Field sensitivity of targeted neonatal hearing screening by transient-evoked otoacoustic emissions

Modelling the cost-effectiveness of a newborn hearing screening programme; usability and pitfalls

OBJECTIVE

The EUSCREEN project concerns the study of European vision and hearing screening programmes. Part of the project was the development of a cost-effectiveness model to analyse such programmes. We describe the development and usability of an online tool to enable stakeholders to design, analyse or modify a newborn hearing screening (NHS) programme.

DESIGN

Data from literature, from existing NHS programmes, and observations by users were used to develop and refine the tool. Required inputs include prevalence of the hearing impairment, test sequence and its timing, attendance, sensitivity, and specificity of each screening step. Outputs include the number of cases detected and the costs of screening and diagnostics.

STUDY SAMPLE

Eleven NHS programmes with reliable data.

RESULTS

Three analyses are presented, exploring the effect of low attendance, number of screening steps, testing in the maternity ward, or screening at a later age, on the benefits and costs of the programme. Knowledge of the epidemiology of a staged screening programme is crucial when using the tool.

CONCLUSIONS

This study presents a tool intended to aid stakeholders to design a new or analyse an existing hearing screening programme in terms of benefits and costs.

A programme of studies including assessment of diagnostic accuracy of school hearing screening tests and a cost-effectiveness model of school entry hearing screening programmes

BACKGROUND

Identification of permanent hearing impairment at the earliest possible age is crucial to maximise the development of speech and language. Universal newborn hearing screening identifies the majority of the 1 in 1000 children born with a hearing impairment, but later onset can occur at any time and there is no optimum time for further screening. A universal but non-standardised school entry screening (SES) programme is in place in many parts of the UK but its value is questioned.

OBJECTIVES

To evaluate the diagnostic accuracy of hearing screening tests and the cost-effectiveness of the SES programme in the UK.

DESIGN

Systematic review, case-control diagnostic accuracy study, comparison of routinely collected data for services with and without a SES programme, parental questionnaires, observation of practical implementation and cost-effectiveness modelling.

SETTING

Second- and third-tier audiology services; community.

PARTICIPANTS

Children aged 4-6 years and their parents.

MAIN OUTCOME MEASURES

Diagnostic accuracy of two hearing screening devices, referral rate and source, yield, age at referral and cost per quality-adjusted life-year.

RESULTS

The review of diagnostic accuracy studies concluded that research to date demonstrates marked variability in the design, methodological quality and results. The pure-tone screen (PTS) (Amplivox, Eynsham, UK) and HearCheck (HC) screener (Siemens, Frimley, UK) devices had high sensitivity (PTS ≥ 89%, HC ≥ 83%) and specificity (PTS ≥ 78%, HC ≥ 83%) for identifying hearing impairment. The rate of referral for hearing problems was 36% lower with SES (Nottingham) relative to no SES (Cambridge) [rate ratio 0.64, 95% confidence interval (CI) 0.59 to 0.69; p < 0.001]. The yield of confirmed cases did not differ between areas with and without SES (rate ratio 0.82, 95% CI 0.63 to 1.06; p = 0.12). The mean age of referral did not differ between areas with and without SES for all referrals but children with confirmed hearing impairment were older at referral in the site with SES (mean age difference 0.47 years, 95% CI 0.24 to 0.70 years; p < 0.001). Parental responses revealed that the consequences to the family of the referral process are minor. A SES programme is unlikely to be cost-effective and, using base-case assumptions, is dominated by a no screening strategy. A SES programme could be cost-effective if there are fewer referrals associated with SES programmes or if referrals occur more quickly with SES programmes.

CONCLUSIONS

A SES programme using the PTS or HC screener is unlikely to be effective in increasing the identified number of cases with hearing impairment and lowering the average age at identification and is therefore unlikely to represent good value for money. This finding is, however, critically dependent on the results of the observational study comparing Nottingham and Cambridge, which has limitations. The following are suggested: systematic reviews of the accuracy of devices used to measure hearing at school entry; characterisation and measurement of the cost-effectiveness of different approaches to the ad-hoc referral system; examination of programme specificity as opposed to test specificity; further observational comparative studies of different programmes; and opportunistic trials of withdrawal of SES programmes.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN61668996.

FUNDING

This project was funded by the NIHR Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 20, No. 36. See the NIHR Journals Library website for further project information.

Current audiological diagnostics

Today's audiological functional diagnostics is based on a variety of hearing tests, whose large number takes account of the variety of malfunctions of a complex sensory organ system and the necessity to examine it in a differentiated manner and at any age of life. The objective is to identify nature and origin of the hearing loss and to quantify its extent as far as necessary to dispose of the information needed to initiate the adequate medical (conservative or operational) treatment or the provision with technical hearing aids or prostheses. Moreover, audiometry provides the basis for the assessment of impairment and handicap as well as for the calculation of the degree of disability. In the present overview, the current state of the method inventory available for practical use is described, starting from basic diagnostics over to complex special techniques. The presentation is systematically grouped in subjective procedures, based on psychoacoustic exploration, and objective methods, based on physical measurements: preliminary hearing tests, pure tone threshold, suprathreshold processing of sound intensity, directional hearing, speech understanding in quiet and in noise, dichotic hearing, tympanogram, acoustic reflex, otoacoustic emissions and auditory evoked potentials. Apart from a few still existing gaps, this method inventory covers the whole spectrum of all clinically relevant functional deficits of the auditory system.

Detection of Deafness in Puppies Using a Hand-Held Otoacoustic Emission Screener

The purpose of this study was to evaluate the use of a hand-held otoacoustic emissions screener to detect deafness in puppies. Specifically, distortion product otoacoustic emissions were recorded from 34 puppies (both sexes) of a variety of breeds, from 6-10 wk of age, and the results were compared to brainstem auditory evoked responses (BAER) recorded from the same puppies. Recordings were obtained from both ears in awake or lightly anesthetized puppies, and the results from each ear were compared. In all 62 ears that had normal BAERs, the distortion product otoacoustic emissions screener gave a response of "Pass." The three puppies that had flat BAER recordings in one or both ears provided a screener result of "Refer." In two ears with unusual BAERs (waveforms with reduced amplitudes and prolonged latencies) and a "Refer" response from the screener, there was compacted debris in one external ear canal, and the other ear canal was normal. The screener technology has proven application in human infants and is an attractive alternative to BAER testing in puppies because of expense and ease of use.

Economic evaluation of long-term impacts of universal newborn hearing screening

OBJECTIVE

Little is known about the long-term efficacious and economic impacts of universal newborn hearing screening (UNHS).

DESIGN

An analytical Markov decision model was framed with two screening strategies: UNHS with transient evoked otoacoustic emission (TEOAE) test and automatic acoustic brainstem response (aABR) test against no screening. By estimating intervention and long-term costs on treatment and productivity losses and the utility of life years determined by the status of hearing loss, we computed base-case estimates of the incremental cost-utility ratios (ICURs). The scattered plot of ICUR and acceptability curve was used to assess the economic results of aABR versus TEOAE or both versus no screening.

STUDY SAMPLE

A hypothetical cohort of 200,000 Taiwanese newborns.

RESULTS

TEOAE and aABR dominated over no screening strategy (ICUR = $-4800.89 and $-4111.23, indicating less cost and more utility). Given $20,000 of willingness to pay (WTP), the probability of being cost-effective of aABR against TEOAE was up to 90%.

CONCLUSIONS

UNHS for hearing loss with aABR is the most economic option and supported by economically evidence-based evaluation from societal perspective.

Gentamicin extended interval regimen and ototoxicity in neonates

OBJECTIVES

To assess the extended interval regimen gentamicin associated ototoxicity in neonatal intensive care unit using hearing tests.

METHODS

Two hundred and twenty neonates admitted to neonatal intensive care were assessed; 110 neonates who had received gentamicin and 110 neonates who had not received gentamicin served as control group. Gentamicin group were further subdivided according to the duration of treatment into 50 neonates who had received gentamicin for 5 days or less and 60 neonates who had received gentamicin for more than 5 days. TEOAEs were used for hearing screening. Auditory brain response was performed 3 months later for failed cases to confirm the hearing impairment.

RESULTS

Three neonates failed TEOAEs screening in each group but hearing impairment was confirmed in one neonate only (0.9%) in each group (gentamicin and control groups). Neonates who received gentamicin for more than 5 days showed comparable results as regard TEOAEs or ABR results with those who received gentamicin for 5 days or less, and control group.

CONCLUSIONS

Extended interval dosing of gentamicin therapy in neonates does not increase the incidence of hearing loss. This suggests that hearing loss in neonatal intensive care unit may be attributed to factors other than gentamicin treatment.

The economics of screening infants at risk of hearing impairment: an international analysis

OBJECTIVE

Hearing impairment in children across the world constitutes a particularly serious obstacle to their optimal development and education, including language acquisition. Around 0.5-6 in every 1000 neonates and infants have congenital or early childhood onset sensorineural deafness or severe-to-profound hearing impairment, with significant consequences. Therefore, early detection is a vitally important element in providing appropriate support for deaf and hearing-impaired babies that will help them enjoy equal opportunities in society alongside all other children. This analysis estimates the costs and effectiveness of various interventions to screen infants at risk of hearing impairment.

METHODS

The economic analysis used a decision tree approach to determine the cost-effectiveness of newborn hearing screening strategies. Two unique models were built to capture different strategic screening decisions. Firstly, the cost-effectiveness of universal newborn hearing screening (UNHS) was compared to selective screening of newborns with risk factors. Secondly, the cost-effectiveness of providing a one-stage screening process vs. a two-stage screening process was investigated.

RESULTS

Two countries, the United Kingdom and India, were used as case studies to illustrate the likely cost outcomes associated with the various strategies to diagnose hearing loss in infants. In the UK, the universal strategy incurs a further cost of approximately £2.3 million but detected an extra 63 cases. An incremental cost per case detected of £36,181 was estimated. The estimated economic burden was substantially higher in India when adopting a universal strategy due to the higher baseline prevalence of hearing loss. The one-stage screening strategy accumulated an additional 13,480 and 13,432 extra cases of false-positives, in the UK and India respectively when compared to a two-stage screening strategy. This represented increased costs by approximately £1.3 million and INR 34.6 million.

CONCLUSIONS

The cost-effectiveness of a screening intervention was largely dependent upon two key factors. As would be expected, the cost (per patient) of the intervention drives the model substantially, with higher costs leading to higher cost-effectiveness ratios. Likewise, the baseline prevalence (risk) of hearing impairment also affected the results. In scenarios where the baseline risk was low, the intervention was less likely to be cost-effective compared to when the baseline risk was high.

Is discordance in TEOAE and AABR outcomes predictable in newborns?

OBJECTIVE

To determine the perinatal predictors of discordant screening outcomes based on a two-stage screening protocol with transient-evoked otoacoustic emissions (TEOAE) and automated auditory brainstem response (AABR).

METHODS

A cross-sectional study of infants tested with TEOAE and AABR under a hospital-based universal newborn hearing screening program in Lagos, Nigeria. Maternal and infant factors associated with discordant TEOAE and AABR outcomes were determined with multivariable logistic regression analyses adjusting for potential confounding factors.

RESULTS

Of the 4718 infants enrolled under the program 1745 (36.9%) completed both TEOAE and AABR. Of this group, 1060 (60.7%) passed both TEOAE and AABR ("true-negatives"); 92 (5.3%) failed both TEOAE and AABR ("true-positive"); 571 (32.7%) failed TEOAE but passed AABR ("false-positives") while 22 (1.3%) passed TEOAE but failed AABR ("false-negatives"). Infants with false-positives were likely to be admitted into well-baby nursery (p=0.001), belong to mothers who attended antenatal care (p=0.010) or who delivered vaginally (p<0.001) compared to infants with true-negatives while infants with true-positives were also more likely to be delivered vaginally (p=0.002) or admitted into well-baby nursery (p=0.035) compared to infants with false-negatives. Infants with true-positives were significantly more likely to be delivered vaginally (p<0.001) and have severe hyperbilirubinemia (p=0.045) compared with infants with true-negatives. No association was observed between false-negatives and true-negatives. Antenatal care status, mode of delivery and nursery type were useful predictors of discordant outcomes among all infants undergoing screening (c-statistic=0.73).

CONCLUSIONS

Given the available screening technologies, discordant TEOAE and AABR may be inevitable for some categories of hearing loss among apparently healthy newborns whose mothers received prenatal care. The potential limitations of perinatal morbidities as basis of targeted screening for such cases therefore merit further consideration.

Language screening in the pediatric office setting

This article discusses screening of language development in a pediatric office setting. It describes the relationship between language delay and various developmental disorders. It provides recent recommendations regarding the efficacy of formal language screening instruments, and suggests developmental guidelines for clinical observations. Referrals for specialty evaluations and services for evaluation and treatment are presented. The article offers suggestions for counseling parents when a language disorder is suspected.

A Multisite Study to Examine the Efficacy of the Otoacoustic Emission/Automated Auditory Brainstem Response Newborn Hearing Screening Protocol

Purpose: Most newborns are screened for hearing loss, and many hospitals use a 2-stage protocol in which all infants are screened first with otoacoustic emissions (OAEs). In this protocol, no additional testing is done for those passing the OAE screening, but infants failing the OAE are also screened with automated auditory brainstem response (A-ABR). This study evaluated how many infants who failed the OAE and passed the A-ABR had permanent hearing loss (PHL) at 8–12 months of age.

Method: A total of 86,634 infants were screened at 7 birthing centers using a 2-stage OAE/A-ABR hearing screening protocol. Of infants who failed the OAE but passed the A-ABR, 1,524 were enrolled in the study. Diagnostic audiologic evaluations were performed on 64% of the enrolled infants (1,432 ears from 973 infants) when they were 8–12 months old.

Results: Twenty-one infants (30 ears) who passed the newborn A-ABR hearing screening were identified with PHL when they were 8–12 months old. Most (71%) had mild hearing loss.

Conclusions: If all infants were screened for hearing loss using a typical 2-stage OAE/A-ABR protocol, approximately 23% of those with PHL at 8–12 months of age would have passed the A-ABR.

A Multisite Study to Examine the Efficacy of the Otoacoustic Emission/Automated Auditory Brainstem Response Newborn Hearing Screening Protocol

Purpose: This 3rd of 4 articles on a study of the efficacy of the 2-stage otoacoustic emission/automated auditory brainstem response (OAE/A-ABR) newborn hearing screening protocol describes (a) the behavioral audiometric protocol used to validate hearing status at 8–12 months of age, (b) the hearing status of the sample, and (c) the success of the visual reinforcement audiometry (VRA) protocol across 7 sites.

Method: A total of 973 infants who failed OAE but passed A-ABR, in one or both ears, during newborn screening were tested with a VRA protocol, supplemented by tympanometry and OAE screening at age 8–12 months.

Results: VRA audiograms (1.0, 2.0, and 4.0 kHz) were obtained for 1,184 (82.7%) of the 1,432 study ears. Hearing loss was ruled out in another 100 ears by VRA in combination with OAE, for a total of 88.7% of the study sample. Permanent hearing loss was identified in 30 ears of 21 infants. Sites differed in their success with the VRA protocol.

Conclusions: Continued monitoring of hearing beyond the newborn period is an important component of early detection of hearing loss. Using a structured protocol, VRA is an appropriate test method for most, but not all, infants. A battery of test procedures is often needed to adequately delineate hearing loss in infants. Examiner experience appears to be a factor in successful VRA.

Decreasing test time in newborn hearing screening

This study examined the effect of reducing the number of accepted responses in transiently evoked otoacoustic emissions based on the results of a universal neonatal hearing screening program. Our intention was to decrease the test time of newborns. A total of 464 ears were examined by using a universal newborn hearing-screening program implemented in a private maternity hospital. ILO88 Otodynamics Analyzer Quickscreen program was used for all testing and a two-stage procedure was adopted. In the first stage, the results were continuously evaluated for the 'pass' criteria, during the test, after at least 20 low-noise sweeps had been presented. As soon as the criteria were met, the test was interrupted and the results were recorded. In the second stage of the procedure, the test was continued and finally terminated after 260 quiet samples had been recorded. The results of each stage of this procedure were compared and evaluated. A total of 402 ears had normal otoacoustic emissions and only 62 ears had absent emissions. It was concluded that after a minimum of 20 averaged quiet responses, which we consider necessary in order to record reliable emissions and as soon as the 'pass' criteria were fulfilled the test could be terminated without affecting the 'pass-fail' rates which were similar in both stages. However, we believe that for diagnostic and clinical purposes, all 260 quiet samples must be used, because the results after the second stage indicated statistically better scores in response and reproducibility measures.

The Effect of Different ‘Pass-Fail’ Criteria on the Results of a Newborn Hearing Screening Program

'Pass' criteria in newborn hearing screening programs are important, since they affect the operating characteristics of the programs. In the present study, we intended to compare the results of two screening procedures, using different 'pass' criteria, in two samples from the same pool of screened newborns. The subjects were divided into two study groups, screened consecutively during 6 months. Testing and all procedures were exactly the same in both groups, differing only in the 'pass' criteria. In the first group a signal-to-noise ratio of at least 3 dB in the frequency bands of 1-2, 2-3 and 3-4 kHz was considered necessary for a 'pass', whereas a signal-to-noise ratio > or =6 dB was used in the second group, at the same frequency bands. During the period of the study, no other minor or major modification of the protocol was applied. The comparison of the screening predischarge results between the two groups showed no statistically significant differences in the 'pass-refer' results. Thus, it appears that the 3-dB signal-to-noise ratio is as valid as the 6-dB criterion, and it may be confidently used, especially in settings where rescreening is not available.

Otoacoustic Emissions in Universal Hearing Screening: Which Day after Birth Should We Examine the Newborns?

Transiently evoked otoacoustic emissions (TEOAEs) have been widely used in universal newborn hearing screening programs. Although there is consensus with regard to the avoidance of early screening, especially during the first hours after birth, the optimum testing day is not yet unanimously accepted. The aim of the present study was to compare the 'pass-refer' results between 4 groups of newborns tested during the 4 postbirth days and determine the most appropriate day for assessing newborn hearing. Our results suggest that, although TEOAEs can be recorded in very high rates from the first 24 h of life, 'refer' scores are lower on the third and fourth days after birth. It may be thus concluded that the optimum time of assessing newborn hearing in universal hearing screening programs seems to be the third or fourth postbirth day, provided that other social or financial reasons do not suggest an earlier discharge from the hospital.

Methodological factors involved in neonatal screening using transient-evoked otoacoustic emissions and automated auditory brainstem response testing

The methodological factors involved in screening neonates for hearing loss, using transient-evoked otoacoustic emissions (TEOAEs) and automated auditory brainstem responses, have been evaluated from a large sample of neonates. The risk factors, commonly used to select babies for a targeted screen, have very little correlation with failing TEOAE testing. The parameters used to determine passing or failing the TEOAE test and the false alarm rate change markedly with age in the first few days of life as, of course, did the percentage of babies who failed the test. The stimulus level used was the default setting for the Otodynamics equipment but the stimulus level measured in the ear canal decreased over the first 140 h of life. It is thought that this reflects the impedance changes in outer and middle ears and possible changes in middle ear dynamics. The methodological variables investigated here can illuminate some of the differences in previous reports of neonatal screening, in particular the reported hit and false alarm rates.

Click-evoked ABR at high stimulus repetition rates for neonatal hearing screening

A new, fast screening algorithm based on auditory brainstem response (ABR) recorded at a high click repetition rate is proposed. Response detection is carried out in the frequency domain by a statistical test procedure which includes the fundamental frequency and the harmonics below 800 Hz. First, the method was tested in 25 young adults. ABRs were recorded in the repetition rate range 20/s to 400/s. With a mean response detection time of 31 s, a click repetition rate of 140/s was found to be the optimum rate among the adult group. The method was then tested using a group of 114 neonates in whom the repetition rate range 60/s to 200/s was examined. At the repetition rate 90/s, which was found to be the optimum rate in neonates, the mean detection time was 24.6 s. In addition to the fast ABR detection, the proposed screening algorithm also allows simultaneous hearing screening of both ears using a one-channel data recording.

Universal neonatal hearing screening: past, present, and future

After a brief review of the history of newborn hearing screening including the Downs behavioral testing procedure, the Crib-o-gram and similar devices, and the use of auropalpebral reflex and otoacoustic emissions, there is a discussion of key issues that need to be resolved before universal hearing screening is introduced. Included are questions regarding the target population(s) of screening programs, well baby versus NICU screening, dealing with false-positives and the effects on parent-child relationships, and finally, the availability of resources for screening and follow-up. The results of a recent study in the United Kingdom that assessed the current state of audiology services and found there is a difference between existing standards and what is actually being done in practice, are presented and considered in terms of current trends in the United States to move ahead with universal screening without a solid database of information regarding the preparedness of clinical centers to deal with the need for services that will result from the initiation of universal programs. Caution is urged.

Cost and cost-effectiveness of universal screening for hearing loss in newborns

OBJECTIVE

To estimate the cost and cost-effectiveness of universal newborn hearing screening.

STUDY DESIGN AND SETTING

Decision analysis model utilizing the hospital perspective. This model evaluated 4 distinct protocols for screening a fixed and defined hypothetical cohort of newborn infants.

OUTCOME MEASURES

Cost of screening and the number of infants with hearing loss identified through universal screening.

RESULTS

Otoacoustic emissions testing at birth followed by repeat testing at follow up demonstrated the lowest cost ($13 per infant) and had the lowest cost-effectiveness ratio ($5100 per infant with hearing loss identified). Screening auditory brainstem evoked response testing at birth with no screening test at follow-up was the only protocol with greater effectiveness, but it also demonstrated the highest cost ($25 per infant) and highest cost-effectiveness ratio ($9500 per infant with hearing loss identified). These findings were robust to sensitivity analysis, including best-case and worst-case estimation. The prevalence of hearing loss and the fraction of infants returned for follow-up testing had a large impact on the absolute level, but not relative level of protocol cost and cost-effectiveness.

CONCLUSION

The otoacoustic emissions testing protocol should be selected by screening programs concerned with cost and cost-effectiveness, although there are certain caveats to consider.

SIGNIFICANCE

The most significant barriers to implementation of universal newborn hearing screening programs have been financial, and this study compares the most common protocols currently in use. This study can assist program directors not only in the decision to initiate universal screening but also in their choice of screening protocol.

[Importance of impedance audiometry on infant hearing screening test with otoacoustic emissions]

The target of this paper is to evaluate the importance of Impedanciometry in a protocol with transient evoked otoacoustic emissions on high risk infant hearing screening. We used tympanometry and stapedious reflex only when otoacoustic emissions became altered. This method try to decrease the impact in this test caused for middle ear diseases. We realized an Impedanciometry study in all children with abnormal otoacoustic responses and then, we obtained normal otoacoustic responses when tympanometry normalize. In the other hand, the time of test wasn't too large, between 15-20 minutes for child. The analysis of specificity and positive predictive value, of otoacoustic emissions without Impedanciometry was 89% and 45% respectively, however the same values with Impedanciometry was 96% and 75% respectively. In our opinion, high frequency of middle ear diseases in children, and our results, justify the introduction of Impedanciometry in a complete protocol of hearing screening with Otoacoustic Emissions, preventing appearance of false positive caused by these diseases.

The effect of high-pass filtering on TEOAE in 2-month-old infants

The aim of the present study was to investigate the effect of high-pass filtering on TEOAE obtained from 2-month-old infants as a function of filter cut-off frequency, activity states and pass/fail status of infants. Two experiments were performed. In Experiment 1, 100 2-month-old infants (200 ears) in five activity states (asleep, awake but peaceful, sucking a pacifier, feeding, restless) were tested by use of TEOAE technology. Five different filter conditions were applied to the TEOAE responses post hoc. The filter conditions were set at 781 Hz (default setting), 1.0, 1.2, 1.4 and 1.6 kHz. Results from this experiment showed that TEOAE parameters such as whole-wave reproducibility (WR) and signal-to-noise ratio (SNR) at 0.8 kHz and 1.6 kHz, changed as a function of the cut-off frequency. The findings suggest that the 1.6 kHz and 1.2 kHz filter conditions are optimal for WR and SNR pass/fail criteria, respectively. Although all infant recordings appeared to benefit from the filtering, infants in the noisy states seemed to benefit the most. In Experiment 2, the high-pass filtering technique was applied to 23 infants (35 ears) who apparently failed the TEOAE tests on initial screening but were subsequently awarded a pass status based on the results from a follow-up auditory brainstem response (ABR) assessment. The findings showed a significant decrease in noise contamination of the TEOAE with a corresponding significant increase in WR. With high-pass filtering at 1.6 kHz, 21/35 ears could be reclassified into the pass category.

Performance of neonatal and infant hearing screens: sensitivity and specificity

The terms 'sensitivity' and 'specificity' are defined and some of the factors that determine their values are discussed in the context of screening for permanent childhood hearing loss. There is a need to distinguish between the values observed in 'simple experiments' and those that may be obtained under more realistic 'field' conditions. It is not feasible to give a meta-analytic overview of published data because of the variety of methods and objectives used in those studies published in the literature. However, a qualitative synthesis of the data is possible. This suggests that most proposed neonatal hearing screening tests, when implemented in accordance with a programme of quality assurance, can be reasonably accurate at a modest cost. However, the optimal combination of tests and test parameters for given populations has not yet been fully researched. The infant distraction test screen has a lower sensitivity than neonatal hearing screening tests, particularly for moderate impairments, accompanied by a fairly low specificity.

Identification of neonatal hearing impairment: hearing status at 8 to 12 months corrected age using a visual reinforcement audiometry protocol

OBJECTIVES

1) To describe the hearing status of the at-risk infants in the National Institutes of Health-Identification of Neonatal Hearing Impairment study sample at 8 to 12 mo corrected age (chronologic age adjusted for prematurity). 2) To describe the visual reinforcement audiometry (VRA) protocol that was used to obtain monaural behavioral data for the sample.

DESIGN

All neonatal intensive care unit infants and well babies with risk factors (including well babies who failed neonatal tests) were targeted for follow-up behavioral evaluation once they had reached 8 mo corrected age. Three thousand one hundred and thirty-four (64.4%) of the 4868 surviving infants returned for at least one behavioral hearing evaluation, which employed a well-defined VRA protocol. VRA thresholds or minimum response levels (MRLs) were determined for speech and pure tones of 1.0, 2.0, and 4.0 kHz for each ear using insert earphones.

RESULTS

More than 95% of the infants were reliably tested with the VRA protocol; 90% provided complete tests (four MRLs for both ears). Ninety-four percent of the at-risk infants were found to have normal hearing sensitivity (MRLs of 20 dB HL) at 1.0, 2.0, and 4.0 kHz in both ears. Of the infants, 2.2% had bilateral hearing impairment, and 3.4% had impairment in one ear only. More than 80% of the impaired ears had losses of mild-to-moderate degree.

CONCLUSIONS

This may be the largest study to attempt to follow all at-risk infants with behavioral audiometric testing, regardless of screening outcome, in an effort to validate the results of auditory brain stem response, distortion product otoacoustic emission, and transient evoked otoacoustic emission testing in the newborn period. It is one of only a few studies to report hearing status of infants at 1 yr of age, using VRA on a clinical population. Successful testing of more than 95% of the infants who returned for the VRA follow-up documents the feasibility of obtaining monaural behavioral data in this population.

Automated auditory brainstem response in neonatal hearing screening

Severe congenital hearing impairment is an important handicap affecting 0.1% of apparently healthy liveborn infants and 1-2% of graduates of neonatal intensive care units. The prognosis for intellectual, emotional, language and speech development in the hearing-impaired child is improved when the diagnosis is made early and intervention is begun before the age of 6 mo. Universal screening is preferable, since about 50% of infants with hearing loss are not discovered if neonatal hearing screening is restricted to high-risk groups. The automated auditory brainstem response (AABR) screener is a dedicated hearing screening device which provides information not only about the outer/middle ear and cochlea but also about the auditory pathway up to the brainstem. AABR has an agreement with conventional auditory brainstem response up to 98%. It uses a 35 dB near hearing level click. No operator interpretation is needed and it can be used on the ward and during oxygen therapy without disturbance from ambient noise. Reported referral rates in a hospital-based screening programme at the time of discharge vary, with an average of 4%. AABR has also been used in a home-based setting, with the same results. The time necessary for screening varies with the setting, but ranges from 4 to 15 min. Initial costs range from $15 to $25 per test, which is similar to neonatal screening for metabolic diseases. In addition to individual healthcare savings, early diagnosis may lead to savings on costs of intensive speech-language intervention and educational facilities.

Health visitor screening versus vigilance: outcomes of programmes for detecting permanent childhood hearing loss in west Berkshire

The Health Visitor Distraction Test (HVDT) screen for hearing was replaced in West Berkshire in 1989 with a vigilance programme incorporating a questionnaire. The detection of permanent congenital deafness (bilateral > 50 dB HL) for all children born since 1984 has been audited to compare the two regimes. Sixty-two cases met the criteria, giving an ascertainment of 1.0 per 1000. Performance was similar under the two systems for severe/profound losses (> 70 dB HL), but there was a longer tail of late-detected moderate losses (50-70 dB HL) under the vigilance regime. The sensitivity of the Health Visitor questionnaire in referring those with permanent hearing loss was very similar to that of the HVDT (39% compared with 42%). Coverage for the questionnaire was approximately 87%, but only 78% for the known cases. Referral rate was lower under the vigilance programme, at approximately 3%. The results suggest that a vigilance programme is likely to perform as well as the HVDT but no better. Despite subsequent modifications to the vigilance programme, the poor pickup of moderate losses probably indicates the limitations of parental and professional observation in detecting partial hearing problems. The evidence adds support to the recent recommendations for universal neonatal screening.

Transient evoked otoacoustic emissions (TEOAEs) in new-borns: normative data

OBJECTIVE

Early diagnosis and rehabilitation of congenital hearing loss are mandatory in order to achieve a satisfactory linguistic and cognitive development. A universal hearing screening in order to identify congenital hearing losses before 3 months of age is required.

METHODS

TEOAEs are an easy to perform, short lasting, not invasive and low-cost test with a high sensitivity. 320 at term new-borns (640 ears) without any risk factor for hearing loss underwent TEOAEs. The new-borns were screened 3 days after birth. Those who failed the first test were retested when possible before the discharge from the hospital. ABR was performed 3 months later in cases who failed TEOAE.

RESULTS

The median TEOAE sampling time was 98 s, the median test duration was 14 min. The mean stimulus amplitude was 80 dB peSPL in the left ear and 81 dB peSPL in the right ear, noise levels within the external meatus during sampling were 44 dB SPL on the right ear and 43 dB SPL on the left one, noise contained within the response (A-B difference) was 8.65 dB SPL in the left ear and 8.74 dB SPL in the right ear, mean TEOAEs amplitudes were 21.49 dB SPL and 21.78 dB SPL in the right and left ear respectively, the mean lower and upper limit of the spectrum being 678 and 5720 Hz. According to these criteria 494/640 ears (77.2%) passed the test at the first recording, while TEOAEs resulted to be absent in 146/640 ears (22.8%). A retest was performed successfully before the discharge from the Hospital in 30/640 ears (4.7%). An ABR recording within the third month of life was scheduled as out-patient in the 58 new-borns (116 ears, 18.2%) who failed the test. 18 of them (36 ears, 5.6%) did not complete the program, 19 new-borns (38 ears, 11.8%) showed a normal ABR, while two new-borns (four ears, 0.6%) failed ABR after 3 months. A second ABR performed after 6 months was normal.

CONCLUSIONS

TEOAEs recording seems at now the test of choice for a universal hearing screening. However, a greater standardization of criteria both in performing the test and in evaluating the results is needed.

Factors influencing the efficacy of universal newborn hearing screening

The debate surrounding the issue of universal hearing screening is being carried out on several levels. Although little disagreement exists over the educational, vocational, and quality-of-life benefits that would result from early identification and timely intervention of congenital hearing loss, the pragmatic issues, such as the effectiveness and the cost benefits associated with universal screening, cannot be ignored. This means that sensitivity, specificity, prevalence, and predictive value remain important factors. Determining the number of infants born with hearing loss in the United States each year, the prevalence issue is key to calculating the predictive value of newborn hearing screening. Emerging from current studies is that estimates of prevalence in the universal newborn population vary from 0.9 in 1000 for permanent bilateral hearing loss of more than 35 dB, to 3.24 in 1000 for bilateral hearing loss, to 5.95 in 1000 when unilateral and moderate hearing loss infants are counted. By comparison, the incidence of hearing loss in the NICU or at-risk population is accepted as high, somewhere between 2% and 4% or 20 to 40 in 1000. Incidence in the NICU varies depending on admission policies and level of care. In general, however, by screening the NICU and targeted at-risk populations, estimated to make up 10% to 16% of the newborn population, half or more of all newborns with severe to profound educationally disabling hearing loss are identified. Data from several well-conducted clinical studies, dating back to the first studies on the use of ABR to screen in the NICU, provide ample justification for the recommendation that all infants admitted to an NICU for longer than 24 hours should be screened for hearing impairment regardless of whether they have any of the at-risk indicators for hearing loss. In the author's opinion, screening in the NICU should be modeled on the operator-controlled ABR protocol outlined by Galambos and colleagues, with the addition that every ABR fail be screened by OAE before discharge. Unlike the targeted NICU population, the question remains for well infants, is sufficient clinical data or evidence available to justify screening all well newborns, specifically those with none of the at-risk for hearing loss indicators cared for in the well-infant nursery and who are discharged home within 24 or 48 hours? With the steady increase in the number of hospital-based universal newborn hearing screening programs implemented since the NIH Consensus Statement, additional data should become available to help resolve several outstanding issues, including prevalence and the predictive value of the various test protocols currently in use or proposed.

Newborn and infant hearing loss: detection and intervention.American Academy of Pediatrics. Task Force on Newborn and Infant Hearing, 1998- 1999

This statement endorses the implementation of universal newborn hearing screening. In addition, the statement reviews the primary objectives, important components, and recommended screening parameters that characterize an effective universal newborn hearing screening program.

Anomalous screening outcomes from click-evoked otoacoustic emissions and auditory brainstem response tests

Transiently evoked otoacoustic emissions (TEOAE) and auditory brainstem response (ABR) tests were used in parallel to screen 862 neonates with risk factors for hearing loss. Seven neonates (0.81%) passed on TEOAE screen yet failed on ABR in one ear (six neonates) or both ears (one neonate). This combination of results has been termed 'anomalous'. Examination of audiometric results obtained on follow-up shows that in one ear of one neonate the result was consistent with the later confirmed audiogram shape. The explanation for the anomalous results in the remaining ears is unclear although neural maturation and the effects of hyperbilirubinaemia are possibilities. There were no instances of progressive or retrocochlear hearing loss identified. None of the seven neonates had better ear hearing loss of > or = 40 dB on long term follow-up.

Field sensitivity of targeted neonatal hearing screening using the Nottingham ABR Screener

OBJECTIVE

A population of children having bilateral moderate to profound hearing impairment was investigated to find out the number who passed (false negatives) and who failed (true positives) a screening test based on the auditory brain stem response (ABR).

DESIGN

This study of the ABR is a parallel report to Lutman, Davis, Fortnum, and Wood (1997), where the transient evoked otoacoustic emission was examined on a similar but not identical population of at-risk neonates. Hearing screening was undertaken in at-risk neonates (targeted screening) at seven hospitals in the UK using the Nottingham ABR Screener. During the period from January 1988 to December 1993, a total of 6983 neonates had been tested. Assessment of audiological records from the participating centers ascertained that 201 children born between January 1988 and December 1993 had hearing threshold levels in both ears of 50 dB or more, averaged over the speech frequencies 0.5, 1, 2, and 4 kHz. Of these, 51 had completed the ABR screening test. Examination of the ABR records, which included both machine and visual scoring, indicated whether they had passed or failed the screening test.

RESULTS

Forty-six of the 51 hearing impaired babies failed the neonatal ABR screen either on one or both ears (five false negatives), and 42 failed the test on both ears (nine false negatives). This resulted in field sensitivities of 90% and 82%, respectively. Two cases of suspected progressive hearing loss have been included in the numbers of false negative results. The percentage of babies passing the screen on both ears and subsequently having normal hearing (specificity) is typically 93%.

CONCLUSIONS

Neonatal hearing screening in an at-risk population using a highly automated ABR test is a viable and effective tool for identification of hearing impairment. Although the field sensitivity of the test is high, it is unable to identify all babies with a criterion level of hearing loss. There are a number of possible explanations as to the origin of false negative results: configuration of the pure-tone audiogram, a progressive hearing loss, acquired sensorineural or conductive loss, retrocochlear deafness, or an incorrect interpretation of the screening test. Passing a neonatal screening test, therefore, does not exclude the possibility of subsequent hearing impairment and highlights the need for further surveillance.

Epidemiology of permanent childhood hearing impairment in Trent Region, 1985-1993

This retrospective study of permanent childhood hearing impairment (PCHI) > or = 40 dB HL in children born between 1985 and 1993 and resident in Trent Health Region, achieved an ascertainment of 92.9% of that expected from previous studies and 100% for the subset of children born between 1985 and 1990. The prevalence rate of all permanent hearing impairment > or = 40 dB HL for the birth cohort 1985-90 is 133 (95% confidence interval, (ci) 122-145) per 100,000 live births (1 in 750). Sixteen per cent of PCHI were postnatally acquired, late-onset or progressive impairments. Excluding these, the prevalence rate for congenital impairments is 112 (ci 101-123) per 100,000 (1 in 900). The rate for profound impairments > or = 95 dB HL is 24 (ci 20-30) per 100,000 live births (1 in 4150). Prevalence was increased sixfold for children with a history of neonatal intensive care and 14-fold for children with a family history, compared with children with no risk factors. A more than two-fold increase in prevalence was seen in Asian children. For the congenitally-impaired children born between 1985 and 1990, 29% had a stay in neonatal intensive care > or = 48 hours, 30% had a family history of permanent childhood hearing impairment, and 12% had a cranio-facial abnormality (CFA). Over 59% were potentially detectable by a targeted neonatal screening programme using these three high-risk factors. For 1985-1993, the overall yield of the targeted neonatal screening programmes available in three of the 11 health districts was 15% but increased over time. The overall yield from the Health Visitor distraction test was 30% but lower in districts with neonatal screening programmes. Only 59% of children had a stated aetiology, classified by time of onset into genetic, including syndromes and CFA (41%), pre- or peri-natal (10%), post-natally acquired (6%), and uncertain onset (2%). Just under 40% of the children were said to have another clinical or developmental problem, about half of whom had at least two additional problems. The median age at referral, confirmation of the impairment, prescription of the hearing aid and fitting of the hearing aid were, respectively, 10.4 months, 18.1 months, 24.4 months and 26.3 months. A more severe impairment was associated with earlier age. Small improvements in the median age of hearing aid prescription and fitting were seen over time. Twenty-five per cent of children were referred for genetic counselling, the proportion increasing systematically with the severity of the impairment. Based on evidence of the yield from hearing screens we suggest a wider implementation of neonatal screening and further consideration of the role of the health visitor distraction test in the identification of children with PCHI. To facilitate further assessment of services for hearing-impaired children we suggest implementation of a co-ordinated shared list of children with permanent hearing impairment on a region-wide basis to provide adequate numbers for comparison over time, and the routine collection of a minimum set of data for each child.