the American Academy of Audiology, 12 (2),
101-112.
Govaerts, P. J., Yperman, M. et al (2001). 'A two-stage bipodal screening model for
universal neonatal hearing screening', Otology & Neurotology, 22 (6), 850-854.
260
Universal neonatal hearing screening
Habib, H. S. & Abdelgaffar, H. (2005). 'Neonatal hearing screening with transient evoked
otoacoustic emissions in Western Saudi Arabia', International Journal of Pediatric
Otorhinolaryngology, 69 (6), 839-842.
Hadjikakou, K. & Bamford, J. (2000). 'Prevalence and age of identification of permanent
childhood hearing impairment in Cyprus', Audiology, 39 (4), 198-201.
Hahn, M., Lamprecht-Dinnesen, A. et al (1999). 'Hearing screening in healthy newborns:
feasibility of different methods with regard to test time', International Journal of Pediatric
Otorhinolaryngology, 51 (2), 83-89.
Hatzopoulos, S., Qirjazi, B. & Martini, A. (2007). 'Neonatal hearing screening in Albania:
Results from an ongoing universal screening program', International Journal of Audiology, 46
(4), 176-182.
Helfand, M., Thompson, D. C. et al (2001). Newborn hearing screening, Agency for
Healthcare Research and Quality, Rockville.
Hergils, L. & Hergils, A. (2000). 'Universal neonatal hearing screening - Parental attitudes
and concern', British Journal of Audiology, 34 (6), 321-327.
Ho, V., Daly, K. A. et al (2002). 'Otoacoustic emissions and tympanometry screening
among 0-5 year olds', Laryngoscope, 112 (3), 513-519.
Hunter, M. F., Kimm, L. et al (1994a). 'Feasibility of otoacoustic emission detection
followed by ABR as a universal neonatal screening test for hearing impairment', British
Journal of Audiology, 28 (1), 47-51.
Hunter, M. F., Kimm, L. et al (1994b). 'Feasibility of otoacoustic emission detection
followed by ABR as a universal neonatal screening test for hearing impairment', British
Journal of Audiology, 28 (1), 47-51.
Huynh, M. T., Pollack, R. A. & Cunningham, R. A. J. (1996). 'Universal newborn hearing
screening: Feasibility in a community hospital', Journal of Family Practice, 42 (5), 487-490.
Iley, K. L. & Addis, R. J. (2000). 'Impact of technology choice on service provision for
universal newborn hearing screening within a busy district hospital', Journal of Perinatology,
20 (8 Part 2), S122-127.
Isaacson, G. (2000). 'Universal newborn hearing screening in an inner-city, managed care
environment', Laryngoscope, 110 (6), 881-894.
Iwasaki, S., Hayashi, Y. et al (2003). 'A model of two-stage newborn hearing screening
with automated auditory brainstem response', International Journal of Pediatric
Otorhinolaryngology, 67 (10), 1099-1104.
Jacobson, J. T. & Jacobson, C. A. (1994). 'The effects of noise in transient EOAE
newborn hearing screening', International Journal of Pediatric Otorhinolaryngology, 29 (3), 235248.
Universal neonatal hearing screening
261
Jakubikova, J., Kabatova, Z. & Zavodna, M. (2003). 'Identification of hearing loss in
newborns by transient otoacoustic emissions', International Journal of Pediatric
Otorhinolaryngology, 67 (1), 15-18.
Kanne, T. J., Schaefer, L. & Perkins, J. A. (1999). 'Potential pitfalls of initiating a
newborn hearing screening program', Archives of Otolaryngology -- Head and Neck Surgery, 125
(1), 28-32.
Kemp, D. T. (2002). 'Otoacoustic emissions, their origin in cochlear function, and use',
British Medical Bulletin, 63, 223-241.
Kemp, D. T. (2003). OAE's and the cochlea [Internet]. Otodynamics. Available from:
http://www.oae-ilo.co.uk/page_view.asp?id=59 [Accessed 02/05/03].
Kemper, A. R. & Downs, S. M. (2000). 'A cost-effectiveness analysis of newborn hearing
screening strategies', Archives of Pediatrics and Adolescent Medicine, 154 (5), 484-488.
Kennedy, C., McCann, D. et al (2005). 'Universal newborn screening for permanent
childhood hearing impairment: an 8-year follow-up of a controlled trial', Lancet, 366
(9486), 660-662.
Kennedy, C. R. (1999). 'Controlled trial of universal neonatal screening for early
identification of permanent childhood hearing impairment: coverage, positive predictive
value, effect on mothers and incremental yield. Wessex Universal Neonatal Screening
Trial Group', Acta Paediatrica, 88 (432 (Supplement)), 73-75.
Kennedy, C. R., Kimm, L. et al (1998). 'Controlled trial of universal neonatal screening
for early identification of permanent childhood hearing impairment', Lancet, 352 (9145),
1957-1964.
Kennedy, C. R., McCann, D. C. et al (2006). 'Language ability after early detection of
permanent childhood hearing impairment', N Engl J Med, 354 (20), 2131-2141.
Keren, R., Helfand, M. et al (2002). 'Projected cost-effectiveness of statewide universal
newborn hearing screening', Pediatrics, 110 (5), 855-864.
Kezirian, E. J., White, K. R. et al (2001). 'Cost and cost-effectiveness of universal
screening for hearing loss in newborns', Otolaryngology - Head and Neck Surgery, 124 (4),
359-367.
Khairi, M. D., Din, S. et al (2005). 'Hearing screening of infants in Neonatal Unit,
Hospital Universiti Sains Malaysia using transient evoked otoacoustic emissions', J
Laryngol Otol, 119 (9), 678-683.
Khan, K. S., Ter Riet, G. et al (2001). Undertaking systematic reviews of research on effectiveness.
CRD's guidance for those carrying out or commissioning reviews, NHS Centre for Reviews and
Dissemination, University of York, York.
Khandekar, R., Khabori, M. et al (2006). 'Neonatal screening for hearing impairment-The Oman experience', International Journal of Pediatric Otorhinolaryngology, 70 (4), 663-670.
262
Universal neonatal hearing screening
Kolski, C., Le Driant, B. et al (2007). 'Early hearing screening: What is the best strategy?'
International Journal of Pediatric Otorhinolaryngology, 71 (7), 1055-1060.
Kubba, H., MacAndie, C. et al (2004). 'Is deafness a disease of poverty? The association
between socio-economic deprivation and congenital hearing impairment', International
Journal of Audiology, 43 (3), 123-125.
Lemons, J., Fanaroff, A. et al (2002). 'Newborn hearing screening: costs of establishing a
program', Journal of Perinatology, 22 (2), 120-124.
Leveque, M., Schmidt, P. et al (2007). 'Universal newborn hearing screening: a 27-month
experience in the French region of Champagne-Ardenne', Acta Paediatrica.
Lin, C. Y., Huang, C. Y. et al (2004). 'Community-based newborn hearing screening
program in Taiwan', International Journal of Pediatric Otorhinolaryngology, 68 (2), 185-189.
Lin, H. C., Shu, M. T. et al (2005). 'Comparison of hearing screening programs between
one step with transient evoked otoacoustic emissions (TEOAE) and two steps with
TEOAE and automated auditory brainstem response', Laryngoscope, 115 (11), 1957-1962.
Low, W. K., Pang, K. Y. et al (2005). 'Universal newborn hearing screening in Singapore:
the need, implementation and challenges', Ann Acad Med Singapore, 34 (4), 301-306.
Martines, F., Porrello, M. et al (2007). 'Newborn hearing screening project using transient
evoked otoacoustic emissions: Western Sicily experience', International Journal of Pediatric
Otorhinolaryngology, 71 (1), 107-112.
Mason, J. A. & Herrmann, K. R. (1998). 'Universal infant hearing screening by automated
auditory brainstem response measurement', Pediatrics, 101 (2), 221-228.
Maxon, A. B., White, K. R. et al (1995). 'Referral rates and cost efficiency in a universal
newborn hearing screening program using transient evoked otoacoustic emissions',
Journal of the American Academy of Audiology, 6 (4), 271-277.
Maxon, A. B., White, K. R. et al (1993). 'Using transient evoked otoacoustic emissions
for neonatal hearing screening', British Journal of Audiology, 27 (2), 149-153.
McNellis, E. L. & Klein, A. J. (1997). 'Pass/fail rates for repeated click-evoked
otoacoustic emission and auditory brain stem response screenings in newborns',
Otolaryngology - Head and Neck Surgery, 116 (4), 431-437.
McPherson, B., Kei, J. et al (1998). 'Feasibility of community-based hearing screening
using transient evoked otoacoustic emissions', Public Health, 112 (3), 147-152.
Mehl, A. L. & Thomson, V. (2002). 'The Colorado newborn hearing screening project,
1992-1999: on the threshold of effective population-based universal newborn hearing
screening', Pediatrics, 109 (1).
Messner, A. H., Price, M. et al (2001). 'Volunteer-based universal newborn hearing
screening program', International Journal of Pediatric Otorhinolaryngology, 60 (2), 123-130.
Universal neonatal hearing screening
263
Mohr, P. E., Feldman, J. J. et al (2000). 'The societal costs of severe to profound hearing
loss in the United States', International Journal of Technology Assessment in Health Care, 16 (4),
1120-1135.
Molini, E., Ricci, G. et al (2004). 'Identifying congenital hearing impairment. Personal
experience based on selective hearing screening', Acta Otorhinolaryngol Ital, 24 (3), 109-116.
MSAC (2000). Funding for new medical technologies and procedures: application and assessment
guidelines, Commonwealth of Australia, Canberra, ACT.
MSAC (2005). Guidelines for the assessment of diagnostic technologies [Internet]. Medical Services
Advisory Committee, Commonwealth of Australia. Available from: www.msac.gov.au
[Accessed
Muir, J. A. (2001). Evidence-based healthcare. second edition Churchill Livingstone,
Edinburgh.
Mukari, S. Z., Tan, K. Y. & Abdullah, A. (2006). 'A pilot project on hospital-based
universal newborn hearing screening: Lessons learned', International Journal of Pediatric
Otorhinolaryngology, 70 (5), 843-851.
Murray, J., Cuckle, H. et al (1999). 'Screening for cystic fibrosis.' Health Technology
Assessment, 3 (8), i-iv, 1 - 104.
Mytton, J. & Mackenzie, I. (2005). 'Observed and expected prevalence of permanent
childhood hearing impairment in Oldham', Journal of public health (Oxford, England), 27 (3),
298-302.
NCHAM (2003). What does a newborn hearing screening program cost? [Internet]. National
Center for Hearing Assessment and Management (NCHAM). Available from:
www.infanthearing.org/resources/cost/index.html [Accessed 1/07/03].
Neary, W., Kent, S. et al (2003). 'The Role of Audiological Testing and Computed
Tomography in the Aetiological Investigation of Children with Permanent Unilateral
Hearing Loss', Audiological Medicine, 1 (4), 215-223.
Nekahm, D., Weichbold, V. et al (2001a). 'Improvement in early detection of congenital
hearing impairment due to universal newborn hearing screening', International Journal of
Pediatric Otorhinolaryngology, 59 (1), 23-28.
Nekahm, D., Weichbold, V. & Welzl Muller, K. (2001b). 'Epidemiology of permanent
childhood hearing impairment in the Tyrol, 1980-94', Scandinavian audiology, 30 (3), 197202.
Neumann, K., Gross, M. et al (2006). 'Effectiveness and efficiency of a universal
newborn hearing screening in Germany', Folia Phoniatr Logop, 58 (6), 440-455.
Ng, P. K., Hui, Y. et al (2004). 'Feasibility of implementing a universal neonatal hearing
screening programme using distortion product otoacoustic emission detection at a
university hospital in Hong Kong', Hong Kong Medical Journal, 10 (1), 6-13.
264
Universal neonatal hearing screening
NHMRC (2000a). How to review the evidence: systematic identification and review of the scientific
literature, National Health and Medical Research Council, Canberra.
NHMRC (2000b). How to use the evidence: assessment and application of scientific evidence,
National Health and Medical Research Council, Canberra.
NHS Centre for Reviews and Dissemination (2001). Undertaking systematic reviews of research
on effectiveness: CRD's guidance for those carrying out or commissioning reviews, CRD Report 4. 2nd
edition.
Nussbaum, D. (1999). About hearing loss [Internet]. series number 5001. Laurent Clerc
National Deaf Education Center, Galludet University. Available from:
http://clerccenter.galludet.edu/SupportServices/series/5001.html [Accessed 24 July
2003].
OECD (2003). Prices and Purchasing Power Parities [Internet]. Organisation for Economic
Co-operation and Development. Available from:
http://www.oecd.org/topicstatsportal/0,2647,en_2825_495691_2759297_1_1_1_1,00.ht
ml#500300 [Accessed 20/08/03].
Oudesluys-Murphy, A. M. & Harlaar, J. (1997). 'Neonatal hearing screening with an
automated auditory brainstem response screener in the infant's home', Acta Paediatrica, 86,
651 - 655.
Owen, M., Webb, M. & Evans, K. (2001). 'Community based universal neonatal hearing
screening by health visitors using otoacoustic emissions', Archives of disease in childhood.
Fetal and neonatal edition, 84 (3), F157-162.
Owens, R. (1996). Language development. Fourth edition Allyn and Bacon, New York.
Pastorino, G., Sergi, P. et al (2005). 'The Milan Project: A newborn hearing screening
programme', Acta Paediatrica, International Journal of Paediatrics, 94 (4), 458-463.
Prieve, B., Dalzell, L. et al (2000). 'The New York State universal newborn hearing
screening demonstration project: outpatient outcome measures', Ear and Hearing, 21 (2),
104-117.
Pugh, M. (2000). Stedman's medical dictionary. 27th edition, Lippincott Williams & Wilkins,
Philadelphia.
Rao, A., Anderson, G. & Zack, J. (2002). 'A collaborative early hearing detection and
intervention program for five rural hospitals', International Pediatrics, 17 (2), 102-106.
Robles, L. & Ruggero, M. A. (2001). 'Mechanics of the mammalian cochlea', Physiological
Reviews, 81 (3), 1305-1352.
Rouev, P., Mumdzhiev, H. et al (2004). 'Universal newborn hearing screening program in
Bulgaria', International Journal of Pediatric Otorhinolaryngology, 68 (6), 805-810.
Russ, S. A., Poulakis, Z. et al (2003). 'Epidemiology of congenital hearing loss in Victoria,
Australia', International Journal of Audiology, In Press.
Universal neonatal hearing screening
265
Russ, S. A., Rickards, F. et al (2002). 'Six year effectiveness of a population based two tier
infant hearing screening programme', Archives of Disease in Childhood, 86 (4), 245-250.
Schauseil-Zipf, U. & Von Wedel, H. (1988). 'Hearing screening by means of acoustic
evoked brain stem potentials in newborns and infants', Klinische Padiatrie, 200 (4), 324-329.
Schersten, T., Baile, M. A. et al (1999). Prostate cancer screening., The Swedish Council on
Technology Assessment in Health Care (SBU); Basque Office for Health Technology
Assessment (OSTEBA); Department of Health Basque Government., Vitoria-Gasteiz.
Scott, M. E. & Bhattacharyya, N. (2002). Auditory brainstem response audiometry [Internet]. emedicine. Available from: www.emedicine.com/ENT/topic473.htm [Accessed
28/04/03].
Shoup, A. G., Owen, K. E. et al (2005). 'The Parkland Memorial Hospital experience in
ensuring compliance with universal newborn hearing screening follow-up', Journal Of
Pediatrics, 146 (1), 66-72.
Smyth, V., McPherson, B. et al (1999). 'Otoacoustic emission criteria for neonatal hearing
screening', International Journal of Pediatric Otorhinolaryngology, 48 (1), 9-15.
Stata Corporation. (2001). Intercooled stata 7.0 for windows 98/95/NT [computer program],
Stata Corporation, College City, Texas.
Stewart, D. L., Mehl, A. et al (2000). 'Universal newborn hearing screening with
automated auditory brainstem response: a multisite investigation', Journal of Perinatology, 20
(8 Pt 2), S128-131.
Stone, K. A., Smith, B. D. et al (2000). 'Universal newborn hearing screening', Journal of
Family Practice, 49 (11), 1012-1016.
Swanepoel, D., Ebrahim, S. et al (2007). 'Newborn hearing screening in a South African
private health care hospital', International Journal of Pediatric Otorhinolaryngology, 71 (6), 881887.
Tatli, M. M., Bulent Serbetcioglu, M. et al (2007). 'Feasibility of neonatal hearing
screening program with two-stage transient otoacoustic emissions in Turkey', Pediatr Int,
49 (2), 161-166.
Thompson, D. C., McPhillips, H. et al (2001). 'Universal newborn hearing screening:
summary of evidence', Jama, 286 (16), 2000-2010.
Tsuchiya, H., Goto, K. et al (2006). 'Newborn hearing screening in a single private
Japanese obstetric hospital', Pediatr Int, 48 (6), 604-607.
UK National Screening Committee (2000). Second report of the UK national screening
committee., Health Departments of the United Kingdom, London.
Uus, K. & Bamford, J. (2006). 'Effectiveness of population-based newborn hearing
screening in England: Ages of interventions and profile of cases', Pediatrics, 117 (5), e887e893.
266
Universal neonatal hearing screening
Van Naarden, K., P., D. & Caldwell, K. (2000). 'Prevalence and characteristics of children
with serious hearing impairment in metropolitan Atlanta, 1991-1993', Pediatrics, 103 (3),
570-575.
Vohr, B. R., Carty, L. M. et al (1998). 'The Rhode Island Hearing Assessment Program:
experience with statewide hearing screening (1993-1996)', J Pediatr, 133 (3), 353-357.
Vohr, B. R., Oh, W. et al (2001). 'Comparison of costs and referral rates of 3 universal
newborn hearing screening protocols', J Pediatr, 139 (2), 238-244.
Wake, M., Hughes, E. K. et al (2003). 'Outcomes of children with mild-profound
congenital hearing loss at 7-8 years: a population study', Ear and Hearing, (In Press).
Wake, M., Poulakis, Z. et al (2005). 'Hearing impairment: a population study of age at
diagnosis, severity, and language outcomes at 7-8 years', Archives of Disease in Childhood, 90
(3), 238-244.
Walter, S. D. (2003). 'Mammographic screening: case-control studies', Ann Oncol, 14 (8),
1190-1192.
Watkin, P. M. & Baldwin, M. (1999). 'Confirmation of deafness in infancy', Archives of
Disease in Childhood, 81 (5), 380-389.
Watkin, P. M., Baldwin, M. et al (1998). 'Maternal anxiety and attitudes to universal
neonatal hearing screening', British Journal of Audiology, 32 (1), 27-37.
Watkin, P. M., Baldwin, M. & Laoide, S. (1990). 'Parental suspicion and identification of
hearing impairment', Archives of Disease in Childhood, 65 (8), 846-850.
Weichbold, V., Nekahm-Heis, D. & Welzl-Mueller, K. (2006). 'Ten-year outcome of
newborn hearing screening in Austria', International Journal of Pediatric Otorhinolaryngology, 70
(2), 235-240.
Weichbold, V. & Welzl Mueller, K. (2001). 'Maternal concern about positive test results
in universal newborn hearing screening', Pediatrics, 108 (5), 1111-1116.
Weinrich, J. E. (1972). 'Direct economic costs of deafness in the United States', American
Annals of the Deaf, 117 (4), 446-454.
Weirather, Y. P., Korth, N. et al (1997). 'Cost analysis of TEOAE-based universal
newborn hearing screening', Journal of Communication Disorders, 30 (6), 477-493.
White, K. R. (1997). 'Universal newborn hearing screening issues and evidence',
Conference Proceeding: Early Hearing Detection and Intervention, Atlanta, Georgia, In:
Universal newborn hearing screening issues and evidence,
White, K. R., Vohr, B. R. et al (2005). 'A multisite study to examine the efficacy of the
otoacoustic emission/automated auditory brainstem response newborn hearing screening
protocol: research design and results of the study', Am J Audiol, 14 (2), S186-199.
Universal neonatal hearing screening
267
White, S. J. & White, R. E. (1987). 'The effects of hearing status of the family and age of
intervention on receptive and expressive oral language skills in hearing-impaired infants',
AHSA Monographs, (26), 9-24.
Whiting P, R. A., Reitsma JB, Bossuyt PM, Kleijnen J. (2003). 'The development of
QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in
systematic reviews.' BMC Med Res Methodol., 3 (1), 25.
Yee-Arellano, H. M., Leal-Garza, F. & Pauli-Muller, K. (2006). 'Universal newborn
hearing screening in Mexico: Results of the first 2 years', International Journal of Pediatric
Otorhinolaryngology, 70 (11), 1863-1870.
Yoshinaga-Itano, C., Coulter, D. & Thomson, V. (2001). 'Developmental outcomes of
children with hearing loss born in Colorado hospitals with and without universal
newborn hearing screening programs', Semin Neonatol, 6 (6), 521-529.
Yoshinaga Itano, C., Coulter, D. & Thomson, V. (2000). 'The Colorado Newborn
Hearing Screening Project: effects on speech and language development for children with
hearing loss', Journal of Perinatology, 20 (8 part 2), S132-137.
Zaputovic, S., Stimac, T. et al (2005). 'Molecular analysis in diagnostic procedure of
hearing impairment in newborns', Croatian Medical Journal, 46 (5), 797-800.
268
Universal neonatal hearing screening
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement