Preschool hearing, speech, language and vision ... - University of York

APRIL 1998 VOLUME 4 NUMBER 2 ISSN: 0965-0288EffectiveHealth CareBulletin onthe effectivenessof health serviceinterventions fordecision makersNHS Centre for Reviewsand Dissemination,University of YorkPre-school hearing,speech, language andvision screening■ Child health screening focuseson the early detection ofchildhood disorders in orderto reduce disability. Thisincludes screening for hearing,speech and language andvision problems in pre-schoolchildren.■ Screening for permanentchildhood hearing impairmentis usually carried out byhealth visitors using the infantdistraction test (HVDT) at 6–9months. This test fails todetect a significant number ofhearing problems sufficientlyearly.■ There is good evidence thatuniversal neonatal hearingscreening is more effectiveand cost-effective than HVDTat detecting significantcongenital hearing loss.■ A significant number of preschoolchildren show signs ofspeech and language delay atsome point. Delay whichwould not spontaneouslyimprove can be effectivelytreated. However, it is not yetclear how to identify childrenwho will fail to progresswithout treatment.■ Pre-school vision screening forrefractive errors and nonobvioussquints may not bebeneficial because theseminor conditions, bythemselves, do not appear tobe problematic and becausethe advantages of treatingthem have not beendemonstrated. Major defectsare identified withoutscreening.■ The direct research evidencethat amblyopia in youngchildren is disabling and canbe effectively treated is weak.However, there is strongclinical opinion thatidentifying and treatingamblyopia at an early ageproduces benefits. Betterresearch is required to providea clearer picture of the likelybenefits, harms and costs ofpre-school vision screening.■ NHS organisations may wishto review their hearingscreening in the light of thisevidence but further change isnot recommended in advanceof the advice from theNational Screening Committeeand/or further research.The contents of this bulletin are likely to be valid for around one year, by which time significant new research evidence may have become available.

A. Pre-schoolscreeningChild health surveillance (CHS) ispart of a broad set of activitiesinitiated and provided byprofessionals. The objective is toreduce childhood disability byidentifying and managing amultiplicity of conditions at anearly stage and by working withfamilies to promote child healthand well-being. 1 CHS includes anumber of screening programmeswhich are focused on thedetection of specific disorders.CHS began as a series of activitieswhich tended to come together inan ad hoc manner. The value ofsurveillance and monitoring ofchild health, growth anddevelopment used to be regardedas self-evident. The Hall reportsemphasised the importance ofapplying rigorous criteria forscreening programmes incommunity child health andhelped to produce a more coordinatednational programme. 2–4However, there is still significantvariation both within and betweenhealth authorities in the content,timing and delivery of CHS.This bulletin summarises theresearch evidence about hearing,speech and language and visionscreening and is based on recentsystematic reviews commissionedby the NHS Health TechnologyAssessment Programme. Detailsof the methods and the results areavailable in the full reports. 5–7 Thisbulletin aims to provide easyaccess to this research intelligence,to promote informed debate andhelp decision makers.These three reviews focus on onlysome of the components of CHSand, within those parts, have notnecessarily examined all thedisorders being sought throughscreening. Some of the disabilitiesfor which children are screened inthe early years can exist alongsideone another and with otherswhich have not been reviewed,such as conductive hearing lossand developmental disability.Some important componentsremain to be scrutinised as doesthe value of the programme in itsentirety. Therefore, changes tocomponent parts of theprogramme should only be madeafter consideration has been givento the consequences of thesechanges to the programme overall.The Department of Health’sNational Screening Committee hasrecently established a sub-group toexamine neonatal and childhoodscreening and surveillance. Thiswill be making recommendationson the direction of futurescreening and surveillance. TheDepartment of Healthrecommends that no changes aremade to existing child healthscreening programmes until thiscommittee has reported and/orfurther research has beencompleted and considered.B. Evaluationof screeningThe objective of universalscreening in childhood is toidentify impairments which arenot obvious or apparent, whichwill cause significant disabilityor handicap and for which earlytreatment is more effective.Screening does not includeTable 1 Major criteria for assessing a screening programmesituations in which potentialproblems are noticed and thenreferred for detailed evaluations.Screening uses considerableresources, and imposes tests onchildren who are not ill. Inaddition, it has been argued thatsome screening programmes couldbe potentially harmful due to theunnecessary worry, referrals andprocedures which may result.There is, therefore, an ethicalresponsibility to ensure thatscreening is only carried out wherethere is confidence that it willresult in more good than harm. ‘Itis unethical to offer screening testswhich cannot stand up to criticalexamination’. 4 A number of criteriaare helpful when consideringwhether or not to carry outscreening (Table 1). 8,9This rational approach toscreening is at odds withconventional views held by somepractising clinicians and parentsthat any disorders should bedetected early if possible. Theseoften powerfully held views do notjustify the establishment ormaintenance of a screeningprogramme which is notsupported by the researchevidence. Ultimately, policydecisions have to be made byintegrating the available evidence,incomplete though it may be, andassessing the relative values ofdifferent opportunities to improvehealth. This needs to take accountDoes the screening programme do more good than harmand at acceptable cost?• Is the impairment sufficiently common to justify screening all children?• Does the impairment cause significant disability or handicap?• Is there agreement about what is meant by a case?• Is there a screening test which accurately identifies children who may have an impairment?• Is there an agreed and available effective intervention with which to treat the impairment orreduce the disability after identification?• Is there an advantage in detecting and/or treating the impairment earlier, before it becomesclinically observable?• Is the cost of screening justified by the net benefit?2 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screening APRIL 1998

of the fact that screeningprogrammes can also provideopportunities for healthprofessionals to undertake otheractivities, for example, healthpromotion. The removal of specificactivities may unwittingly result inan overall loss of service withoutnecessarily resulting in acorresponding time or cost saving.In the final analysis, policy may bedriven by social or politicalconcerns, informed by thescientific evidence, and individualchoices made by parents afterhaving been adequately informedabout likely benefits and harms.Even when there is good evidenceto support a screening activity itsbenefits are only fully realised ifthe programme is well managedand there is quality assurance.This may be difficult since manychildhood screening programmesnow involve several NHS trustsand professional groups and arealso undertaken in GP settings.Routine data collection is neededto monitor programmes in termsof coverage, sensitivity andspecificity and the progress ofthose children screened positive –a requirement rarely met atpresent.C. HearingscreeningC.1 Epidemiology and naturalhistory of congenital hearingimpairment: Approximately 840congenitally hearing-impairedchildren are born in the UK eachyear (1.12 per 1000 live births)with a permanent bilateralmoderate, severe or profoundhearing impairment of > 40dBhearing loss in the better ear. 10There are two types of hearingimpairment: sensorineural hearingloss (SNHL) caused by lesions inthe cochlea or auditory nerve andits central connections (unilateralor bilateral) and conductive hearingloss due to pathology of the middleear e.g. glue ear. The vast majorityof permanent childhood hearingimpairment is sensorineural,which does not resolve.Almost 85% of all permanentchildhood hearing impairment willbe present at birth with around160 cases a year being acquired(often following meningitis). Theimpact of permanent hearingimpairment on children and theirfamilies can be considerable. Lateidentification may compoundproblems in communication,language acquisition and affectother areas of development.C.2 Screening tests: The mostcommon pre-school hearingscreening test used in the UK isthe infant distraction test carriedout by two health visitors (HVDT),or by a health visitor and a trainedassistant (Table 2). It isadministered at about 6–9 monthsof age and assesses the infant’sability to turn and locate a soundsource. It is used as a universalhearing screen in about 98% ofhealth districts and achievesTable 2 Key screening tests used to detect permanent childhood hearing impairmentTestsComments1 Infant distraction test (IDT)1a Traditional Health Visitor distraction test(HVDT) universal in most districtsTest carried out at 6–9 months, usually in protected time. Cost about £25 per test includingfollow-up a1b Targeted IDTProposed in tandem with universal neonatal screening on equity grounds1c BeST test New one person IDT, with calibrated sound source2 Transient Evoked Otoacoustic Emissions Quick test carried out within days of birth. Measures acoustic energy generated by the healthy(TEOAE)cochlea in response to wide band clicks using a lightweight ear-canal probe. Cost £14 per test.Presently most used for well babies. Need agreed criteria for Pass/Refer3 MLS TEOAE New very quick version of TEOAE that may have advantages in noisy situations4 Distortion Product Otoacoustic Emissions Many implementations, need to monitor literature as to outcome(DPOAE)5 Auditory Brainstem Response (ABR) Test carried out within days of birth. Wide band clicks are presented to one ear and theresulting electrical potentials of the early auditory pathways are measured using surfaceelectrodes. Some ABR machines make pass or refer decisions, others need trained operators.High recurrent costs or long test times on some implementations. Presently most used inNICU/SCBU children6 Portable Auditory Response Cradle (PARC) Automated, quick, behavioural test which presents a 70–80dB SPL high pass noise to one orboth of the baby’s ears via an earphone or probe. The baby’s response is measured by acradle and associated computer software which compares head turns and body movements inperiods with the sound on and off. An automated decision algorithm is used to pass or refer.Probably good for severe and profound impairmentsa The full cost of health visitor (HV) time does not allow for the fact that HVs may be visiting the home anyway. However, if done in conjunction withseveral other activities its accuracy and therefore its value is likely to be reduced.APRIL 1998Pre-school hearing, speech, language and vision screening EFFECTIVE HEALTH CARE 3

coverage of about 90% of allinfants. This varies by socioeconomicstatus. There is alsovariability in the way it is carriedout, the sound generators used,the number and level of training ofthe people doing the testing, andthe adequacy of soundproofing ofthe room. This leads to concernsabout the number of children withproblems who are not identifiedduring a screen under presentarrangements.The published evidence on testperformance from clinic-basedretrospective studies and case-notereviews indicates poor andvariable sensitivity (detection rate)and specificity (true negative rate)for the HVDT. 5 The cumulativeyield is low, being about 50% by18 months of age. The average ageof confirmation of hearingimpairment via the HVDT isbetween 12–20 months, withsubsequent age of hearing-aidfitting following HVDT beingabout 18 months.Alternatively, a number ofneonatal screening tests that canbe applied within the first few daysafter birth are available (Table 2).These methods include thePortable Auditory Response Cradle(PARC), the Auditory BrainstemResponse (ABR), and TransientEvoked Otoacoustic Emissions(TEOAE). TEOAE is currently thepreferred technique for wellbabies, and automated ABR forthose in neonatal intensive care orspecial care baby units. Althoughthe PARC has been extensivelytested, its implementation has notbeen as well evaluated in multicentrestudies as TEOAE.One controlled trial has beencarried out to compare screeningmethods. 11 This trial in Wessexcompared 21,000 babies givenTEOAE screening (ABR was usedfor those failing the test) with29,000 babies who received onlythe HVDT at 6–8 months. Interimresults show that the neonatalscreening test has a specificity ofaround 98% and gave a yield of1.1 per 1,000 births by 4 monthsof age. This corresponds to theexpected prevalence rate, thusindicating a high sensitivity. Thehigh specificity and sensitivity ofthe neonatal screen is confirmedby another UK longitudinalstudy. 12–14 The cumulative yield inthe HVDT-only group was lower at0.7 per 1,000 by 18 months oldsuggesting that false negatives willemerge later on. Only 0.1 hearingproblems per 1000 births wereactually detected by the HVDTsince most were identified due toparental or professional concern,or passed the HVDT incorrectly. Inthe neonatal screening group 96%were identified under 9 months ofage compared to around half inthe HVDT-only group.In the UK, where universalneonatal screening programmeshave been implemented with goodcoverage alongside the HVDTscreen, the extra yield of theHVDT is very low (e.g. 0.1 per1000 births).C.3 Interventions for congenitalhearing impairment:Interventions includeamplification, cochlear implants orhelping the child to learn anappropriate sign language (Table3). For children with a profoundimpairment a cochlear implantmay enable the auditory neuralpathway to be stimulated directly.This is currently being evaluatedby an MRC study.While there is a growing body ofliterature on the benefits of earlyintervention, few studies are ofhigh quality. Three of the 18studies identified providereasonable evidence that earlyintervention is better than late. Ina study of 69 children identifiedby a Colorado neonatal screeningprogramme, those habilitatedbefore 3 months of age scored87% of normal for expressivelanguage, compared to only 66%for those habilitated between 3and 12 months. 15 Similarly, in thesame study, 46 children whosehearing impairments wereidentified before the age of 6months were found to have better3-year vocabulary and expressiveand receptive language than 63children whose impairment wasidentified after 6 months (afterhaving taken into account anydifferences in non-verbal cognitiveskills). 16 In another study,subjective assessments by teachersof speech intelligibility of 153children (matched for age, sex, ageof onset of hearing loss, degree ofdeafness and schooling) found thatthose fitted with hearing aidsbefore 6 months achieved higherscores than any groups of childrenfitted with hearing aids later in life. 17The benefits of early identificationin hearing-impaired children aresupported by other studies whichshow earlier onset of babbling 18 orbetter communication skills 19,20 theearlier the children were fittedwith hearing-aids. One study,however, found that the initialbenefits of early intervention onreceptive language did notpersist. 21Overall, this research supports theview that these children(particularly those with moresevere impairments) have verypoor outcomes at presentcompared with normal-hearingchildren. Earlier identification isassociated with better outcomes,particularly in the domain oflanguage acquisition andcommunication. However, theextent to which even betteroutcomes may be achieved withvery early identification is not yetclear, although the early researchresults from Colorado point to thisbeing the case. 16C.4 Cost-effectiveness: There is asignificant difference in the cost ofneonatal and HVDT screeningapproaches. The cost (includingfollow up) for universal neonatalscreening programmes is about£14,000 per 1,000 births; forHVDT it is about £25,000 per1,000 children, when done inprotected time or on a separatevisit. 22 This translates into a ‘costper child with a hearing problemidentified’ of around £17,000 forneonatal screening and £80,000for HVDT screening. These figuresdo not take into account any ofthe benefits to the child of earlier4 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screening APRIL 1998

Table 3 Key interventions for moderate to profound permanent childhood hearing impairment and ways they will be affected if UniversalNeonatal Screening is introducedInterventionFamily support, advice and informationProvision of hearing aidsProvision of communication support(spoken and/or signed)Provision of pre-school educational supportCochlear implantsProvision of other devices e.g. radio aids,tactile aids, other assistive devicesEffect of Universal Neonatal ScreeningNeeds to be effective from screen refer and onwards. Requires better multi-agencyco-operationBetter early diagnostic testing and aid fitting. Requires evaluations for mild impairments if thescreen is to be extended to this groupEarlier support neededEarlier support needed. Different skill mix needed for children in first 18 monthsEarlier implantation will be possibleNo effectdetection and habilitation nor theextra costs of the earlier treatmentand educational support whichthey will receive with neonatalscreening. Conversely, it does nottake into account other healthpromotion activities which may beundertaken by health visitors atthe same contact. However, in themajority of districts, hearing testsare carried out by health visitors inseparate clinics or during protectedtime. Therefore, many healthauthorities will be able to free someresources in moving from HVDT touniversal neonatal hearing screeningas well as improve the service.Approximately two-thirds ofdistricts have some sort ofneonatal hearing screeningprogramme: two have universalneonatal screening and theremainder target babies with oneor more risk factors for hearingimpairment. Because this meansscreening fewer but higher riskbabies, targeted screening is morecost-effective at around £14,000per case detected. However, onlyabout 50–60% of hearing-impairedchildren will have risk factors andin practice the yield will be lower. 5,22childhood 23 with a prevalence ofaround 6% of children. Thedemand for services, particularlyfor children under 4 years of age,is increasing. 24,25 Since the agedistribution at which ‘normal’children learn to speak is probably‘bell-shaped’, prevalence estimatesare dependent to a great extent onthe cut-off point used. Few dataare available on bilingual orethnically diverse groups and theassociation with social class is alsounclear.Spontaneous remission of speechand language delays identified inthe pre-school period can be high,particularly for children withspecific expressive delays, wheresome 60% of cases may resolvewithout treatment by 3 years ofage. 6 The picture for older childrenis unclear because of a lack ofresearch, but it is evident that ifchildren go on to have difficultiesin the first year of primary school,they are at risk of experiencingproblems throughout theirschooling. In addition, 41–75% ofchildren who present with earlyexpressive language delay werefound to exhibit readingdifficulties at 8 years of age.Risk factors for persistent problemsinclude the initial severity of thedelay, the extent to which thedifficulties are generalised acrossspeech and language, and theTable 4 Principal methods of screening for speech and language delay• Level of concern elicited from parent by a professional (the Parental Evaluation ofDevelopmental Status)• Parent provides information about speech and language milestones and the clinicianinterprets the results (the Early Language Milestone Scale, the Clinical Linguistic AuditoryMilestone Scale)• Parent reports on child’s current level of speech/language functioning and the clinicianinterprets the results (e.g. the Minnesota Child Development Inventory, the Ward InfantLanguage Screening Test, the Language Development Survey)• Clinician makes a judgement of child’s performance based on mixed observation/reportedD. Speech andlanguage delayD.1 Epidemiology and naturalhistory of speech and languagedelay: Speech and language delayis one of the most common neurodevelopmentaldifficulties in earlydata (The Denver Developmental Screening Test)• Clinician tests child’s speech and language performance by means of specific activitiessuch as:the child’s response to requests graded in terms of difficulty (e.g. the Hackney LanguageScreening Test, the Mayo Early Language Screening Test, the Uppsala General LanguageScreening)the child’s capacity to imitate words and sentences (Sentence Repetition Screening Test)the child’s capacity to retell storiesAPRIL 1998 Pre-school hearing, speech, language and vision screening EFFECTIVE HEALTH CARE 5

extent to which other cognitiveand developmental skills are alsodelayed. The extent to whichfactors such as social class, familyhistory, temperament and gendercontribute to relative risk is unclear.However, there is reasonableevidence to suggest that speechand language development areaffected by how well parentsinteract verbally with their childrenand by the general level ofstimulation within the homeenvironment. However, it isuncertain whether parental factorscan actually create a ‘clinical’ levelof difficulty.D.2 Screening tests: Severalscreening measures are used in theUK (Table 4). No randomisedcontrolled trials (RCTs) ofscreening programmes wereidentified by the review. Screeningtest performance variesconsiderably with sensitivitywithin the range 17–100% andspecificity in the range 43–100%.Sensitivity was generally lowerthan specificity, particularly in thebetter-quality studies. Thissuggests that it may be easier toindicate those children who arenot cases than to be clear aboutthose who are. Few studies haveattempted to compare theapplication of two or morescreening tests to a singlepopulation or to compare a singlescreening measure across differentpopulations. It is, therefore,difficult to make a judgementabout the relative value of differentprocedures or to single out anyone measure as outperforming theothers. In general, however, screensthat used parents as informantswere as accurate as those thatused formal testing procedures.The majority of the screeningprocedures currently available areapplicable after the age of 2 yearswhen the reported accuracy ofscreening is greater. However,work currently in progress isexploring a method of identifyingthose at risk of subsequentdifficulties based on auditory skillsat 9 months old. 26 Given thevariability in the natural history ofspeech and language delay, andthe high level of subsequentspontaneous improvement,particularly in the very early years,the use of a single measure at thisstage in a child’s development isunlikely to be valuable. Tests thatcan identify those children whowill fail to progress withouttreatment need to be developed.D.3 Interventions for speech andlanguage delay: Several types ofinterventions have been used forhelping children with speech andlanguage delays (Table 5). TenRCTs and 12 controlled studieswere identified which evaluatedtreatment, mostly for problems ofarticulation/phonology andexpressive language. 27–45These studies show thatinterventions are effective inenhancing speech, expressivelanguage, receptive language andauditory discrimination, relative tountreated controls. The size of thebenefits represented progress fromthe 5th to the 25th percentile on anorm-referenced test. ThisTable 5 Key intervention approaches for speech and language delaycorresponds to an overallstandardised effect size of around1.0 – an increase in the averageperformance equivalent to 1standard deviation of thedistribution of performance scores.These results are supported bydata from 26 single-caseexperimental designs. 6 No studiesspecifically compared the effects ofdifferent timing of interventionson social and educational outcomesand there are little reliable datawith which to identify the bestchoice for any area of delay.One of the interesting issues iswho most effectively provides theinterventions: professionals (i.e.speech and language therapists orspecialist teachers) orparents/others in the child’senvironment. Studies have showncomparable results for both in thecase of expressive language (effectsize from 0.65–1.11 and 1.08–1.16for professionals and parentsrespectively). In speech delay,professionals (effect size from0.94–1.11) were more effectiveThe majority of interventions are primarily behavioural in nature and may be provided byspeech and language therapists or specialist teachers, either intensively within a specialist unitor less intensively, but at regular intervals, in a clinical setting, a school or a day care setting.The three main intervention types are:Didactic interventionThe child is given a model of a sound, a word, a communication behaviour or a syntacticconstruction and an attempt made to elicit the child’s production of that model using positivereinforcement. This approach is usually carried out by the therapist or teacher.Naturalistic interventionThis approach recreates the environment which is known to optimise the child’s languagelearning opportunities, not through explicit instruction, but by making the stimulus relevant to thechild’s focus of attention. This approach is aimed at promoting the acquisition andgeneralisation of functional language and frequently involves parents as active participants. Itcan be carried out directly by a therapist or teacher or indirectly by others in the child’senvironment.Hybrid interventionThis approach combines elements of both didactic and naturalistic interventions. It recognisesthat children with delayed speech and language development may learn language in differentways from one another and from their normal peers, and may need to be exposed to a rangeof different types of environmental modifications.Other intervention approaches include non-directive therapy, auditory training,comprehension monitoring and cognitive therapy.6 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screening APRIL 1998

than parents (-0.02 to 0.20). In the No studies were found whichcase of receptive language the aimed to document the naturalreverse was found – treatment history of these conditions inprovided by parents or other untreated pre-school children. Acarers on the advice of afew studies, however, give someprofessional following diagnosis or information on what would beassessment was more effective expected to happen to the vision(average effect size of 1.43) of children at this age withcompared with direct intervention amblyopia, 56 squints, 57,58 and(average effect size of 0.02). refractive errors 59 in the absence ofintervention. These suggest thatD.4 Cost-effectiveness: No costeffectivenessstudies have examined to non-cosmetically obviousamblyopia in some children (duethe impact of screening programmes squints or mild refractive error at 3in terms of possible savings in or 4 years) may resolve withoutspecial education and other support treatment. However, there areservices provided. One study is many important gaps in the data.currently underway at ErasmusUniversity, Rotterdam. There is, Twenty-one studies were foundhowever, evidence from the US which aimed to investigate whetherwhich suggests that home-based a variety of disabilities wereintervention may be more costeffective.46 Given the positive effects target conditions. The majority ofassociated with any of the threeof indirect intervention identified studies compared the performancein this bulletin, this needs to be of children with visual defects inexamined more closely in the UK tasks such as reading with that ofcontext.their peers with normal vision, orcompared the vision of childrenwith and without disabilities suchas dyslexia. The only strong andconsistent relationship to emergeE. Pre-schoolis that children with myopiaperform better than their peers onvision screeningreading tests, (although this is notdue to myopia enhancing childE.1 Epidemiology and naturaldevelopment).history of asymptomatic vision60–63 Studies thatinvestigated the relationshipproblems: The aim of visionbetween squints and readingscreening at the age of 3–4 years isability produced inconsistentthe prevention or reduction offindings.disability due to one or more of64–67 However, childrenwith squints have been found tothe following target conditions:perform less well than their peersamblyopia (reduced visual acuity,without squints in neurodevelopmentaltests.usually in one eye, in the absenceof organic disease which cannot68–70be improved by spectacles),Amblyopia in one eye can disruptrefractive errors, and the typesdepth perception, but the effectsof squints which are not cosmeticallythat this might have are poorlyobvious and so are unlikely to beunderstood and are currently thedetected without screeningsubject of debate.(phorias and microsquints).71–73 The onlystudy found which investigatedthe perceptual difficultiesNo studies were found which hadassociated with amblyopia inthe primary aim of establishing theadulthood suggested thatprevalence of visual defects at 3–4amblyopia in one eye had littleyears of age. However, data fromimpact on perception of space orstudies of primary orthopticcontrast and was unlikely to affectscreening programmes for this ageeveryday life, although this studygroup reported a range of yieldswas methodologically flawed.for the target conditions of74 Nostudies have been carried out2.4–6.1%. 47–55 using a design which is appropriatefor establishing a causal link.Physiological data from animalstudies showed that blurred visionat a critical stage of neurologicaldevelopment could result inpermanent impairment of therelevant brain functions. This gaverise to the enthusiasm for earlydetection of amblyopia. However,the quality of the literature onvisual defects and disability, andon the natural history of theseconditions in humans, isinsufficient to know with anycertainty what might be expectedto happen in an individual childwith amblyopia, a noncosmeticallyobvious squint or arefractive error if they were leftuntreated. One large RCT in Avoncomparing vision screeningprogrammes in children agedunder 3 should provide usefulinformation on associateddisability in older children. 75 Thereis a very strong professional view,however, that amblyopia isdisabling and should be treated.E.2 Screening tests: The principalaim of the child screening test is toidentify children with amblyopia.However, tests are also carried outfor non-cosmetically obvioussquints and refractive errorsbecause these may be associatedboth with amblyopia and alsosometimes with the aim of treatingthem in their own right (Table 6).No RCTs of screening programmesfor the 3–4 year age group wereidentified. One prospectivecontrolled study was found, whichcompared visual outcomes at theage of 7 years in children whowere screened at age 3 byorthoptists, general practitioners orhealth visitors in Newcastle. 56Children with straight-eyedamblyopia and refractive errorswere identified significantly earlierin the orthoptic screening cohort,but there was no difference in thetime of identification of obvioussquint. Despite the fact that manymore children with amblyopiawere identified and treated in theorthoptic screening cohort, theprevalence of amblyopia at 7 yearsof age was the same in all threecohorts. 56 However, this study hascertain methodologicalweaknesses.APRIL 1998Pre-school hearing, speech, language and vision screening EFFECTIVE HEALTH CARE 7

Table 6 Common contents of pre-school vision screeningCommon pre-school vision screening tests• Checking the appearance of the eyes• Cover/uncover test for squint• Single optotype or linear visual acuity test e.g. Sheridan Gardiner or SnellenAdditional tests performed in primary orthoptic screen• Ocular movements• Convergence• Prism test• Test of stereoacuity e.g. Fisby or Lang stereotestSixteen other studies that aimedto establish the effectiveness ofpre-school vision screening wereeither observational or audits. 47–50,52–55, 76–83Uptake rates for primaryorthoptic screening ranged fromaround 44–80%. Vision screeningby health visitors, generalpractitioners or clinical medicalofficers, undertaken as part of aroutine surveillance contact, had amean uptake rate of 76.2%. Ratesof referral from primary orthopticscreening programmes rangedfrom 4.1–10.6% of the screenedpopulation, with no differencesbetween the various healthprofessional groups.In five studies of orthopticscreening programmes, thepositive predictive value, (thepercentage of those referred whoare true positives) varied from47–66%. 47, 51–53, 55 Positive predictivevalues of over 90% were achievedwhere the definition of a positivecase was more broad. 48, 49, 54 Inprogrammes run by health visitorsor clinical medical officers, thepositive predictive value was muchmore variable, ranging from14–62%. 47, 49, 51, 80, 81 In other words,orthoptists are generally better atidentifying these problems thandoctors or health visitors. Asignificant number of areas useorthoptists (usually on a separateoccasion) to test for visualproblems around 3–4 years of age.E.3 Interventions for visionproblems: The treatments foramblyopia include patching thenon-amblyopic eye and spectaclecorrection of associated refractiveerror, and surgery to correctsquints (Table 7). Five prospectiveRCTs of treatment and six non-RCTs were found. None werespecifically relevant to this agegroup and in no study was thetreatment compared to anuntreated control group, and thusthe absolute effects of treatmentare not known.Three of the RCTs compared theeffect of the CAM, (a visionstimulator grating) withconventional orthoptic treatmentand showed no significantadvantage. 84-86 One small RCTshowed that adding the druglevodopa/carbidopa to orthoptictreatment for amblyopia improvedvisual acuity and contrastsensitivity. However at one monththe intervention group hadregressed slightly and the controlgroup had not maintainedimprovement. 87 This latter findingis supported by a controlled studycomparing different occlusionTable 7 Treatments for visual problems identified at pre-school screen• Amblyopia: intermittent occlusion of the non-amblyopic eye with a patch• Intermittent squints: followed up and may be treated with surgery if they deteriorate• Latent squints with hypermetropia (long sighted): often spectacle correction only• Microsquints and latent squints without hypermetropia: not treated• Refractive errors: left untreated or corrected by spectacles depending on severity andwhether associated with amblyopiaregimes, in which 33% of thosewith improved acuity after treatmentshowed some deterioration after 3months. 88 Drugs and CAM are nowrarely used in the UK.Five controlled trials compareddifferent approaches to amblyopiatreatment. 88–92 All of these havemethodological flaws which limitthe value of their findings. Overall,whilst there is evidence that thevision of children with amblyopiaimproves with treatment, 84–92 theseimprovements may not be84, 86–88sustained.Seven studies evaluating screeningprogrammes reportedimprovements in visual acuity oftwo or more Snellen lines in50–80% of children who weretreated for amblyopia followingscreening. 47, 48, 50, 53, 55, 56, 93 However, asnone of these have a comparisongroup of untreated children, it isdifficult to assess the degree towhich these changes areattributable to treatment. None ofthe studies assessed long-termoutcomes of treatment, evaluatedtreatment in terms of disability orother patient-perceived outcomes.In addition, none of the studiesassessed the potential negativeimpact of orthoptic treatment(such as patching) on children ortheir families, which has beensuggested by recent qualitativework. 94An RCT 95 and a non-RCT 96 foundthat the use of pre-operative prismcorrection improved the outcomeof squint surgery. However, thesetrials only included patients withobvious squints. No controlledstudies of treatment for latent ormicro-squints were found.Spectacles are a highly effectiveway of correcting refractive errorsthat requires no evaluation inolder children. However, thetreatment of refractive error in theabsence of amblyopia or manifestsquint is of unproven benefit andmay even cause harm byinhibiting the normal refractivedevelopment of the eye(emmetropisation). 97 APRIL 19988 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screening

F. ImplicationsThis bulletin has summarised arange of important new researchon three of the component partsof the child health surveillanceprogramme. The three individualscreening tests must be seen inthe context of an overall childhealth programme which, aswell as screening for specifichealth problems, also providesother health care, advice andsupport for families of youngchildren. These individualscreening programmes should,therefore, not be seen in isolation.F.1 Hearing screening: Currentservices are seen by parents andprofessionals alike as poorly coordinatedand fragmented. 5 Asmany as 200 of the 840 childrenper year born with significanthearing impairment may not beidentified until after 3 1 /2 years ofage. On the grounds of equity,responsiveness and costeffectiveness,the transition fromuniversal HVDT to universalneonatal hearing screening, incombination with targeted InfantDistraction Tests, is considered thebest value for money. Some healthauthorities will be able to freesome resources in moving fromHVDT to universal neonatalhearing screening, as well asimprove the service.These results are currently beingconsidered by the NationalScreening Committee inconjunction with plans forimplementation, training andquality assurance. This shouldinclude quality markers such asinitial coverage, sensitivity forcongenital hearing impairment,and a maximum false-positive rate,and indicate relevant data to becollected. There will be a need fordevelopment and co-ordinationof audiological follow-up servicesand the ongoing audiological andeducational care of pre-schooland school-aged hearing-impairedchildren, many of whom mayhave other and/or complexneeds. 10F.2 Speech and language delay:There are insufficient dataavailable to recommend theintroduction of populationscreening for early speech andlanguage delay because there isnot yet adequate agreement as towhich children will not progressunless they are given interventionand on the grounds that the currentscreening measures themselveshave yet to be shown to haveadequate predictive validity.Nonetheless, early primary speechand language delay should remaina cause for concern. This isbecause of the problems it maypose for the individual child, theconcern it causes parents, the factit may serve as a litmus test forother problems which commonlyaccompany it such as cognitiveimpairment, behaviour andconduct disorders, and because ofthe implications that it may havefor literacy and socialisation inschool. However, furtherrefinement of screening measures,with accompanying research, isneeded before universal screeningis recommended. There is also acase for the evaluation of primaryprevention programmes aimed atreducing the risk of speech andlanguage delay. Research shouldexamine interventions whichencourage parents to read withtheir children (e.g. Bookstart) andteach parents to listen to andrespond appropriately to theirchild’s communication attempts.F.3 Pre-school vision screening:Amblyopia can cause a significantreduction in visual acuity, asmeasured by the Snellen test, butthis may not be the best outcomemeasure. Equally, the physical,psychological and social implicationsof reduced visual acuity in one eyeare not well understood. Thus it isnot clear that amblyopia should beseen as the cause of significantdisability or handicap. No studyhas addressed adequately thepossible negative aspects ofamblyopia treatment. Furtherresearch is needed to ascertainboth the importance of thiscondition and the most effectiveand acceptable treatment.Pre-school screening for refractiveerrors and non-obvious squint,without associated amblyopia,does not seem to be justified sincethese conditions do not appear tobe problematic by themselves andtheir treatment at an asymptomaticstage has not been shown to conferbenefit. The fact that existingresearch fails to show that screeningfor, and treatment of, amblyopiaare beneficial is not however proofthat a screening and treatmentprogramme confers no benefit.Research is needed to establishwhether pre-school screening is ofbenefit. Pending further research,and the recommendations of theNational Screening Committee, itwould seem inappropriate toinitiate new programmes, orextend or dismantle existing ones.Given the current uncertainty overthe potential benefits and harms oftesting and some correctivemeasures, it is particularly importantthat professionals give adequateand accurate information to parents.References1. Hall D, Hill P, Elliman D. The childsurveillance handbook. 2nd edn.Oxford: Radcliffe Medical Press, 1994.2. Hall DMB. Health for all children:a programme for child healthsurveillance; the report of the JointWorking Party on Child HealthSurveillance. 1st edn. Oxford: OxfordUniversity Press, 1989.3. Hall DMB. Health for all children: aprogramme for child healthsurveillance: the report of the JointWorking Party on Child HealthSurveillance. 2nd edn. Oxford:Oxford University Press, 1992.4. Hall DMB. Health for all children. Thereport of the Joint Working Party onChild Health Surveillance. 3rd edn.Oxford: Oxford University Press, 1996.5. Davis A, Bamford J, Wilson I, et al.A critical review of the role ofneonatal hearing screening in thedetection of congenital hearingimpairment. Health Technol Assess1997;1(10).6. Law J, Boyle J, Harris F, et al.Screening for speech and languagedelay: a systematic review of theliterature. CRD Report No 12.University of York: NHS Centre forReviews and Dissemination, 1998.APRIL 1998 Pre-school hearing, speech, language and vision screening EFFECTIVE HEALTH CARE 9

7. Snowdon SK, Stewart-Brown SL.Preschool vision screening: results of asystematic review. CRD Report No. 9.University of York: NHS Centre forReviews and Dissemination, 1997.8. Wilson JMG, Jungner G. Principlesand practice of screening for disease.Geneva: World Health Organisation,1968.9. Cochrane A, Holland W. Validationof screening procedures. Br Med Bull1971;27:3–8.10. Fortnum H, Davis A. Epidemiology ofpermanent childhood hearing impairmentin Trent Region 1985–1993.Br J Audiol 1997;31:409–46.11. Hunter MF, Kimm L, Cafarelli DeesD, et al. Feasibility of otoacousticemission detection followed by ABRas a universal neonatal screening testfor hearing impairment. Br J Audiol1994;28:47–51.12. Watkin PM. Neonatal otoacousticemission screening and theidentification of deafness. Arch DisChild 1996;74:F16–F25.13. Watkin PM. Optimising the specificityof a neonatal otoacoustic emissionscreen. Unpublished report, 1996.14. Watkin PM. Outcomes of neonatalscreening for hearing loss byotoacoustic emissions. Arch Dis Child1996;75:F158–68.15. Downs MP. Universal newbornhearing screening – the Coloradostory. Int J Pediatr Otorhinolaryngol1995;32:257–9.16. Yoshinaga-Itano C, Sedey A, ApuzzoM, et al. The effect of earlyidentification on the development ofdeaf and hard-of-hearing infants andtoddlers. Paediatrics 1998; in press.17. Markides A. Age at fitting of hearingaids and speech intelligibility. Br JAudiol 1986;20:165–7.18. Eilers RE, Oller DK. Infantvocalisations and the early diagnosisof severe hearing impairment. JPediatr 1994;124:199–203.19. Ramkalawan TW, Davis AC. Theeffects of hearing loss and age ofintervention on some languagemetrics in young hearing-impairedchildren. Br J Audiol 1992;26:97–107.20. Ramkalawan TW. Factors thatinfluence the language andcommunication of hearing impairedchildren. Unpublished. PhD thesis,1997.21. Musselman CR, Wilson AK, LindsayPH. Effects of early intervention onhearing impaired children. ExceptChild 1988;55:222–8.22. Stevens JC, Hall DMB, Davis A, et al.A survey of the costs of hearingscreening in the first year of life inEngland and Wales. Arch Dis Child1998;78:14–19.23. Drillien C, Drummond M.Developmental screening and thechild with special needs. London:Heinemann, 1983.24. Jowett S, Evans C. Speech andlanguage therapy services for children.Slough: NFER, 1996.25. Reid J, Millar S, Tait L, et al. Pupilswith special education needs: the roleof speech and language therapists.Edinburgh: SCER, 1996.26. Ward S, Birkett D. The Ward InfantLanguage Screening Test, assessment,acceleration and remediation.Manchester: Manchester Health CareTrust, 1994.27. Almost D, Rosenbaum P.Effectiveness of speech interventionfor phonological disorders: arandomised controlled trial.Unpublished, 1997.28. Conant S, Budoff M, Hecht B, et al.Language intervention: a pragmaticapproach. J Autism Dev Disord1984;14:301–17.29. Fey ME, Cleave PL, Long SH, et al.Two approaches to the facilitation ofgrammar in children with languageimpairment: an experimentalevaluation. J Speech Hear Res1993;36:141–57.30. Fey ME, Cleave PL, Ravida AI, et al.Effects of grammar facilitation on thephonological performance ofchildren with speech and languageimpairments. J Speech Hear Res1994;37:594–607.31. Gibbard D. Parental-basedintervention with pre-schoollanguage-delayed children. Eur JDisord Commun 1994;29:131–50.32. Girolametto L, Pearce PS, WeitzmanE. Interactive focused stimulation fortoddlers with expressive vocabularydelays. J Speech Hear Res1996;39:1274–83.33. Girolametto L, Pearce PS, WeitzmanE. The effects of focused stimulationfor promoting vocabulary in youngchildren with delays: a pilot study. JChild Commun Dev 1995;17:39–49.34. Lancaster G. The effectiveness ofparent administered input trainingfor children with phonologicaldisorders. MSc Thesis. London: CityUniversity, 1991.35. Matheny N, Panagos JM. Comparingthe effects of articulation and syntaxprograms on syntax and articulationimprovement. Lang Speech Hear Servin School 1978;9:57–61.36. McDade A, McCartan PA.Partnership with parents: a pilotstudy. Unpublished, 1996.37. Reid J, Donaldson ML, Howell J, et al.The effectiveness of therapy for childphonological disorder. TheMetaphon approach. In: Aldridge M(ed.). Child Language. Clevedon,Avon: Multilingual Matters,1996:165–75.38. Schwartz RG, Chapman K, Terrell BY,et al. Facilitating word combinationin language-impaired childrenthrough discourse structure. J SpeechHear Disord 1985;50:31–9.39. Shelton RL, Johnson AF, RuscelloDM, et al. Assessment of parentadministeredlistening training forpreschool children with articulationdeficits. J Speech Hear Disord1978;43:242–54.40. Stevenson P, Bax M, Stevenson J. Theevaluation of home-based speechtherapy for language delayedpreschool children in an inner cityarea. Br J Disord Commun1982;17:141–8.41. Ward S. Validation of a treatmentmethod. In the proceedings of the2nd Conference of the ComitePermanente de Liaison desOthophonistes-Logopedes del’UE[CPLOL], Antwerp, 1994.42. Warrick N, Rubin H, Rowe-Walsh S.Phoneme awareness in languagedelayedchildren: comparativestudies and interventions. AnnDyslexia 1993;43:153–73.43. Whitehurst GJ, Fischel JE, LoniganCJ, et al. Treatment or earlyexpressive language delay: if, when,and how. Top Lang Disord1991;11:55–68.44. Wilcox MJ, Leonard LB. Experimentalacquisition of Wh- questions inlanguage-disordered children. JSpeech Hear Res 1978;21:220–39.45. Zwitman DH, Sonderman JC. Asyntax program designed to presentbase linguistic structures tolanguage-disordered children. JCommun Disord 1979;12:323–35.46. Barnett WS, Escobar CM, RavstenMT. Parent and clinic earlyintervention for children withlanguage handicaps: a costeffectiveness analysis. J DivisionEarly Child 1988;12:290–8.47. Bolger P, Stewart Brown S,Newcombe E, et al. Vision screeningin preschool children: comparison oforthoptists and clinical medicalofficers as primary screeners. BMJ1991;303:1291–4.48. Beardsell R. Orthoptic visualscreening at 3.5 years byHuntingdon Health Authority. BrOrthop J 1989;46:7–13.49. Edwards R. Orthoptists as pre-schoolscreeners: a 2 year study. Br Orthop J1989;46:14–19.10 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screeningAPRIL 1998

50. Ingram R, Holland W, Walker C, etal. Screening for visual defects inpreschool children. Br J Ophthalmol1986;70:16–21.51. Jarvis S, Tamhne R, Thompson L, etal. Preschool vision screening. ArchDis Child 1990;65:288–94.52. Milne C. An evaluation of casesreferred to hospital by the Newcastlepreschool orthoptic service. BrOrthop J 1994;51:1–5.53. Newman D, Hitchcock A, McCarthyH, et al. Preschool vision screening:outcome of children referred to thehospital eye service. Br J Ophthalmol1996;80:1077–82.54. Wormald R. Preschool visionscreening in Cornwall: performanceindicators of community orthoptists.Arch Dis Child 1991;66:917–20.55. Williamson T, Andrews R, Dutton G,et al. Assessment of an inner cityvisual screening programme forpreschool children. Br J Ophthalmol1995;79:1068–73.56. Bray L, Clarke M, Jarvis S, et al.Preschool vision screening: aprospective comparative evaluation.Eye 1996;10:714–18.57. Good W, da Sa L, Lyons C, et al.Monocular visual outcome inuntreated early onset esotropia. Br JOphthalmol 1993;77:492–4.58. Aurell E, Norrsell K. A longitudinalstudy of children with a familyhistory of strabismus. Br JOphthalmol 1990;74:589–94.59. Hard A, Williams P, Sjostrand J. Dowe have optimal screening limits inSweden for vision testing at the ageof 4 years? Acta Ophthalmol Scand1995;73:483–5.60. Stewart-Brown S, Haslum M, ButlerN. Educational attainment of 10 yearold children with treated anduntreated visual defects. Dev MedChild Neurol 1985;27:504–13.61. Teasdale T, Fuchs J, Goldschmidt E.Degree of myopia in relation tointelligence and educational level.Lancet 1988:1351–54.62. McManus I, Mascie-Taylor C.Biosocial correlates of cognitiveabilities. J Biosoc Sci1983;15:289–306.63. Peckham C, Gardiner P, Goldstein H.Acquired myopia in 11 year oldchildren. BMJ 1977;1:542–5.64. Grosvenor T. Are visual anomaliesrelated to reading ability? J AmOptom Assoc 1977;48:510–17.65. Simons H, Gassler P. Visionanomalies and reading skill: a metaanalysisof the literature. Am J OptomPhysiolog Optics 1988;65:893–904.66. Bishop D, Jancey C, Steel A.Orthoptic status and readingdisability. Cortex 1979;15:659–66.67. Hall P. The relationship betweenocular functions and readingachievement. J Pediatr OphthalmolStrabismus 1991;28:17–9.68. Alberman E, Butler N, Gardiner P.Children with squints – ahandicapped group? Practitioner1971;206:501–6.69. Bax M, Whitmore K.Neurodevelopmental screening inthe school-entrant medicalexamination. Lancet1973;825:368–70.70. McGee R, Williams S, Simpson A, etal. Stereoscopic vision and motorability in a large sample of sevenyear old children. J Hum MovementStudies 1991;13:343–52.71. Fielder A, Moseley M. Doesstereopsis matter in humans? Eye1996;10:233–8.72. Jones R, Lee D. Why two eyes arebetter than one: the two views ofbinocular vision. J Exp Psychol, HumPercept Perform 1981;7:30–40.73. Servos P, Goodale M, Jakobson L.The role of binocular vision inprehension: kinematic analysis.Vision Res 1992;32:1513–21.74. Kani W. Human amblyopia and itsperceptual consequences. Durham:University of Durham, 1980.75. Williams C, Harvey I, Frankel S, et al.Preschool vision screening – resultsof a randomised controlled trial.Invest Ophthalmol Vis Sci1996;37:111.76. Nolan J. Data on preschool visionscreening in South Birmingham,April 1995–March 1996. Personalcommunication, 1996.77. Allen J, Bose B. An audit of preschoolvision screening. Arch Dis Child1993;67:1292–93.78. Cameron H, Cameron M. Visualscreening of pre-school children.BMJ 1978:1693–94.79. Gallaher R. Community orthopticvisual screening: first year report,1994–1995. South BuckinghamshireNHS Trust, 1995.80. Kohler L, Stigmar G. Vision screeningof four-year-old children. ActaPaediatr Scand 1973;62:17–27.81. Kohler L, Stigmar G. Visual disordersin 7-year-old children with andwithout previous vision screening.Acta Paediatr Scand 1978;67:373–7.82. Seng C, Curson J. Study of theoutcomes of referrals from theorthoptic vision screening programme.Luton: Bedfordshire Health, 1991.83. James J. Data on preschool visionscreening in Highworth, Swindon,1983–1991. Personalcommunication, 1996.84. Keith C, Howell E, Mitchell D, et al.Clinical trial of the use of rotatinggrating patterns in the treatment ofamblyopia. Br J Ophthalmol1980;64:597–606.85. Nyman K, Singh G, Rydberg A, et al.Controlled study comparing CAMtreatment with occlusion therapy. BrJ Ophthalmol 1983;67:178–80.86. Tytla M, Labow-Daily L. Evaluationof the CAM treatment for amblyopia:a controlled study. Invest OphthalmolVis Sci 1981;20:400–6.87. Leguire L, Walson P, Rogers G, et al.Longitudinal study oflevodopa/carbidopa for childhoodamblyopia. J Pediatr OphthalmolStrabismus 1993;30:354–60.88. Terrell Doba A. Cambridgestimulator treatment for amblyopia.An evaluation of 80 consecutivecases treated by this method. AustralJ Ophthalmol 1981;9:121–7.89. Lennerstrand G, Samuelsson B.Amblyopia in 4-year-old childrentreated with grating stimulation andfull-time occlusion – a comparativestudy. Br J Ophthalmol1983;67:181–90.90. Malik S, Gupta A, Grover V.Occlusion therapy in amblyopia witheccentric fixation. Br J Ophthalmol1970;54:41–5.91. Sullivan G, Fallowfield L. Acontrolled test for the CAMtreatment for amblyopia. Br Orthopt J1980;37:47–55.92. Veronneau Troutman S, Dayanoff S,Stohler T, et al. Conventionalocclusion vs pleoptics in thetreatment of amblyopia. Am JOphthalmol 1974;78:117–20.93. Fathy V, Elton P. Orthoptic screeningfor three- and four-year olds. PublicHealth 1993;107:19–23.94. Snowdon S, Stewart-Brown S.Amblyopia and disability: aqualitative study. Oxford: HealthServices Research Unit, University ofOxford, 1997.95. Group PASR. Efficacy of prismadaptation in the surgicalmanagement of acquired esotropia.Arch Ophthalmol 1990;108:1248–56.96. Ohtsuki H, Hasebe S, Tadokoro Y, etal. Preoperative prism correction inpatients with acquired esotropia.Graefes Arch Clin Exp Ophthalmol1993;231:71–5.97. Troilo D. Neonatal eye growth andemmetropisation – a literaturereview. Eye 1991;6:154–60.APRIL 1998 Pre-school hearing, speech, language and vision screening EFFECTIVE HEALTH CARE 11

The Research Team:This bulletin summarises theresults of three recent systematicreviews, commissioned by the NHSHealth Technology AssessmentProgramme, which were led by thefollowing people:HearingJohn Bamford, University of ManchesterAdrian Davis, MRC Institute of HearingResearch, NottinghamSpeech and LanguageJim Boyle, Strathclyde UniversityJames Law, City UniversityVisionSarah Chapman, University of OxfordSarah Stewart-Brown, University ofOxfordThe bulletin was written and produced by:Professor Trevor Sheldon, Paul Wilson,Mary Turner Boutle and Frances Sharp,NHS Centre for Reviews andDissemination, University of York.Acknowledgements:Effective Health Care would like toacknowledge the helpfulassistance of the followingindividuals who commented onthe text.■ Sheila Adam, Health ServicesDirectorate, Department of Health■ Mark Baker, North Yorkshire HealthAuthority■ Simon Balmer, Nuffield Institute forHealth, Leeds■ Allan Colver, Community ChildHealth, Gateshead■ Alison Evans, University of Leeds■ Jenny Firth-Cozens, NHS ExecutiveNorthern and Yorkshire■ Alexander Foss, Queen’s MedicalCentre, Nottingham■ David Hall, Sheffield Children’sHospital■ Steve Jarvis, Community Child Health,Gateshead■ Thomas Klee, University of Newcastle■ Dee Kyle, Bradford Health Authority■ Simon Lenton, Community ChildHealth, Bath■ Roderick MacFaul, Paediatrics andChild Health, Department of Health■ Colin Pollock, Wakefield HealthAuthority■ D Roberts, Child Health Services,NHS Executive■ Sue Roulstone, Bristol■ Diana Sanderson, York HealthEconomics Consortium■ Colin Waine, Sunderland HealthAuthority■ John Wright, Bradford Royal Infirmary■ Keith Young, Health ServicesDirectorate, Department of HealthThe Effective Health Care bulletins arebased on systematic review andsynthesis of research on the clinicaleffectiveness, cost-effectiveness andacceptability of health serviceinterventions. This is carried out by aresearch team using establishedmethodological guidelines, with advicefrom expert consultants for each topic.Great care is taken to ensure that thework, and the conclusions reached,fairly and accurately summarise theresearch findings. The University ofYork accepts no responsibility for anyconsequent damage arising from theuse of Effective Health Care.Effective Health Care BulletinsVol. 21. The prevention and treatment ofpressure sores2. Benign prostatic hyperplasia3. Management of cataract4. Preventing falls and subsequent injuryin older people5. Preventing unintentional injuries inchildren and young adolescents6. The management of breast cancer7. Total hip replacement8. Hospital volume and health careoutcomes, costs and patient accessVol. 31. Preventing and reducing the adverseeffects of unintended teenagepregnancies2. The prevention and treatment of obesity3. Mental health promotion in high riskgroups4. Compression therapy for venous legulcers5. Management of stable angina6. The management of colorectal cancerVol. 41. Cholesterol and coronary heart disease:screening and treatmentSubscriptions and enquiries:Effective Health Care bulletins are published in association with FT Healthcare. The Department of Health funds a limited number of thesebulletins for distribution to decision makers. Subscriptions are available to ensure receipt of a personal copy. 1998 subscription rates, includingpostage, for bulletins in Vol. 4 (6 issues) are: £42/$68 for individuals, £68/$102 for institutions. Individual copies of bulletins from Vols 1–3 areavailable priced £5/$8 and from Vol. 4 priced £9.50/$15. Discounts are available for bulk orders from groups within the NHS in the UK and toother groups at the publishers discretion.Please address all orders and enquiries regarding subscriptions and individual copies to Subscriptions Department, Financial Times Professional,PO Box 77, Fourth Avenue, Harlow CM19 5BQ (Tel: +44 (0) 1279 623924, Fax: +44 (0) 1279 639609). Cheques should be made payable toFinancial Times Professional Ltd. Claims for issues not received should be made within three months of publication of the issue.Enquiries concerning the content of this bulletin should be addressed to NHS Centre for Reviews and Dissemination, University of York, YorkYO1 5DD; Fax (01904) 433661 email revdis@york.ac.ukCopyright NHS Centre for Reviews and Dissemination, 1998. NHS organisations in the UK are encouraged to reproduce sections of thebulletin for their own purposes subject to prior permission from the copyright holder. Apart from fair dealing for the purposes of research orprivate study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may only be produced,stored or transmitted, in any form or by any means, with the prior written permission of the copyright holders (NHS Centre for Reviews andDissemination, University of York, York YO1 5DD).The NHS Centre for Reviews and Dissemination is funded by the NHS Executive and the Health Departments of Scotland, Wales and NorthernIreland; a contribution to the Centre is also made by the University of York. The views expressed in this publication are those of the authors andnot necessarily those of the NHS Executive or the Health Departments of Scotland, Wales or Northern Ireland.Printed and bound in Great Britain by Latimer Trend & Company Ltd., Plymouth. Printed on acid-free paper. ISSN: 0965-028812 EFFECTIVE HEALTH CARE Pre-school hearing, speech, language and vision screeningAPRIL 1998