Archives ofDisease in Childhood 1990; 65: 479-485

Early identification of hearing loss: screening andsurveillance methods

P E Scanlon, J M Bamford

AbstractService monitoring data on the outcomes ofhealth visitors' screening for hearing loss at 8months in West Berkshire indicate low sensi-tivity and low positive predictive value,despite efforts to improve the conduct of thescreen. Nevertheless, data on a recent seriesof severely hearing impaired children indicatesignificantly earlier diagnosis than previously,due in part in the introduction of other servicechanges including neonatal 'at risk' screeningand surveillance using parental observation.For a trial period the traditional screeningmethod for the detection of hearing loss inbabies will be discontinued and effort concen-trated on these alternative procedures.

West BerkshireHealth Authority,Department ofCommunity ChildHealth, ReadingP E ScanlonAudiology Unit,Royal BerkshireHospital, ReadingJ M BamfordCorrespondence to:Dr P E Scanlon,Community Services Unit,3 Craven Road,Reading,Berkshire RG1 SLF.Accepted 17 November 1989

It is widely accepted that congenital sensori-neural hearing impairment can have appreciableeffects on the speech, language, social, psycho-logical, and educational development of theimpaired child. It is also widely accepted thatthe disability associated with severe congenitalhearing impairment can be reduced signifi-cantly by early identification of the conditionleading to appropriate intervention and rehabili-tation. That there is little direct evidence of thisis not necessarily reason to doubt it.' How earlythe identification should occur is not certain,but from a purely theoretical point of view itshould probably occur as soon after birth as pos-

sible, at least for the more severely impairedcases.

In the United Kingdom there is an estab-lished nationwide hearing screening test,based on the distraction test first described bythe Ewings,2 and performed by health visitorswith children aged 7-9 months. This and otheraspects of the service aimed at early identifica-tion achieved a peak of notoriety in the late1970s partly as a result of surveys, the results ofwhich questioned the coverage and sensitivityof the test.3 Efforts before and since thenin a number of districts, most notablyNottingham,4 5 seem to show that given theappropriate resources for training, equipment,and referral, the health visitors' screening testcan be implemented relatively effectively andcertainly more effectively than previously.

Experience in West BerkshireIn the early to mid 1980s, in response to theunacceptably low sensitivity of the health visi-tors' screen, improvements in training, tech-niques, and equipment for the health visitors inWest Berkshire health district were introduced

after the lead given by the Nottingham services.Previously, screening by the distraction methodhad used sounds produced by the Nuffield highfrequency rattle and voice, to provide a low fre-quency 'oo' and a high frequency 's' at approxi-mately 35 dB. During the training days twice ayear for newly appointed health visitors in WestBerkshire, conducted by the senior clinicalmedical officer responsible for audiology, therewas strong emphasis on the need for the soundsto be produced at this intensity, in the correctmanner, and without visual clues and for theneed for stringent self criticism in the conductof each test. It was difficult to maintain thesestandards when sound level meters were notalways available during testing sessions. Thehearing test was passed if the baby responded toall the sounds presented by a full turning of thehead to localise the source of the sound. Failureto do so indicated a need for a retest three tofour weeks later and, if there was further failureat this stage, referral to the community audio-logy clinic for audiometric assessment and oto-logical examination by clinical medical officersexperienced in audiology or to the audiologyunit at the Royal Berkshire Hospital.To bring the screening of hearing by health

visitors to its 'realistic best' modifications weremade in the method of testing, the criteria forreferral, and the training programme.

Calibrated electroacoustic warblers wereintroduced throughout the district. These pro-vided warble tones at 0 5 KHz, 2 KHz, and4 KHz and they were offered at 35 dB soundpressure level to both ears. Two opportunitieswere provided at each frequency to both ears for2-3 seconds on each occasion. Localisation ofthe sound source by a full turning of the head onone occasion out of two was rated a pass. Failureto respond to any frequency for either ear indi-cated the need for a retest three to four weekslater. If failure of the retest occurred, referral tothe community audiology clinic or hospitalaudiology unit was mandatory. The number oftraining days was increased to four per year,each involving both authors so that the scope ofthe training increased to include lectures on thenature and significance of hearing loss in babiesas well as demonstration of the testing methodand practice of the technique to ensure a highlevel of competence in testing by the health visi-tors. Not only newly appointed health visitorsattended these training days but also those longestablished in the district attended for refreshercourses on a rotational basis.There remained a real problem, however,

with the timing of the screen as a first line ofattack for identification of severe or very severe

479

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from

Scanlon, Bamford

hearing loss. For developmental reasons, thetest cannot be implemented before about 6-7months of age and even with an assessmentservice that could offer immediate appoint-ments for suspected severe cases, the age atwhich hearing aids could be fitted would not bebefore about 9 months at very best and morelikely 12 months (allowing time for retest,referral and differential diagnosis, counselling,and manufacture of earmoulds). While thismight have seemed a service triumph a decadeor two ago, we are now aware of the importanceof the early stages of language development andthe consensus has moved towards rehabilitativeintervention within the first few months ofbirth, at least for severe cases. In view of theseconsiderations, other efforts were made withinWest Berkshire to secure earlier diagnosis thancould be offered by the health visitors' screen-ing route even at its realistic best. In particular:

(1) Screening of 'at risk' neonates was intro-duced using initially the auditory responsecradle,6 and later either or both auditory brain-stem response screening and otoacoustic emis-sion screening. The case for 'at risk' screening(mainly 'graduates' of special care baby units)has been documented and reviewed elsewhere.7

(2) Parental observation, at any age, nomatter how early, was used as a guide to rapidreferral. In particular the 'hints for parents'from a paper by McCormick was introducedand issued to all parents by the health visitor atthe earliest opportunity, usually at the 10th dayvisit.'

(3) A concerted effort was made to encouragehealth visitors and general practitioners to referchildren as early as possible whenever hearingwas in doubt. A programme of awareness ofhearing impairment and its consequences wasimplemented for all professionals, includinghealth visitors, general practitioners, clinicalmedical officers, and paediatricians.Thus from 1984, the services in West Berk-

shire offered neonatal screening for 'at risk'babies, heightened awareness, explicit guide-lines for parental observation, and the healthvisitors' screening test performed near its'realistic best' (given the inevitable constraintsof staff turnover, limited resources, etc). Theeffects of these service changes within WestBerkshire are illustrated by data on two series ofseverely or very severely hearing impaired chil-dren, one from 1974 to 1977 the other from1984 to 1988. Table 1 shows the age of issue ofhearing aids to children in these two series.Clearly there is a noticeable improvement in the1984-8 series in that age of identification anddiagnosis (and therefore age of hearing aid

Table I Age of issue of hearing aid

Age (months) 1974-7 1984-8

0-6 - 47-12 - 313-18 3 819-24 3 425-30 10 131-36 - ->3 Years 3 2

Total 19 22

Table 2 Children from 1984-8 series: referral routes(n=22)

Routine No ofchildren

Neonatal screening to full assessment 4Parental concern to general practitioner or health

visitor to full assessment 7Paediatricians to full assessment 5Health visitors' screen followed by accelerated

referral because of concern 5Health visitors' screen to full assessment via

second tier community clinic

fitting) is reduced. The total numbers ofseverely impaired children (19 and 22) are inline with an overall (that is, including mild andmoderate degrees of hearing loss) prevalencerate of between 1-2/1000, given the birth rate of6000/year in West Berkshire.The routes by which the 22 children in the

later period were identified as severely hearingimpaired are shown in table 2. The full audio-logical assessments were carried out in a tertiarylevel audiology unit based in a district generalhospital with ear, nose, and throat opinionavailable and hearing aid provision adjacent.These organisational details are incidental to

the particular topic of this paper. A prerequisiteof any good programme of audiological screen-ing or surveillance, however, is a competent,adequately resourced audiology department,capable of accurate confirmation with minimaldelay and of contributing to the multifacetedsupport required for aural rehabilitation inchildhood.As a result of the service changes imple-

mented in West Berkshire in 1984, the tradi-tional screening by health visitors had becomemore of a safety net than a first line of attack forcongenital severe sensorineural hearing loss.Indentification of the hearing loss was beingmade earlier than could be expected from thehealth visitors' screen, and therefore amplifica-tion was also being provided earlier. Table 2indicates that babies with suspected sensori-neural hearing loss reached the full audiologicalassessment stage by a number of routes, thetraditional screening test being but one of them.

Data are also available on the effectiveness ofthe health visitors' screen itself during the sameperiod of 1984-8. Unfortunately, given theefforts put into improved protocols, improvedtraining, and the use of calibrated electroacous-tic warblers as sound sources, the detection ratefor moderate, severe, and very severe sensor-ineural hearing impairment still seems to berather low. Of 12 such children confirmed ashearing impaired between 1984 and 1988 (thosealready identified or referred were not screenedby health visitors), six had passed the screen. Itis probable, though not certain, that the hearinglosses had been present since or soon after birth.A sensitivity as low as 50% does not necessarilymean that a screen should not be undertaken; itdepends upon cost, alternatives and the conse-quences of the condition going undetected. Inthe case of infant hearing screening, where afalse negative will introduce undue delay andconfusion into the processes leading to eventualconfirmation, it is generally accepted that a

480

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from

Early identiftcation ofhearing loss: screening and surveillance methods

screening system needs to have a higher sensi-tivity to be acceptable. The issue reduces towhether the contribution of the health visitors'screen to the total sensitivity of the combinedset of surveillance arrangements is sufficient tojustify its cost.

Quite why the detection rate for severe casesremains poor is not clear. Other centres suggestbetter results from the improved screening byhealth visitors',5 although direct measures ofdetection rate are rather rare in the literature.

In West Berkshire, to implement the screen-ing programme requires the successful trainingof 130 health visitors to carry out at least 6000hearing tests each year. In view of the relativelyhigh staff turnover the training, including back-ground lectures, demonstrations, practice oftraining techniques, and discussion probablyrequires more than the eight days a year (fourfrom each of the authors) which have been avail-able.

It has been suggested that decision theory andvigilance theory might help to explain the lowsensitivity: the search for one or two cases/thousand is a search for a relatively rare condi-tion. Hence health visitors' criteria could not beexpected (given an imperfectly specified testprocedure) to be optimal.

This could only be a partial explanation,however. Although the eight month screen isorientated towards detection of severe congen-tial losses,7 it also refers many times more chil-dren with milder hearing losses associated withotitis media with effusion.5 Valid referral ofscreening is, therefore, not nearly as rare aswould be the case if only sensorineural hearingloss existed. A further possibility for the poordetection rate of the improved screen may con-cern the nature of the task, which by necessityfor a screen is an explicit, rigid, and quantitativetest procedure, not open to post hoc discretion-ary interpretation.

Health visiting and the community childhealth services have always emphasised surveil-lance and good clinical judgment, taking intoaccount the whole child, family background,and parental observation. In many ways a rigidnon-interpretative screening test concernedwith just one discrete sensory function does notsit easily within the broader role of health visi-tors, especially as the task of assessing a 7months old's responses to sound involves (ashealth visitors are fully aware) a complex set oflargely unquantified and unquantifiable vari-ables. Haggard used the data on children withotitis media with effusion to examine the effi-cacy of the health visitors' screen.5 He arguesthat what applies to the less severe but morecommon hearing losses associated with otitismedia with effusion provides the only practic-able statistical reflection of what will apply todetection of sensorineural hearing losses. Asotitis media with effusion is a fluctuating condi-tion, the usual counting of false negatives(passed screen but with hearing loss at time ofscreen) is not possible. Haggard thereforesuggests the use of two variables by whichchanges in the screen may be monitored,namely failure rate (FR) and positive predictivevalue (PPV). Failure rate is given by:

True positives + false positivesFR=Total

and positive predictive value by:

True positives + false positivesAs otitis media with effusion and any associ-

ated hearing loss can fluctuate so rapidly, theremust still be some doubt about the true andfalse positive rates unless a full assessment (pre-sumed accurate) takes place on the same day asthe screen. Provided the average time betweenscreen and full assessment is static and reason-ably short, however, this source of error canprobably be ignored.

Table 3 shows the failure rate and positivepredictive value for the health visitors' screen inWest Berkshire for 1984-8. Data for 1984 and1985 cover the whole district (birth rate 6000/year), while the data for 1986, 1987, and 1988are for clinics covering two thirds of the districtpopulation (assumed birth rate 4000/year). Allroutine screening failures are referred first tointermediate community health service clinicsstaffed by experienced clinical medical officers.

Cases thought to be urgent or possible sensor-ineural losses may bypass this tier. The com-munity clinics may refer to the audiology unit atthe Royal Berkshire Hospital for ear, nose, andthroat and audiological assessment (usuallyadjacent) or to the general practitioner for treat-ment, they may advise and review, or they maydischarge. A few cases that have failed thehealth visitors' screen may also be referreddirectly to the audiology unit by the generalpractitioner or health visitor. Therefore, thefigures in table 3 are contaminated by somesources of error, but these are likely to be small.A number of points emerge from table 3.

First, compared with 1984 and 1985, the failurerate has increased for recent years. This mayreflect the gradual effects of the changes intro-duced in 1984 to the screening protocol, inparticular the stricter criteria for a pass and theuse of frequency specific electroacousticwarblers. It has been reported that babies aremore responsive to warble tones than to thetraditional wider band noise makers,9 but thereare theoretical and empirical grounds forexpecting narrower band stimuli to elicit fewerresponses.'0 Furthermore, a questionnaire tohealth visitors within the district also supportedthis view that the babies were less responsive to

Table 3 Number ofcases referred by health visitors' screenin West Berkshire to the second tier clinics showing thosedischarged after only one appointment, those referred on to thethird tier at Royal Berkshire Hospital, and hencefailure rateand positive predictive value. Those neither discharged norreferred on were kept on review, and have been included inthe calculation ofpositive predictive value (as true positives)Year No No (%) No (%) Failure Positive

referred discharged referred rate (%) predictivevalue (%)

1984 183 110 (60) 11 (6) 3 051985 156 81 (52) 22 (14) 2 601986 152 88 (58) 29 (19) 3 80 42-11987 177 115 (65) 23 (13) 4-43 35 61988 264 169 (64) 42 (16) 6-60 35 9

"No data available.

True positives

481

PPV=

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from

Scanlon, Bamford

warblers: 75% of respondents claimed thatbabies were 'slightly' or 'much less' responsiveto warblers compared with their responsivenessto high frequency rattles for example.The failure rate in West Berkshire in the

more recent years is double that reported byHaggard and Gannon who examined a similarscreening system that utilises strict protocols,warblers, and appropriate training." This mayreflect the stricter criteria for screening failurein our protocol but it is difficult to be sure with-out a closer examination of the fine details ofboth protocols for testing and training.

Second, the positive predictive value associ-ated with the screen is low and much less thanthat of 78-6% quoted by Haggard for theNottingham services.5 This difference is likelyto be due largely to the (unacceptably) longwaiting time of about six months that hasexisted in West Berkshire for some years for thesecond tier community clinics. As otitis mediawith effusion fluctuates it seems that many ofthe cases had improved spontaneously by thetime they were seen in the second tier clinic,giving very high discharge rates. The timebetween screening and referral (at least for non-urgent cases) in the Nottingham services wasconsiderably shorter and is commonly only sixto eight weeks (B McCormick, personal com-munication). The data from table 3, and exami-nation of the outcomes of those cases referred tothe Royal Berkshire Hospital do incidentallyemphasise the fact that the health visitors'screen was not designed for and is not an effi-cient way of detecting those children with otitismedia with effusion of such severity and/or per-sistence to justify surgical intervention.Although the screening does detect over 30times more children with otitis media with effu-sion than it detects cases of severe sensorineuralloss,7 the number of surgical referrals of otitis

Table 4 Outcomes for the cases in table 3 (excludingsenorineural losses) referred to Royal Berkshire Hospital in1986-8

Year No No who had Advice, review, Notreferred surgical andlor traced

intervention discharge

1986 27 14 8 51987 20 10 6 41988 42 18 21 3

media with effusion requiring surgical treat-ment is really rather small. Table 4 shows thesurgical and therapeutic outcomes for the casesreferred (in table 3) to the Royal BerkshireHospital. Of course, the ear, nose, and throatand the audiology services at the hospital weredealing with a much larger group of childrenwith otitis media with effusion during theseyears and with a surgical intervention rateapproaching 1000 operations/year on childrenbelow the age of 8 years (and mostly below 5years). The referral route for these cases tendsto be parental concern about illness, hearingloss, or speech/language development, togeneral practitioner, and to the ear, nose, andthroat and audiology departments at thehospital.

A different approachFrom 1 January 1989, initially for a trial periodof two years, it has been decided to discontinuetraditional hearing screening tests by healthvisitors at 7-8 months of age. In its place, asystem of surveillance has been introduced thatwill involve parents and professionals in accord-ance with the principles being recommended fordevelopmental surveillance by the Joint Work-ing Party on Child Health Surveillance.'2 Forthe period of the trial: (1) parental observationand clinical concern will determine immediatereferral at any age (however young the baby).(2) Neonatal screening of 'at risk' babies willcontinue. (3) Parental awareness about theimportance of hearing and guidance for parentalobservations will be supported by the Hints ForParents leaflet introduced at the initial visit bythe health visitor.8 (4) The hearing of everybaby will be considered in depth at a convenienttime between 6 and 8 months of age, usually inassociation with the developmental surveillanceprogramme being carried out at that age. Toensure that the appropriate information iselicited, a questionnaire (see table 5) will becompleted by each health visitor and forwardedto the senior clinical medical officer, accom-panied by a referral for a hearing test in thecommunity clinic if appropriate.

Professional awareness of the importance andrelevance of hearing will be maintained amongall health visitors by study days during which

Table S West Berkshire health authority-community health services questionnaire about hearing for 7-8 month old babies

Name ofbaby: .......................................................... Dat e of birth: .

Address: ......................................................... General practitioner:.

.................................................................................................... H a t vi to :......................................................Hat iio:

Home telephone No: . .........................................................CaseloadNo: .

Date of completion of questionnaire: .............................................................................................................................

Is there any history of deafness in young people in the family?Was there any possibility of maternal rubella during the pregnancy?Were there any problems associated with the birth?Did the baby have to go into the special care baby unit?Has the baby had frequent colds or ear infections?Do loud noises make the baby jump?Does the baby respond to some quiet sounds?Does the baby anticipate a person approaching without visual clues?Is double babble present?Are there any concerns about the baby's general development?Is there ANY parental concern about the baby's hearing?Has a referral been completed?

Yes/noYes/noYes/noYes/noYes/noYes/noYes/noYes/noYes/noYes/noYes/noYes/no

23456789101112

482

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from

Early identification ofhearing loss: screening and surveillance methods

Table 6 Advisory leaflet for parents. Children with mild hearing losses: how to help

1 Speak clearly. Look at your child when you are talking.2 Try to keep background noises down: turn the TV off, tell others to be quiet while you are talking with your child.3 Make the 'little bits' of a conversation clear-they are just as important as the rest.4 Sometimes your child will nod, or smile, or say 'yes' even when he/she has not heard you.5 Try to be patient and understanding, it's no fun not hearing well.Remember-your child's hearing may be good one day, bad the next.

their knowledge of hearing loss and its implica-tions will be extended. Opportunities will alsobe provided for discussion and exchange ofideas about the conduct of the programme.

General practitioners and paediatricians willalso be kept informed about the progress of thetrial and reinforcements given about the import-ance of early referral in cases of suspected hear-ing loss. The aim of this programme is toprovide a responsive service with waiting timesfor appointments reduced to a minimum. It ishoped to see all cases of possible sensorineuralhearing loss within a few days and cases ofpossible conductive hearing loss associated withmiddle ear effusion within a month or two ofreferral.

If there is any parental concern about hearingloss, it is recognised that there may be con-

tinuing anxiety while waiting for an appoint-ment. In order to help advise parents of sensiblemanagement, a simple advisory leaflet has beenprepared that can be given to parents of chil-dren who have been referred through the non-

urgent route (see table 6).When the trial system is fully implemented, it

is expected that:(1) With 90% or more coverage for 'at risk'

neonatal screening, about half of all the cases

with sensorineural hearing loss will be identifiedby this screen.7

(2) The programme of surveillance by healthvisitors, general practitioners, and paediatri-cians and responses to parental observation willhave a detection rate at least as good as the pasthealth visitors' screen for the remaining severe

or profound congenital cases. A trial period oflonger than two years, however, will berequired to verify this with any certainty.

(3) With appropriate support and training,health visitors will refer via general practitioneror community audiology clinics a set of severeand/or persistent cases of otitis media with effu-sion with referral rates and a positive predictivevalue more appropriate than those estimated forthe traditional hearing screen. The effect of thisshould be to reduce the long waiting lists thatcurrently exist in the community clinics, whichare unnecessary (in that most cases are dis-charged) and which delay the referral of more

appropriate or more urgent cases.

The training programme will no longer beconstrained by the need to teach the techniqueof distraction testing. The emphasis will be on

the development of communication and therelevance of normal and abnormal hearing tothe acquisition of these skills. Video recordingsof patterns of communication between parentsof normal and hearing impaired children haveproved to be very useful teaching aids, and con-

siderable use will be made of such recordings.It may be argued that the absence of a fre-

quency specific hearing test will fail to identify

mild, U shaped, and high frequency hearingloss or moderate losses with recruitment until achild's delay in language acquisition raises sus-picion of a significant loss. The question ofwhether a surveillance based system will besharp enough to identify babies who havemoderate or even severe hearing loss combinedwith very considerable recruitment will beanswered empirically. The authors think thatwith suitable training and support for healthvisitors these cases will be identified earlyenough; however, doubt must remain until theresults of the trial period are available. The casefor very early intervention with mild, U shaped,and high frequency hearing loss have yet to bemade convincingly and we argue, therefore,that any programme for early diagnosis shouldbe primarily directed towards the more severelyimpaired at this stage. The possible case forearly intervention with children with mild sen-sorineural hearing losses should not, in ourview, be confused with the more obvious casesof intervention with mild fluctuating middle earhearing losses associated with chronic otitismedia with effusion. With the latter can go dis-comfort, illness, poor auditory attention, anddifficult behaviour and it may be that togetherthese give rise to a greater disability than doesthe static mild sensorineural hearing loss.'3

Financial considerationsIt is thought that the changes that are beingintroduced will not effectively cut the cost of theservice. Although it will not be necessary to pro-vide a tester and a distractor for each hearingtest, the time taken for discussion of hearingand communication is time consuming. In addi-tion, production costs of leaflets, question-naires, videos, and information sheets forgeneral practitioners and health visitors all addto the financial commitment.

Furthermore, a surveillance based scheme,coupled with neonatal screening, should only becontemplated: (a) on a trial basis with effectivemonitoring and data collection and (b) in dis-tricts offering a good paediatric audiologyservice.

ConclusionEfforts have been made in West Berkshirehealth district to secure identification anddiagnosis of severe congenital sensorineuralhearing loss as early as possible. Neonatalscreening, use of parental observation, ques-tionnaires, better training for health visitors,general practitioners, and paediatricians, moreresponsive services and the 7 months screeningtest by health visitors improved to its 'realisticbest' have been introduced. Data indicate thatthe screening by health visitors remains a pro-

483

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from

484 Scanlon, Bamford

blem, with low detection rate for sensorineuralcases and over referral of cases with transientotitis media with effusion. For an initial trialperiod of two years, the screen is to be discon-tinued and a programme of surveillance intro-duced to secure a more efficient system. Thedata will be monitored and published in duecourse.

The authors wish to thank Barbara Dubois for data collection andanalyses, particularly the data summarised in tables 3 and 4.

1 Markides A. Age at fitting of hearing aids and speech intelli-gibility. Br J Audiol 1986;20:165-8.

2 Ewing IR. Screening tests and guidance clinics for babies andyoung children. In: Ewing AWG, ed. Educational guidanceand the deaf child. Manchester: Manchester UniversityPress, 1957.

3 Martin JAM, Bentzen 0, Colley JRT, et al. Childhood deaf-ness in the European community. Scand Audiol 1981;10:165-74.

4 McCormick B, Wood SA, Cope Y, et al. Analysis of recordsfrom an open access audiology service. BrJ Audiol 1984;18:127-32.

5 Haggard MP. Monitoring the efficiency of hearing screens forthe first year of life. Audiology in Practice 1986;3:3-5.

6 Bennett MJ. Trials with the auditory response cradle:I-neonatal response to auditory stimuli. BrJ Audiol 1979;13:125-34.

7 Sancho J, Hughes E, Davis A, et al. Epidemiological basis forscreening hearing. In: McCormick B, ed. Paediatricaudiology 0-5 years. London: Taylor and Francis,1988:1-35.

8 McCormick B. Health screening by health visitors: a criticalappraisal of the distraction test. Health Visitor 1983;56:449-51.

9 McCormick B. Evaluation of a warbler in hearing screeningtests. Health Visitor 1986;59:143-4.

10 Bench J, Mentz L. Stimulus complexity. State and infant'sauditory behavioural responses. British Journal of Com-munication Disorders 1975;10:52-60.

11 Haggard MP, Gannon MM. Analyses from service records ofscreening systems for hearing impairment in pre-schoolchildren. Medical Research CouncillInstitute of HearingResearch internal report series A. No 3. London: MRC, 1985.

12 Hail DMB, ed. Health for all children: a programme for childhealth surveillance. Oxford: Oxford University Press, 1989.

13 Bamford JM, Saunders E. Hearing impairment, auditoryperception and language disability. London: Edward Arnold,1985.

CommentaryThe appearance of this paper is very timely.Reports of studies demonstrating successfuloutcome from the use of the distraction test in ascreening context are unfortunately very rare,' 2and it is more common to see adverse reports.3This situation reflects the extreme variability inthe standard of application of the test techniqueand in the quality of services. There isabsolutely no doubt that a properly performeddistraction test can provide an effectivemethod for screening hearing and Scanlon andBamford acknowledge this.The question at issue is whether it is worth-

while allocating the necessary resources toachieve this good level of performance orwhether it might be more appropriate toredirect these resources in an attempt to identifyimpairment at an earlier age than is presentlyachieved. It is their hope that their alternativeapproach may detect deafness earlier and reducethe requirement to follow up cases with inter-mittent or less significant degrees of hearing lossin the first year of life.

Scanlon and Bamford are taking a bold stepwith their revised system, but the reader mustbe under no misapprehension about theresources required to achieve their alternativeservice. It is first necessary to establish a fullneonatal hearing screening test programme witha back up diagnostic audiology service for

babies of a few weeks of age: this should beavailable in all districts already but unfor-tunately it is not and it is confined to only a few.The next requirement is to train health visitorsin hearing surveillance and to develop anongoing support service to ensure that such avigilant service can continue to operate. Scanlonand Bamford stress that this requires new formsof training and a reallocation of resources. Theyare not talking about a cost cutting exercise butrather additional resources to establish thesetwo basic introductions to their service beforeabandoning the distraction test for a trialperiod.While respecting the need for the approach

taken by Scanlon and Bamford there must be aconcern as to whether they will acquire suffi-cient data within the time scale allocated. Therewill only be 10-12 000 births over that period intheir district giving a yield of only six to eightseverely or profoundly deaf children and it willbe necessary to allow a period of three or fouryears to elapse after their two year trial period tosee the appearance of any late detected casesmissed by their method. Serious considerationwill have to be given as to what they should doduring that period given the evidence alreadyavailable to show how effective the distractiontest can be (given appropriate training andcorrect technique). The questions that reallyshould be addressed before their study is under-taken are why do they not achieve better successalready, why are their standards not at the levelexpected and achieved by others, and why dothey not allocate resources to further improvethe distraction test?From the data they present there is clear evi-

dence that their current level of performance ofthe distraction test is poor in terms of its sensi-tivity, positive predictive value, and referralrate and all of these values are unacceptable andfall far short of those reported by Haggard andGannon for the Nottingham service.4 It may beinstructive here to quote the results from theNottingham service where every effort has beenmade over the years to refine and improve thedistraction test in addition to incorporatingneonatal hearing screening for at risk babies andintroducing a surveillance method utilising thehints for parents form 'Can your baby hearyou,. 6

Included in the table are the details of theinitial factor leading to referral for babies/children with later confirmed congenital orneonatally acquired severe and profound hear-ing loss. The babies were born over the threeyear period 1984-6 from a total birth populationof approximately 36 000. Babies born in morerecent years have not been included because anylate detected cases might not have emerged anda true picture would not be portrayed. Caseswith mild, moderate, and acquired hearinglosses have not been included.

Despite the availability of neonatal screening(for at risk cases) and a parent check listsurveillance system the distraction test resultwas the singly most important factor leading toreferral of severely and profoundly deaf chil-dren requiring hearing aids.

This method of analysis is useful because it

on August 31, 2020 by guest. P

rotected by copyright.http://adc.bm

j.com/

Arch D

is Child: first published as 10.1136/adc.65.5.479 on 1 M

ay 1990. Dow

nloaded from