|Year : 2016 | Volume
| Issue : 4 | Page : 237-242
Automated auditory brainstem response: Its efficacy as a screening tool for neonatal hearing screening in the postnatal ward
Chavakula Rajkumar, Ann Mary Augustine, Anjali Lepcha, Achamma Balraj
Department of ENT, Christian Medical College, Vellore, Tamil Nadu, India
|Date of Web Publication||13-Oct-2016|
Ann Mary Augustine
Department of ENT, Christian Medical College, Vellore, Tamil Nadu
Source of Support: None, Conflict of Interest: None
Introduction: This observational study was carried out to determine the sensitivity and specificity of MB11 BERAphone® , when used for neonatal hearing screening in a postnatal ward setting in comparison against the gold standard, auditory brainstem response (ABR). Materials and Methods: Thirty-seven consecutive newborns (74 ears) who either unilaterally or bilaterally failed hearing screening with MB11 BERAphone in the postnatal ward were recruited and a second screening with BERAphone was performed after 1 week along with confirmatory testing using ABR. Results: MB11 BERAphone showed sensitivity of 92.9%, specificity of 50%, positive predictive value of 30.23%, and negative predictive value of 96.77% for the diagnosis of hearing loss. The prevalence of confirmed hearing impairment was 18.9%. The rate of unilateral impairment was 10.8%, and the rate of bilateral impairment was 13.5%. The average ambient noise levels in the postnatal ward setting was 62.1 dB. Conclusion: Although the sensitivity of MB11 BERAphone is good, the specificity is significantly lower when the test is performed in the postnatal ward setting with high ambient noise. Neonates who fail the two-step screening should undergo auditory response for confirming the diagnosis of hearing loss.
Keywords: Auditory brainstem response, congenital hearing loss, neonatal hearing screening
|How to cite this article:|
Rajkumar C, Augustine AM, Lepcha A, Balraj A. Automated auditory brainstem response: Its efficacy as a screening tool for neonatal hearing screening in the postnatal ward. Indian J Otol 2016;22:237-42
|How to cite this URL:|
Rajkumar C, Augustine AM, Lepcha A, Balraj A. Automated auditory brainstem response: Its efficacy as a screening tool for neonatal hearing screening in the postnatal ward. Indian J Otol [serial online] 2016 [cited 2019 May 26];22:237-42. Available from: http://www.indianjotol.org/text.asp?2016/22/4/237/192134
| Introduction|| |
The prevalence rates of congenital hearing impairment ranges from 1.2 to 5.7 per 1000 live births. , Of these, approximately 90% of newborns are born to parents who can hear. , The significant phase for speech and language development is around the time from birth to about 5 years of age. Data from cohort studies indicate that the crucial time for diagnosis and intervention is earlier than 6 months of age since it provides opportunity for the improvement of language and speech.  Delay in the identification and rehabilitation would lead to marked delay in the language and cognitive skills which hinder the overall development of the child. The "National Institutes of Health Consensus Development Conference on Early Identification of Hearing Impairment in Infants and Children" therefore recommends a universal neonatal hearing screening for the early identification of hearing impairment among newborns. 
Two methods are commonly used in universal newborn hearing screening: Otoacoustic emissions (OAEs) and automated auditory brainstem response (aABR), both methods being automated. aABR is superior to OAE, among the screening tests, because the aABR will be able to identify retrocochlear hearing impairments, such as auditory neuropathy which are missed on OAE screening. The aABR test uses a series of click sounds at 35 dB hearing level and detects brainstem responses to these stimuli. Usually, 3 out of 100 babies are referred for a full diagnostic assessment of hearing. aABR has also been found to be time and cost-effective, with a high sensitivity and a low failure rate.  The time needed for screening is 4-15 min.  Further test interpretation by an audiologist is not needed since the automated screener gives a "pass-fail" report. A "refer" report given by the aABR shows that the hearing level is worse than 35 dB. Each ear can be tested individually. It can be used within the first 24 h after birth in the nursery. The low rate of screening failures decreases any potential of false positives during screening, minimizing the costs associated with later follow-up assessments. 
The MB11 BERAphone is an aABR which has been found to be an effective neonatal hearing screening tool. To identify the presence of an ABR response automatically, the device uses a chirp stimulus and takes a maximum test time of 180 s based on an implemented statistical test algorithm (modified q-sample uniform scores test) as given by Cebulla and Shehata-Dieler.  This implemented test algorithm attempts to detect a response from the auditory system. The screening test generates a "pass" result when a response is detected. The screening test generates a "refer" result if there is no response within 180 s.  The device is set at a screening stimulus level of 35 dB HL for a pass.
MB11 has shown a considerable advantage with regard to the testing time with a median testing time of 28 s.  Melagrana et al.  in their study also concluded that screening of babies in an intensive care environment with MB11 screening test was appropriate. van Straaten, in his study using automated ABR, reported a 92% "pass" in the first screening and 98% "pass" after the second screening.  As claimed by the manufacturer, MB11 BERAphone has a test sensitivity and specificity of 99% and 97%, respectively, for a two-stage test.  This is possible only when the test is performed in ideal conditions. This study was hence done to determine the sensitivity and specificity of MB11 BERAphone®, in identifying neonates with congenital hearing loss in a postnatal ward setting (when performed at the bedside of the newborn where the sound levels are over 35 dB) in comparison against the gold standard, ABR, and to discuss the challenges of using the MB11 BERAphone® at the bedside in a ward setting for universal neonatal hearing screening.
| Materials and Methods|| |
This observational study was done between October 2012 and October 2013 at our tertiary care center after approval from the institutional review board and Ethical Committee Clearance. Neonates born in the institution who have been referred once on sequential screening with MB11 BERAphone® were recruited after obtaining informed consent from the parents.
For an expected sensitivity and specificity of 95% each and prevalence (in the second screen positives) of 0.5 and a precision of ± 0.07%, a sample size of 74 ears was required for the estimation of both sensitivity and specificity. Thirty-seven consecutive healthy neonates (74 ears) who were "referred" on the first screening with MB11 BERAphone® were recruited for the study which was performed in the postnatal ward setting and in the audiovestibular unit. Neonates with upper respiratory tract infections, discharging ears, obvious anomalies, and very sick neonates were excluded from the study.
The screening protocol included two stages (two-step protocol) [Figure 1]. The first screening was done in the postnatal ward before discharge within 48 h of delivery at normal ambient sound levels which were above 35 dB. The neonatal screenings both the first and second were done using the device MAICO MB11 BERAphone® (MAICO Diagnostic GmbH, Berlin/Germany).
The MB11 BERAphone® is a screening device that is fitted with a handheld "headphone unit" that incorporates all the ingredients required for ABR recording such as the loudspeaker, a preamplifier, and a set of three integrated reusable electrodes. Two electrodes record the ABR from the vertex and mastoid while a third electrode serves as the ground reference. The position of the vertex electrode is adjustable to enable its optimal placement on different head sizes [Figure 2]a-c.
|Figure 2: (a) BERAphone unit, (b) Headphone unit, (c) Hand-held device with 3 integrated electrodes and transducer|
Click here to view
The mothers or carers were counseled about the test and consent was obtained. A complete case history including prenatal, perinatal, and family history was obtained from the parent or primary caretaker of the infant. History pertaining to various risk factors for hearing loss (TORCH infections, diabetes, hypothyroidism, ototoxic drug intake, and other illnesses) was recorded. Family history with regard to deafness, blindness, and other systemic diseases was noted. This was followed by a complete physical and ENT examination.
The screening process was organized and supervised by the audiovestibular unit of the Department of ENT. The test with the MB11 BERAphone® was performed by support staff recruited from the Department of Audiology. They had received prior training on the use of the device.
When the neonate was asleep or quiet, the special handheld headphone unit of MB11 BERAphone® which emits a series of soft clicks was positioned over the newborn's ears after application of electrode gel. The vertex electrode is adjustable to ensure its proper placement on various head sizes. These electrodes pick up the neural activity in response to the stimulus. Comparison of the neural responses with normal response templates is done by the software and a "pass" or "refer" result is displayed. If a response is detected and verified at 35 dBHL, the test result is "pass." The machine indicates a "refer" when there is no response at 35 dBHL at all frequencies. The electrodes were cleaned with antiseptic solutions before and after each use.
The second stage screening along with all audiological diagnostic tests was performed in an electrically and acoustically shielded room in the Department of Audiology where the noise level is below 20 dB. The mothers/carers after counseling were asked to report to the Department of Audiology in the outpatient department after 1 week for the second screening and for the confirmatory test. This was done to overcome the possibility of vernix in the ear of the neonate interfering with the first screening soon after birth.
All the 37 infants who had been "referred" or "passed" in the second screening with "MB11 BERAphone® were subjected to a confirmatory diagnostic testing with ABR which was used as the gold standard for confirming hearing loss in the neonates recruited for the study. The second screening with the MB11 BERAphone® was done at the same sitting before the ABR test. The subset of children who were "referred" in the first screening and have "passed" in the second screening was also subject to the confirmatory ABR testing to provide the second arm of the 2 × 2 table for calculation of sensitivity and specificity.
Following the second screening, the neonates underwent the confirmatory test ABR audiogram using the Intelligent Hearing Systems ABR Machine (ISO 13485, Miami, Florida). The ABR test was performed in a soundproof room by an experienced audiologist who was blinded to the results of the first and second screening by MB11 BERAphone®. The procedure of the test was explained to the mothers/carers, and once the neonate was comfortable, dry, breastfed, and naturally sleeping, the infant was placed in a secure area for testing. The skin for electrode application was prepared by carefully cleaning with a mildly abrasive skin preparation gel. The positive "noninverting" electrode was placed on the vertex at the midline, and the negative "inverting" electrode was placed on the mastoid. The ground electrode was placed on the opposite ear. It was ensured that the electrode impedance was kept to a minimum (<5 kΩ). The signal was delivered through insert earphones. The stimulus used was a click signal which was presented at various intensities to obtain the air conduction threshold of the ear. The threshold of hearing was taken as the lowest intensity at which a replicable wave V was obtained.
The ambient sound level was tested in the ward, the audiology room, and the ABR room with sound level meter, CENTER 322-Sound Level Meter (Data logger) (Center Technology Corp, Victoria, Australia).
The data of all the patients were collected systematically using the software EPIDATA version 3.1 (EpiData Association, Denmark). All statistical analyses were performed using the statistical software STATA version 10.0 (STATA Corporation, Texas, USA). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated.
| Results|| |
Of the 37 newborns screened during the 13-month period, 23 (62.2%) were males and 14 (37.8%) were females. Eight (21.6%) were born of consanguineous marriage and 29 (78.3%) of nonconsanguineous marriage. Among the 8 neonates born of consanguineous marriage, 5 (23.3%) neonates were bilaterally referred and 3 (19.4%) showed bilateral pass in the second BERAphone screening. Only one neonate was a confirmed case of bilateral hearing loss. From the 29 neonates born out of nonconsanguineous marriage, 23 (79.3%) were "referred" and 7 (24.1%) showed confirmed hearing loss.
Out of the 37 (100%) neonates in the study, 17 mothers were primigravida and 20 were multigravida. Eight neonates of primigravida mothers showed "refer" and 3 were confirmed to have hearing loss. Among the neonates of multiparous mothers, 19 were "referred" and 3 were confirmed to have hearing loss.
Gestational ages at birth of all the newborns recruited in this study were above 36 weeks of gestation, and hence, all the study newborns were not born preterm. Two mothers had associated maternal risk factors. One mother had both gestational diabetes and hypothyroidism and underwent delivery by cesarean section. The neonate of this mother showed bilateral "refer" in the second screening and hearing loss was confirmed in the left ear by ABR. One mother had only gestational diabetes. The second screening for this neonate showed bilateral "refer" and bilateral hearing loss was confirmed by ABR. There were 12 neonates born by cesarean section; 9 of these neonates were "referred" and 4 neonates had confirmed hearing loss.
Of the 37 neonates in the study, 3 had birth weight <2.5 kg and all 3 showed confirmed hearing loss. Two babies had neonatal jaundice and one had seizures; all 3 were "referred" on second screening and 1 neonate who had jaundice was confirmed to have hearing loss. One had bacterial infection but had no hearing loss.
The first screening consisted of all 37 neonates who were "referred" and recruited in the study which made up a total of 74 ears. Of these, 13 (35.10%) infants had bilateral "refer" and 24 (64.80%) neonates had unilateral "refer." All of them were subjected to the second BERAphone screening.
The 37 newborns who participated in the second stage of the screening process provided 74 ears for comparative testing against the ABR. Of the 74 ears of newborns referred for the second stage screening, 31 (41.9%) ears passed and 43 (58.1%) ears were referred again. Therefore, the pass rate for the second screening was 41.9% and the referral rate was 58.1%. At the second screening, 15 infants showed bilateral "refer," 13 infants showed unilateral "refer," and 9 infants showed bilateral "pass."
On confirmatory testing by ABR, of the 74 ears tested, 14 ears (18.9%) had confirmed hearing loss and 60 ears (81.1%) were normal which represents a percentage of confirmed hearing loss of 18.9% and a pass level of 81.1%. Of the 9 neonates with hearing loss, 5 neonates had bilateral impairment and 4 neonates had a unilateral impairment.
The purpose of the study was to calculate the specificity and sensitivity of the neonatal screening tool MB11 BERAphone® (MAICO Diagnostic, Germany) when used in the ward setting. [Table 1] shows the results of the second BERAphone screening when confirmed with ABR.
|Table 1: BERAphone second screening with auditory brainstem response confirmation|
Click here to view
The diagnostic results are shown in [Table 2]. Of the 28 neonates (43 ears) who failed the second stage screening, 9 neonates had confirmed hearing loss on ABR. Of these, 5 neonates had a bilateral impairment and 4 neonates had a unilateral impairment. The remaining 28 neonates were found to have bilaterally normal hearing. The MB11 BERAphone screening yielded 1 false negative (7.1%) and 30 false positive (50%). Efficacy of MB11 showed a sensitivity of 92.9% (95% confidence interval [CI]: 66.06% to 98.81%) and a specificity of 50% (95% CI: 36.81% to 63.19%). The PPV was 30.23% (95% CI: 17.20% to 46.13%) and NPV was 96.77% (95% CI: 83.24% to 99.46%) for the diagnosis of hearing loss.
|Table 2: Sensitivity, specificity, positive predictive value, and negative predictive value|
Click here to view
The observed prevalence of confirmed hearing impairment was 18.9%. The rate of unilateral impairment was 10.8%, and the rate of bilateral impairment was 13.5%. The average ambient noise levels in the postnatal ward setting was 62.1 dB.
| Discussion|| |
The average number of live babies born per month in this tertiary care setup during the period of study was 1107. Considering the large numbers, implementation of the "universal screening program" in the ideal setting (sound-treated room) was not possible; hence, the screening was undertaken in the postnatal ward setting. Studies by van Straaten  also showed that the screening test could be used in the ward setting without disturbance from ambient noise and during oxygen therapy.
Hearing screening immediately after birth often results in high false positives due to vernix plugs in the external auditory canals. Therefore, an effective way of minimizing over-referral due to false positives is to implement a two-stage screening protocol. , This protocol also improves the specificity. The two-step screening is a standard practice in western countries for the universal screening program.  Benito-Orejas et al.  followed a two-step screening protocol where the first screening was undertaken during the first 48 hours of life, before discharge from the hospital and the infants referred from the first screening underwent a second screening after discharge and before 1 month. They showed that aABR gave 2.6% failed results in the first screening, and in the second screening step, 0.32% of those screened with aABR were referred. They concluded that two-step screening with aABR had a significantly less referral rate than two-step screening with OAE and was therefore more cost-effective. Another study by Lin et al.  showed a significant reduction in referral rate using a two-step protocol. The referral rate of 5.8% in the first screening was reduced to 1.8% after the second screening. Yet, another study by Iley and Addis  concluded that the adoption of aABR as an initial screening is less expensive, more practical, and acceptable to the parents.
The second screening of infants who were "referred" in the first screening was done after a period of 1 week in a sound-treated room in the Department of Audiology with an ambient sound level of 44 dB. Our experience was that if there was any delay in the second screening, a large number of neonates were lost to follow-up. Mathur et al. suggest that screening all neonates within the first 48 hours of life is impractical and that it should be delayed up to 3 months of age.  Our study could complete the second screening as per the protocol.
The manufacturer reports that the "BERAphone" has a sensitivity of more than 0.99 and test specificity of 0.87 for a one-time test, and as a two-stage test, the specificity is greater than 0.96. Perhaps, such high values are only possible when tested under ideal conditions. In tertiary care centers such as ours, with more than 13,000 annual deliveries, testing under ideal conditions is not feasible. Testing had to hence be moved to the bedside in the ward for high coverage rates using the screening tool (MB11 BERAphone®).
Of the 74 newborn ears referred for the second stage screening, 31 (41.9%) passed and 43 (58.1%) were referred again in comparison with the study done in Germany by Cebulla and Shehata-Dieler,  where 27.4% passed on second screening and 72.6% were referred again.
All the infants who underwent the second screening were also subjected to the confirmation test by ABR, and 18.9% was the prevalence of confirmed hearing loss. The rate of unilateral impairment was 10.8%, and the rate of bilateral impairment was 13.5%. All the nine infants with confirmed hearing loss were placed on regular follow-up. This study shows a higher prevalence of confirmed hearing loss than that reported in the literature. This could be due to the fact that only newborns that were "referred" on the first screening were recruited for the study.
The MB11 BERAphone test yielded 1 false negative (7.1%) and 30 false positive (50%). The MB11 BERAphone as used in our study protocol showed a lower sensitivity and s pecificity than that claimed by the manufacturer for diagnosis of hearing loss. A study by Melagrana et al.  showed no false negatives and 10 false positives among 388 ears of infants who were evaluated after the second month of life but when done under ideal conditions and showed sensitivity of 100% (95% CI: 93.9-100%), specificity of 96.8% (95% CI: 94.8-98.7%), PPV of 88.2% (95% CI: 79-93.9%), and NPV of 100% (95% CI: 98.4-100%) for diagnosis of hearing loss. Stewart et al. showed that the refer rate using aABR in universal hearing screening was <2% and the false positive rate was between 0.3% and 2.5% although it was performed by various personnel in the nursery setting. 
Richmond et al. indicate that ambient noise levels can cause alterations in the screening results of the infants.  Most studies using different screening protocols have been done under quiet or sound treated conditions and of these only a few are from the developing countries. The uniqueness of our study is because, the screening tool was used in the ward with a high volume of deliveries and a high ambient noise, making it possible in a country like India to set up universal screening on a wider scale.
Good screening tests should have a high sensitivity and the two-stage screening with BERAphone has shown high sensitivity. Those tests with very high sensitivity are useful "to rule out" diseases or characteristics if the result is negative. On the other hand, good confirmatory tests have high specificity and two-stage screening with BERAphone has shown a low specificity. High specificity is useful "to rule in" diseases or characteristics if the result is positive. Both sensitivity and specificity are intrinsic properties of a given test which do not depend on the given population. PPV and NPV however depend on the prevalence rate of the characteristic in a given population. The rarer the characteristic (prevalence), the lower the PPV and the higher the NPV. The two-stage screening protocol with BERAphone has a low PPV and a high NPV. The prevalence of hearing loss in the study group was 18.9%. The BERAphone would have had a higher PPV had the prevalence been higher, i.e., in a situation with a higher prevalence (e.g., neonatology intensive care) and also if a third screening with BERAphone was done before confirmation with ABR.
| Conclusion|| |
Two-stage screening with the MAICO MB11 BERAphone results in a pass rate of 41.9% and a referral rate of 58.1%. The screening tool showed sensitivity of 92.86%, specificity of 50%, PPV of 30.23%, and NPV of 96.77% for the diagnosis of hearing loss when used in the ward setting with high ambient noise making it useful as a first level screening tool in a developing country. Its results must however be verified with ABR which is the gold standard test for the diagnosis of hearing loss among infants.
Financial support and sponsorship
The fund was provided by Fluid Research Grant from Christian Medical College, Vellore.
Conflicts of interest
The investigators declare that the study was performed independently and neither the authors nor the departments have received any monetary support in any form from any industry or other external source related to the material discussed in this manuscript.
| References|| |
Olusanya BO, Emokpae A, Renner JK, Wirz SL. Costs and performance of early hearing detection programmes in Lagos, Nigeria. Trans R Soc Trop Med Hyg 2009;103:179-86.
Kemper AR, Downs SM. A cost-effectiveness analysis of newborn hearing screening strategies. Arch Pediatr Adolesc Med 2000;154:484-8.
Gracey K. Current concepts in universal newborn hearing screening and early hearing detection and intervention programs. Adv Neonatal Care 2003;3:308-17.
McMahon E, Wintermark P, Lahav A. Auditory brain development in premature infants: The importance of early experience. Ann N Y Acad Sci 2012;1252:17-24.
Cebulla M, Shehata-Dieler W. ABR-based newborn hearing screening with MB11 BERAphone®
using an optimized chirp for acoustical stimulation. Int J Pediatr Otorhinolaryngol 2012;76:536-43.
van Straaten HL. Automated auditory brainstem response in neonatal hearing screening. Acta Paediatr Suppl 1999;88:76-9.
Melagrana A, Casale S, Calevo MG, Tarantino V. MB11 BERAphone and auditory brainstem response in newborns at audiologic risk: Comparison of results. Int J Pediatr Otorhinolaryngol 2007;71:1175-80.
Benito-Orejas JI, Ramírez B, Morais D, Almaraz A, Fernández-Calvo JL. Comparison of two-step transient evoked otoacoustic emissions (TEOAE) and automated auditory brainstem response (aABR) for universal newborn hearing screening programs. Int J Pediatr Otorhinolaryngol 2008;72:1193-201.
Lin HC, Shu MT, Lee KS, Ho GM, Fu TY, Bruna S, et al.
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 2005;115:1957-62.
Iley KL, Addis RJ. Impact of technology choice on service provision for universal newborn hearing screening within a busy district hospital. J Perinatol 2000;20 (8 Pt 2):S122-7.
Mathur NN, Dhawan R. 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. J Laryngol Otol 2007;121:639-43.
Stewart DL, Mehl A, Hall JW 3 rd
, Thomson V, Carroll M, Hamlett J. Universal newborn hearing screening with automated auditory brainstem response: A multisite investigation. J Perinatol 2000;20 (8 Pt 2):S128-31.
Richmond KH, Konkle DF, Potsic WP. ABR screening of high-risk infants: Effects of ambient noise in the neonatal nursery. Otolaryngol Head Neck Surg 1986;94:552-60.
[Figure 1], [Figure 2]
[Table 1], [Table 2]