vol. 18, no. 3, pp. 200-211 a model of early detection and...

Infants & Young ChildrenVol. 18, No. 3, pp. 200-211© 2005 Uppincott Williams & Wilkins, Inc.

A Model of Early Detection andDiagnosis of Autism SpectrumDisorder in Young Children

Suniti Cbakraharti, MD; Christine Haubus, RLCST;Sally Dugmore, RGN; Gill Orgill, RGN; Frank Devine, RMN

Autism and autism spectrum disorder (ASD) are a group of severe developmental disorders thatare characterized by 3 core sets of developmental abnormalities: impairment of social interaction,verbal and nonverbal communication, and restricted, repetitive patterns of behavior. The disorderis far more common than previously thought. There is no cure for autism but it is apparent thatearly detection followed by early intervention is likely to provide the best chance of long-term ben-eficial outcome in this condition. Unfortunately, until recently, there had been no validated methodof comprehensive early detection of ASD, nor a tool with adequate sensitivity and specificity tobe recommended for universal screening of preschool children with ASD. We describe a modelof comprehensive early detection and diagnosis of ASD that is achieved by using the resourcesof primary care workers and a multidisciplinary team with skill and experience in assessing de-velopmental problems in young children and specific expertise in ASD. Both early detection anddiagnosis may be carried out by this team in collaboration with parents and primary care pro-fessionals and can result in high rates of detection and diagnosis of ASD. Key words: aspergersyndrome, autism, autism spectrum disorder, diagnosis, earty detection, earty intervention,pervasive devetopmentai disorder not otherwise specified, preschoot chitdren, screening

THERE is general agreement that early de-tection and early intervention in disease

or disability are essential for the long-termbenefit of the child and family and also, inthe long run, of society (eg, Guralnick, 1997;Keating & Hertzman, 1999] Shonkoff &Meisels, 2000). Parents want to be told atthe earliest possible opportunity if thereis any concern about their child's develop-ment or well-being (Cunningham, Morgan, &McGucken, 1984; Quine & Pahl, 1987). Thisallows intervention services to be delivered to

From the Chitd Devetopment Center, Centrat Ctinic,Stafford, United Kingdom.

We are gratefut to att the parents and chitdren as tvett asto our cotteagues at the Chitd Devetopment Center andin the Community without whose hetp and support thiswork woutd not have been possible.

Corresponding author: Suniti Chakrabarti, MD, ChitdDevetopment Center, Centrat Ctinic, Stafford ST16 3AE,UntCedKingdom (e-mail:[email protected]).

200

the child and support provided for the fam-ily. Parents are empowered to become knowl-edgeable about the nature and source of theirchild's difficulties and plans can be made forappropriate educational intervention for thechild. It also enables the family to plan aheadfor their future, especially in conditions thathave genetic implications. Beside this, it islikely that the window of opportunity to bringabout beneficial changes in at least some do-mains of children's development may be timesensitive (Nelson, 2000). For all these reasonsthere is increasing emphasis on early identifi-cation of developmental problems and institu-tion of appropriate early intervention to max-imize the life chances of the child and thefamily.

These same principles hold true for chil-dren with autism and autism spectrum dis-order (ASD). Autism and ASD, also knownas pervasive developmental disorder (PDD),are a group of severe developmental disor-ders that are characterized by 3 core sets of

A Model of Early Detection and Diagnosis of Autism Spectrum Disorder 201

developmental abnormalities: impairment ofsocial interaction and communication, and re-stricted and repetitive patterns of behaviorand interest (American Psychiatric Associa-tion [APA], 1994). The developmental abnor-malities manifest themselves early in infancyand may be present from birth. Autism is themost well defined and most intensively re-searched of all childhood developmental dis-orders. It is also not as uncommon as was pre-viously thought (Wing & Potter, 2002).

Current best estimates of the prevalenceof ASD in children under the age of 8 yearsis of the order of 60 per 10,000 children(Fombonne, 2003). Autism spectrum disordercan occur in children at all levels of intel-ligence. However, irrespective of the associ-ated intellectual level, nearly all children withASD have significant long-term disability. Thecause or causes of autism are not known butit is recognized that there is a strong geneticcomponent toward the liability to developASD (Rutter, Silberg, O'Connor, & Simonoff,1999).

There is increasing recognition that earlydetection followed by early intervention tar-geting the core developmental deficits is likelyto provide the best chance of long-term ben-eficial outcome for these children (Hoyson,Jamieson, & Strain, 1984; McEachin, Smith,& Lovaas, 1993; National Research Coun-cU, 2001; Rogers & Lewis, 1989). There istherefore an urgency to develop tools andmethodologies for early detection and di-agnosis of ASD that are prerequisites forthe institution of high-quality, evidence-basedinterventions.

Efforts at devising screening and diagnos-tic instruments for autism go back a longtime. Historically the earliest systematic at-tempt at this is the Rimland Diagnostic Formfor Behavior Disturbed Children, Form E-1.This was rapidly modified and replaced byForm E-2 (Rimland, 1968), which is still avail-able from the Autism Research Institute in SanDiego. Rimland's instrument is based on LeoKanner's original conceptualization of autismthat has undergone substantial changes in theintervening years and there are no studies to

show^ how well it fares in diagnosing autismand ASD according to currently accepted di-agnostic concepts. Other previous importantinstruments for screening/diagnosis of autismand ASD are Behavior Rating Instrument forAutistic and Atypical Children (Ruttenberg,Dratman, Frakner, & Wenar, 1966), Handi-cap, Behavior and Skills schedule (Wing &Gould, 1978), the Behavior Observation Scale(Freeman, Ritvo, Guthrie, et al., 1978), as wellas the Autism Behavior Checklist (Krug, Arick,& Almond, 1980) and the highly influen-tial Childhood Autism Rating Scale (Schopler,Reichler, & Renner, 1988).

Of the newer diagnostic instrumentsthat conform to the current ICD-10 (WorldHealth Organization, 1992) and DSM-IViAPA,1994) diagnostic criteria of autism and ASD(PDD), the most important instruments arethe Autism Diagnostic Interview - Revised(ADI-R; Lord, Rutter, & Le Couter, 1994),Autism Diagnostic Observation Schedule-Generic (Lord et al., 2000), and the Diagnos-tic Interview for Social and CommunicationDisorders (Wing, Leekam, Libby, Gould, &Larcombe, 2002). The screening instrumentsfor use speciflcally for very young childrenare CHAT (Checklist for Autism in Toddlers;Baron-Cohen, Allen, & Gillberg, 1992), Per-vasive Developmental Disorder ScreeningTest (PDDST; Seigel, 1998), Screening Toolfor Autism in Two-year-olds (STAT; Stone,Coonrod, & Ousley, 2000), and the ModifiedCHAT (M-CHAT; Robins, Fein, Barton, &Green, 2001). Of these screening tools,CHAT is the most rigorously researched andvalidated tool for use in very young children.CHAT is designed to screen for autism in chil-dren at 18 months of age and was originallyused for screening a normal population of16,235 children. CHAT consists of 2 parts,a parent questiormaire consisting of 9 itemsand a child practitioner observation scheduleconsisting of 5 items. Initial research showedCHAT to have very high specificity in identify-ing cases of autism among young children butsubsequent follow-up study has shown thatCHAT has poorer sensitivity in identifyingless severe cases of ASD (Baird et al., 2000).

202 INFANTS & YOUNG CHILDREN^ULY-SEPTEMBER 2005

The conclusion of the US multidisciplinaryconsensus panel for screening and diagnosisof ASD on CHAT is that "as a tool for identi-fying 18-month-olds at risk of autism from anormal population, the CHAT appears to be auseful tool, but not an entirely sufficient tool,for identifying the majority of children whowill fall within the autistic spectrum" (Filipeket al., 1999). Nevertheless, apart from its useas a screening tool, CHAT has generated con-siderable additional benefits in training pro-fessionals, raising awareness among primaryhealthcare workers, and showing for the firsttime that it is possible to consistently identifychildren as young as 18 months with autismfrom a general population of children.

One of the requirements of the success-ful operation of CHAT screening is the pres-ence of a system of universal surveillance ofthe general population of young children anda group of primary care workers, such ashealth visitors, to carry out this surveillance,as is availabJe jn the United Kingdom (VK).To overcome the absence of a healthcarevisiting professional group to carry out thepractitioner observation component of CHATscreening and to maximize the identificationof children who show a period of regressionbetween 18 months and 2 years of age, re-searchers in the United States led by Robins etal. have devised a modified version of CHATdesignated M-CHAT. M-CHAT consists of a23-item parent questionnaire, which incorpo-rates the original 9 items of the CHAT parentquestionnaire as well as additional items. Thequestionnaire takes about 10 minutes to com-plete and is used purely as a screening instru-ment and not for diagnosis. Initial results ofstudy using M-CHAT shows excellent sensitiv-ity and specificity (Robins et aL, 2001). Fur-ther studies are being conducted to determinethe usefulness of M-CHAT as a screen in a gen-eral population of children.

Another screening questionnaire, PDDST, isalso in the format of a parent questionnairethat is available in 3 stages and can be used fordifferent levels of screening from the primarycare to the specialist clinic. In initial stud-ies, PDDST shows high sensitivity but moder-

ate specificity (Watson, Baranek, & Di Lavore,2003) in identifying children with ASD butmore work on the use of this tool iscontinuing.

A further screening instrument, STAT, is anobservational tool for use as a second-levelscreen for children referred with developmen-tal problems. The tool is used by trained ob-servers for children between 2 and 3 yearsof age in a semistructured play environmentto distinguish children with autism from chil-dren with other developmental disorders. Thetest takes about 20 minutes to complete. Pre-liminary data on sensitivity and specificity ofthe STAT screen are excellent but work is con-tinuing. Currently apart from CHAT, none ofthe other tools have been used for populationscreening of young children with autistic dis-order and their performance in this context isnot yet known.

In this article, we describe a model ofearly detection and diagnosis of ASD that isachieved by using the resources of primarycare workers, and a multidisciplinary teamwith experience in assessing children's de-velopmentaJ problem and specific expertisein ASD. A more complete discussion can befound in Chakrabarti and Fombonne (2001),who applied this model in a region in theUK in which 15,500 children were screened.Both early detection and diagnosis may becarried out by this team in collaborationwith parents and primary care professionals(Figure 1).

THE APPROACH IN THEUNITED KINGDOM

Stage 1

In the UK, there has been a long tradition ofsurveillance and screening for medical and de-velopmental problems in young children. Thescreening and surveillance program is carriedout by health visitors, general practitioners(family physicians), and pediatricians. Healthvisitors in the UK are community-based spe-cialist nurses with general nursing and oftenmidwifery or obstetric training followed by

A Model of Early Detection and Diagnosis of Autism Spectrum Disorder

Stage 1

Universal screening/surveillance of young childrenBirth, 6-8 weeks, 6-9 months, 18-24 months, VA-VA years

(UK National Recommendation)

Children identified with a developmental problem

Stage 2

Assessment by Child Development Team (CDT)(Observation of child and elicitation of developmental concerns

from parents and referrer)

203

Problem: Mild/Moderate

a) Mild problem:No extra services neededbut progress monitoredby CDT or referrer.

Unsatisfactory progress

h) Moderate problem:Specific developmentalservices offered, eg, speech therapy,physiotherapy, earlyeducational intervention,progress monitored hy CDT.

Stage 3

Severe or complex problem

Multidisciplinary assessmentand medical investigations:Clinical Diagnosis of ASD

Stage 4

ADI-R & psychometricassessment of children

with clinical diagnosis of ASD

Figure 1. Model of screening and diagnosis of ASD in preschool children.

12 months of specific health visiting train-ing and usually wide experience of workingwith families with young children. A signifi-cant component of the health visiting train-ing consists of theoretic and practical aspectsof working with families with young children

and assessing children's developmental andphysical progress. Health visitors have had ahistoric and crucial role in health and devel-opmental surveillance, anticipatory guidance,health education, advocacy, and supportingfamilies of very young children in the UK.


The program of surveillance and screen-ing begins with the neonatal examination,followed by periodic assessments at 6 to 8weeks, 6 to 9 months, 18 to 24 months, andat 3'A to 3V2 years. The neonatal and 6 to 8weeks' examinations are usually carried outby pediatricians or general practitioners andtend to focus on screening for medical prob-lems, for example, checking for congenitalanomalies, heart disease, dislocation of hips,testicular descent in boys, appropriate weightgain, and head circumference, and elicitingany other parental concerns.

The subsequent screening and surveillanceare most commonly carried out by the healthvisitors and consist of a hearing screening at6 to 9 months and a surveillance program un-til the age of 3V2 years. The surveillance isconducted according to the guidelines of the"Health for All Children" report (Hall, 1996),which emphasizes continuity of care, mak-ing observations, checking history, elicitingparental concerns, offering health advice andguidance, and moving away from prescriptivetests.

Although there is no total uniformity ofpractice, on the whole the 2-year and 3V2-yearhealth visitor checks tend to have 3 distinctcomponents. The health visitor uses the Per-sonal Cbild Health Record (PCHR), which is aunique record for each child and is used uni-versally in the UK to record the child's de-mographic details, birth and medical history,immunization records, and growth and devel-opmental records. The PCHR also containshealth promotion and accident prevention ad-vice for parents and parent questionnaires re-lating to each of the specific developmentalstages at 6 to 8 weeks, 7 months, 18 to 24months, and 3 V4 to 3 V2 years. The parent ques-tionnaire relating to the last 2 time periods,that is, 18 to 24 months and 3'A to 3V2 years,asks parents about their own health, any con-cern about their child's growth, vision, hear-ing, speech, understanding, behavior, walk-ing, immunization, and toilet training. Theother components of the check are a face-to-face assessment of the child and a dialogue

between the health visitor and parents aboutthe result of the check. The direct assessmentof the child by the health visitor takes oneof various forms. It may involve using a stan-dardized developmental screening tool suchas the Denver Developmental Screening Test(DDST; Frankenburg et al., 1992) or a toolcalled Schedule of Growing SkiUs (Bellman,Lingam, & Aukett, 1996) or much more likely,using the template of Mary Sheridan's "Chil-dren's developmental progress" (Sheridan,1973). The important point is that, irrespec-tive of the tool being used, it involves en-gaging the child, face-to-face communicationwith the child, and using some tasks to brieflylook at the child's speech, behavior, social,cognitive, and fine motor skills. The checktakes about 30 minutes to complete and is car-ried out at the local clinic or at home, depend-ing on individual preference and local guid-ance. The dialogue witb parents at the endof the assessment would include a discussionof any concerns raised in the parent question-naire that parents would have already filledin or filled jointly with the health visitor. Onthe basis of this information and the resultof the direct assessment, the parents and thehealth visitor would arrive at an agreed deci-sion about any further course of action if anyproblem is identified.

The popularity and general acceptance ofthese surveillance measures is shown by thefact that in our area the uptake rate of the2-year check is 82% and that of the 3V2-yearcheck is 79% of all eligible children. These areclearly high figures in any general populationsurveillance.

Apart from these scheduled screening andsurveillance, the health visitor often bas moreongoing contact with families who have morethan 1 young child and also families wherean earlier problem or potential problem hadbeen identified. For example, health visitorswould provide closer support and monitoringfor mothers with postnatal depression, fami-lies lacking adequate social support, or par-ents who have learning or mental health prob-lem, or there are child protection concerns or


a child is born prematurely or has a medicalproblem.

The health visitor is also available to lis-ten to, and offer advice about, any specificworries or concerns parents have about theirchild outside the scheduled program of con-tacts, and especially during the nationally rec-ommended immunization sessions when thechildren are 2, 3, 4, and 13 months and3 years old. Two other groups of profes-sionals, apart from the health visitors, areknown to be closely involved with youngpreschool children with actual or potential de-velopmental problems, namely, pediatriciansand speech and language therapists. As de-scribed below, these 3 professional groups,that is, health visitors, pediatricians, andspeech and language therapists, were targetedas the most likely sources of referral of veryyoung children with atypical developmentand in particular of children with potentialASD.

To establish our model, we approachedthese prospective referrers and arranged forbrief training sessions on early identificationand referral of developmental problems. Wealso provided written guidelines for referral.The training program and guidance took theform of publicizing the services available atthe Child Development Center, the differentand complementary mix of expertise availableto assess young children's development, therationale and urgency of early referral of chil-dren who are thought to have developmentalconcerns, and discussion about guidelines forreferral.

The guidance for the referrals was left pur-posefully general and wide to provide maxi-mum sensitivity for ASD case finding. We hy-pothesized that children with ASD, even theones with normal or superior intelligence,would show^ some forms of atypical devel-opment, if not delay as well, and felt thatit was crucial to cast a wide net to findthese children. However, we were also anx-ious not to get overwhelmed with lots of chil-dren with minor or transient developmentaldelays.

To achieve this dual aim, the suggested cri-teria for initial referrals were to include anychild with moderate to severe impairment inone area of development or mild or moderateimpairment in 2 or more areas. Any impair-ment that in the clinical judgment of the pri-mary care worker and/or parents was thoughtto be more than transient was considered assignificant (moderate or severe, according tothe judgment of the referrer). The areas ofdevelopment comprised motor development,speech and language development, socializa-tion and play, behavior, vision, and hearing.The referrers were also advised to refer anychild about whom either parents or profes-sionals had concern, although the concerncould not be described in more specific de-velopmental terms. We put extra emphasis onconcerns expressed by parents as well as onprofessional's clinical judgment, irrespectiveof the results of any screening test used.

Referrals were sought as soon as any prob-lem was identified but tended to clusteraround the specified ages of developmentalsurveillance, mostly around the age of 2 and3 years. All referrals were made with parents'consent.

Stage 2

All parents along with their children re-ferred at this initial stage were invited toa brief meeting lasting about half an hourwith a child development team. The teamconsisted of a pediatrician, a health visitor,speech and language therapist, physical, oc-cupational, and play therapist. The primaryreferrer, frequently a health visitor or speechtherapist, who knew the parents and the chil-dren well, often attended the meeting as well.This was found to be very helpful for parentsand professionals alike. Every effort was madeto encourage the parents to attend the meet-ings, the family health visitor often being in-strumental in helping parents in hard-to-reachfamilies. During the meeting there was op-portunity to observe the child informally inbrief play, encouraged by the play therapist,whereas the team explored the parents' and

206 INFANTS & YOUNG CHILDREN/JULY-SEPTEMBER 2005

the referrer's concerns about the strengthsand weaknesses of the child s development.At the end of the meeting, taking the parents'views and all the information into account,a joint agreement was reached between theparents and the team that the child's progresswas such that

a. no extra services were warranted forthe child's optimal development but thechild's progress needed monitoring or

b. specific developmental services such asspeech therapy, physical therapy, or oc-cupational therapy and/or early educa-tional intervention and close monitoringof the child's development were neces-sary or

c. the nature and/or extent of the child's de-velopmental problems were such that afurther, more in-depth assessment of thechild's development was required.

For cfiildren in pathways a and b, options re-mained open for more in-depth assessment ata later date if on follow-up, this was thoughtnecessary and helpful for the child andfamily.

Stage 3The children who were selected for in-

depth assessment attended the local child de-velopment center in groups of 4 children fora period of 2 weeks. The composition of eachgroup remained constant during the assess-ment period and groups of 4 children were as-sessed consecutively at 2-week intervals. Theassessments were conducted by a multidis-ciplinary team of professionals and spreadover 10 daily sessions of 2 hours each. Dur-ing the assessment, a play therapist led thegroup of 4 children with their participatingparents in structured activities as well as infree play in a specifically built nursery en-vironment. Parents were invited to be withtheir children during these sessions. Mostchildren were accompanied by their moth-ers but a significant number of fathers alsoattended at least some of the sessions. Thechildren were grouped so as to be fairly closein age range to each other, mostly between

the ages of 2 and 3 years, with an occasionalchild below the age of 2 or above the age of4 years.

During play sessions, a particular note w astaken of the children's behavior, social skills,and interaction with the peer group andadults. Note was also made of the children'sspeech and communication in a more nat-uralistic setting than a formal individual as-sessment session. The children were observedfor their motor skills, attention, listening, anddistractibility. Any unusual behaviors, particu-larly unusual sensory behavior (showing dis-tress or covering ears w ith hands to blockout noise or peering at objects through cor-ner of eyes), repetitive behaviors, or stereo-typies (hand flapping, tiptoeing, etc), wererecorded.

A developmental pediatrician experiencedin ASD took a detailed developmental his-tory and conducted a comprehensive medi-cal and neurodevelopmental examination. Allchildren had appropriate biologic investiga-tions according to a standard protocol. Chil-dren were assessed by a speech and languagetherapist experienced in assessing childrenwith ASD and other learning and developmen-tal problems. The children's gross and finemotor skills were assessed by a physical ther-apist and an occupational therapist, respec-tively. The children were also observed andparents interviewed by a specialist nurse inbehavior intervention, with wide experienceof w orking with children and families w ithASD.

The children's hearing was tested by an au-diologic physician and vision screened by anorthoptist. A dietician and nutritionist took adetailed dietary history ofthe children and ob-tained a 3-day food diary from the parents foran analysis of the adequacy of the diet withrespect to intakes of calorie, vitamins, and mi-cronutrients, compared to the recommendedreference nutritional intakes of a child of sim-ilar age. The children's teeth and oral hygienewere inspected by a dental nurse.

All families had a general interview witha social worker who subsequently took up


follow-on work with individual families to ac-cess social care and statutory benefits as ap-propriate and needed by the child and family.At the end ofthe 2-week period of assessment,all the results were collated by the lead pedia-trician and a clinical diagnostic formulation ofthe child's problem was made.

Stage 4

Children who were strongly suspected tohave ASD underwent further assessment witha standardized diagnostic measure, the AutismDiagnostic Interview-Revised (ADI-R), and apsychometric assessment, using the WechslerPreschool and Primary Scale of Intelligence(WPPSI; Wechsler, 1990) or Merrill PalmerScale (Stutsman, 1948).

The ADl-R is a semistructured diagnostic in-terview for use with the parents or main care-givers of the child with possible ASD diag-nosis. The ADI-R algorithm generates scoresfor areas of social interaction, communica-tion (verbal and nonverbal), and repetitive be-havior for which appropriate cutoff pointsfor diagnosis are available. The ADI-R algo-rithm is compatible with the Diagnostic andStatistical Manual of Mental Disorders, 4thedition (D5Af-/V), and International Classifi-cation of Diseases, 10th edition (ICD-10), di-agnostic criteria.

All but a few children with ASD had aformal psychometric assessment. As childrenwith ASD vary enormously in their intellectualand verbal ability, 2 specific instruments wereused for psychometric testing. The WPPSIwas used for verbal children and MerrillPalmer Scale was used for nonverbal children.

OUTCOMES OF THE SCREENINGPROGRAM

Detailed results and their likely inter-pretation are available in Chakrabarti andFombonne (2001). In brief, of the 576 chil-dren referred to the Child Development Cen-ter following stage 1 assessment, 97 chil-dren were diagnosed with PDD (ASD) at thecompletion of stage 4, resulting in a preva-

lence estimate of 62.6 (95% CI, 50.8-76.3) per10,000 children for aU PDDs (ASDs). Of the97 children, 26 were diagnosed with autisticdisorder (16.8 per 10,000), 13 with Aspergersyndrome (8.4 per 10000), 1 girl with Rettsyndrome, 1 boy with childhood disintegra-tive disorder, and 56 children with PDDNOS(36.1 per 10000).

Of the 97 children with PDD, 34% of thechildren were referred by pediatricians, 32.9%by speech and language therapists, 21.6%by health visitors, and 5.1% by family physi-cians (general practitioner in the UK), and6.2% were from various other sources. How-ever, closer inspection of the referral patternshowed that most of the referrals to pediatri-cians and speech therapists were initiated byhealth visitors. Taking these data into account,it was estimated that 79 (81%) of the 97 chil-dren were originally identified by health vis-itors as children with a developmental prob-lem requiring further assessment.

The average age oi referral oi the childrenwas 357 months (range, 11-63 months), andaverage age at initial diagnosis was 41 months(range, 21-78 months). A statistically signif-icant difference was found in both age ofreferral and age of diagnosis depending on di-agnostic category. Children with autism w erereferred (mean age of referral, 30 months)and diagnosed (mean age of diagnosis,34.6 months) earlier than children withPDDNOS (mean age of referral, 37.2 months;mean age of diagnosis, 43.1 months) who,in turn, were referred and diagnosed earlierthan children with Asperger syndrome (meanage of referral, 47.5 months; mean age ofdiagnosis, 51.8 months).

Of the 97 children, 77 were boys (79.4%)and 20 girls (20.6%). Ninety-three of the 97children were administered psychometric as-sessment. Of these, 24 children (25.8%) hadsome degree of mental retardation.

IMPUCATIONS FOR PRACTICE

In this article, we have described a modelof detection and diagnosis of autism and

208 INFANTS & YOUNG CHILDREN/JULY-SEPTEMBER 2005

ASD in young children (children between 2and 4 years of age) which has been usedin the UK context. The niodel we have de-scribed is broadly in line with the main rec-ommendation of the consensus panel (Filipeket al, 1999) on screening and diagnosis ofautism except the use of an autism-specificscreening instrument. We feel that the cur-rently used ASD screening instruments may betoo restrictive in identifying the broad rangeof autistic disorders in young children.

Understandably, the common target of allASD screening instruments is the identifica-tion of autism-specific behaviors and develop-mental patterns in children. Although thereis extensive information and broad agreementabout the presentation, range of behavior,and developmental patterns of children withautism, there is much less research and in-formation about the antecedents of childrenwith eventual diagnosis of milder variants ofASD such as PDDNOS (pervasive develop-mentai disorder not otherwise specified) orAsperger syndrome.

Because of this, we felt that it is importantnot to assume that these children with ASDs,but not autism disorders, would also presentwith behavioral features that have been com-monly targeted in autism screening instru-ments. On the contrary, our view is that it isimportant to keep a broader perspective andtarget any atypical development in children assubjects of interest.

A further difficulty in consistent and com-prehensive detection and diagnosis of ASDis that despite the overall prevalence of thecondition being much higher than previouslythought, it is nevertheless a small proportionof all children whose development may causesome concern. It is estimated that 12% to16% of all children present with some de-velopmental or behavioral difficulty (Boyle,Decoufle, & Yeargin-Allsopp, 1994). Of thesechildren, only about 4% to 5% would be ex-pected to have an ASD. Hence, searchingfor children with ASD is like the prover-bial search for a needle in a haystack, al-though it is a larger needle than previouslythought.

Autism spectrum disorder may occur inchildren of all social classes and economicand demographic backgrounds. Apart from asmall number of recognized medical condi-tions (tuberous sclerosis. Fragile X syndrome,etc) and siblings of index cases of ASD, thereare no other known risk factors that canbe targeted and lend themselves to selectivescreening.

As a consequence, effective strategies forgeneral population screening are necessary toidentify all cases of ASD. We have shown thatwith uptake rates of around 80%, the healthvisitor young child surveillance provides anideal opportunity to sample preschool chil-dren's developmental progress over a courseof time. It also must be remembered that apartfrom the scheduled surveillance, the healthvisitor also carries out some opportunisticsurveillance, particularly in hard-to-reach fam-ilies, and these may not be counted in the sta-tistical record of the surveillance figures. Weare reasonably confident that the true uptakeofthe surveillance is higher than the recordedfigures.

The other important aspect of the surveil-lance is that the process is repeated to sam-ple 2 crucial ages in the children's early de-velopment, at 2 years and 3V2 years. In ourexperience, repeated sampling of children'sdevelopment over a course of time duringchildren's early development is crucial tomaximize case finding in ASD. In children sub-sequently diagnosed with ASD, parents oftenrecognize that "something is not right in theirchildren's development" by the age of 18 and20 months (Di Giacomo & Fombonne, 1998).How ever, this is by no means universal andthere are children with ASD whose devel-opment does not arouse concern in theirparents until the age of 2 or 3 or even4 years. Screening instruments designed foruse at single age points are likely to miss thesechildren.

One of the criticisms of the feasibility ofour model is that it is dependent on a pro-fessional group like health visitors for initialdetection of children with developmentalproblems. Without the health visitors the


success ofthe model would be compromised.However w e would suggest that in countrieswithout the tradition of a health visiting pro-fession, the task of identifying children withdevelopmental problem would fall on otherprimary care workers, be it pediatricians,family physicians who look after children,play group leaders, nursery or kindergartenteachers, or children's early interventionprofessionals.

There is now a whole range of validated,high-quality screening instruments includingParents' Evaluations of Developmental Sta-tus (Glascoe, 1998) for primary identificationof developmental and behavioral problem inyoung children. These could possibly take theplace of health visitor screening. Whether theuptake rates and outcomes from these screenswould be comparable to the screening we de-scribed remains to be seen. There may also besome concern that the suggested 2-week mul-tidisciplinary assessment in our model is toolong for parents to be able to afford, particu-larly in countries where both parents are morelikely to be working. The model may also bethought to be expensive in professional timeand resources.

On the question of parent participation, wefind it invaluable to arrive at an informed,shared understanding of the child which par-ents are comfortable with if at least 1 parentis present during the assessment. This is be-cause the parents actively participate in theassessment process and have a continuous di-alogue w ith the professionals so that the out-come of the assessment is not a rude sur-prise. When the parents are already attuned

to a diagnostic possibility of an ASD, it is eas-ier for parents to identify and become moreknowledgeable about the strengths and de-velopmental problems of their child duringthe assessment. However, as pointed out byCharman and Baird (2002), "the difficultiesof recognition, belief and acceptance are farfrom easy when the professional is givingcompletely unexpected information."In thesecases, parents tell us that they find the 2-weekassessment extremely helpful to attune them-selves and come to some understanding as towhat it is in their child's development that iscausing concern.

Although we welcome both parents andgrandparents if they share the child's careand sometimes siblings, it is mostly motherswho attend all the sessions. In our experi-ence it is extremely rare for parents to de-cline or withdraw from an assessment. Re-garding the cost of the model, we feel that itis likely to be more than compensated in fi-nancial term by the cost saving in better out-comes for the child and family (due to earlyintervention) and less dependence on costlyservices from the state. More than this, thesavings in human terms in reducing frustra-tion and anguish of both parents and childand interminable wait for parents to come toan understanding of their children's difficul-ties may themselves be very significant. Thisis also what parents want (Howlin & Moore,1997).

On all these counts, we feel that our modelhas much to commend itself for comprehen-sive early detection and diagnosis of autismand ASD.

REFERENCES

American Psychiatric Association. (1994). Diagnosticand statisticai manuai oftnentai disorders (4th ed.).Washington, DC: Author.

Baird, G., Charman, T, Baron-Cohen, S., Cox, A., Sweet-enham, J., Wheelwright, S., et al. (2000). A screeninginstrument for autism at 18 months of age: A 6-yearfollow-up study. Journal of the American Academyof Child and Adolescent Psychiatry, 39, 694-702.

Baron-Cohen, S., Allen, J., & Gillberg, C. (1992). Canautism be detected at 18 months? The needle, the

haystack and the CHAT. British Journal of Psychiatry,161, 839-843.

Bellman, M., lingam, S., & Aukett, A. (1996). Schedule ofgrowing skills. Windsor, UK: NFER-Nelson.

Boyle, C. A., Decoufle, P., & Yeargin-Allsopp, M. (1994).Prevalence and health impact of developmental dis-abilities. Pediatrics, 93, 863-865.

Chakrabarti, S., & Fombonne, E. (2001). Pervasive devel-opmental disorders in preschool children._^M4, 285,3093-3099.


Charman, T, & Baird, G. (2002). Practitioner review: Diag-nosis of autism spectrum disorder in 2- and 3-year-oldchildren. Journal of Child Psychology and Psychia-try, 43, 289-305.

Cunningham, C. C, Morgan, P A., & McGucken, R. B.(1984). Down's syndrome: Is dissatisfaction withdisclosure of diagnosis inevitable. DevelopmentalMedicine & Child Neurology, 26, 33-39.

Di Giacomo, A., & Fombonne, E. (1998). Parental recog-nition of developmental abnormalities in autism.European Child and Adolescent Psychiatry, 7, 131-136.

Filipek, P A., Accardo, P J., Baranek, G. T, Cook, E. H., Jr.,Dawson, G., Gordon, B., et al. (1999). The screeningand diagnosis of autistic spectrum disorders. Journalof Autism and Developmental Disorders, 29, 439-484.

Fombonne, E. (2003). The prevalence of autism [Edito-rial].yAM4, 259(1), 87-89.

Frankenburg, W. K., Dodds, J., Archer, P, Shapiro, H., &Bresnick, B. (1992). The Denver II: A major revisionand restandardisation of the Denver DevelopmentalScreening Test. Pediatrics, 89, 91-97.

Freeman, B. J., Ritvo, E. R., Guthrie, D., Schroth, P, «&Ball, J. (1978). The Behaviour Observation Scale forautism: Initial methodology, data analysis and prelimi-nary findings on 89 children . /ouma/ of the AmericanAcademy of Child Psychiatry, 17, 576-588.

Glascoe, E P (1998). Collaborating with parents: Usingparents' evaluation of developmental status to detectand address developmental and behavioural prob-lems. Nashville, TN: Ellsworth & Vandermeer

Guralnick, M. J. (Ed.). (1997). The effectiveness of earlyintervention: Second generation research. Baltimore:Paul H. Brookes.

Hall, D. (Ed.). (1996). Health for all children: Reportof the Third Joint Working Party on Child HealthSurveillance (3rd ed.). Oxford, England: Oxford Uni-versity Press.

Howlin, P., & Moore, A. (1997). Diagnosis in autism: Asurvey of over 1200 patients in UK. Autism, 1, 135-162.

Hoyson, M., Jamieson, B., & Strain, P S. (1984). Individu-alized group instruction of normally developing andautistic-like children: The LEAP curriculum model.Journal ofthe Division of Early Childhood, 8, 157-72.

Keating, D. P, & Hertzman, C. (1999). Developmentalhealth and the wealth of nations: Social, biologicaland educational dynamics. New York: The GuilfordPress.

Krug, D. A., Arick, J., & Almond, P (1980). Behaviourchecklist for identifying severely handicapped individ-uals with high levels of autistic behaviour./oMm«/ ofChild Psychology and Psychiatry, 21, 221-229.

Lord, C, Rutter, M., & Le Couter, A. (1994). Autism Diag-nostic Interview—Revised: A revised version of a di-agnostic interview for caregivers of individuals with

possible pervasive developmental disorders. Journalof Autism and Developmental Disorders, 24, 659-685.

Lord, C, Risi, S., Lambrecht, L., Cook, E. H., andothers (2000). The Autism Diagnostic ObservationSchedule—Generic: A standard measure of social andcommunication deficits associated with the spectrumof autism./oMma/ of Autism and Developmental Dis-orders, 30, 205-223.

McEachin,J. J., Smith, T., &Lovaas, I. O. (1993). Long-termoutcome for children with autism who received earlyintensive behavioural treatment. American Journal ofMental Retardation, 97, 359-372.

National Research Council. (2001). Educating childrenwith autism. Washington, DC: National AcademyPress.

Nelson, C. A. (2000). The neurobiological bases of earlyintervention. In Shonkoff & Meisels (Eds.), Handbookof early childhood intervention (2nd ed., pp. 204-230). Cambridge, MA: Cambridge University Press.

Quine, L., & Pahl, J. (1987). First diagnosis of severe men-tal handicap: A study of parental reactions. Develop-mental Medicine and Child Neurology, 29, 232-242.

Rimland, B. (1968). On the objective diagnosis of infantileautism. Acta Paedopsychiatrica, 35, 146-161.

Robins, D. L, Fein, D., Barton, M. L., & Green, J. (2001).The Modified Checklist for Autism in Toddlers: An ini-tial study investigating the early detection of autismand pervasive developmental disorders. Journal ofAutism and Developmental Disorders, 31, 131-144.

Rogers, S. J., & Lewis, H. (1989). An effective daytreatment model for young children with pervasivedevelopmental disorders. Journal of the AmericanAcademy of Child and Adolescent Psychiatry, 28,207-214.

Ruttenberg, B. A., Dratman, R., Frakner, T. A., & Wenar, C.(1966). An instrument for evaluating autistic children.Journal ofthe American Academy of Child Psychia-try, 5, 453-478.

Rutter, M., Silberg, J., O'Connor, T, & Simonoff, E. (1999).Genetics and child psychiatry: II. Empirical researchfindings. Journal of Child Psychology and Psychia-try, 40, 19-55.

Schopler, E., Reichler, R. j . , & Renner, B. R. (1988). TheChildhood Autism Rating Scale (CARS). Los Angeles:Western Psychological Services.

Seigel, B. (1998). Early screening and diagnosis inautism spectrum disorders: The Pervasive Devel-opmental Disorder Screening Test (PDDST). Paperpresented at the NIH State of the Science inAutism Screening and Diagnosis Working Conference,Bethesda, MD.

Sheridan, M. D. (1973). Children's developmentalprogress, from birth to five years: The Stycar se-quences. Windsor, CT: NFER Publishing Company

Shonkoff, J. P, & Meisels, S. J. (2000). Handbook ofearly childhood intervention (2nd ed.). New York:Cambridge University Press.


Stone, W, L,, Coonrod, E, E,, & Ousley, O, Y. (2000). Briefreport: Screening Tool for Autism in Two-year-olds(STAT): Development and preliminary data, Joumalof Autism and Developmental Disorders, 30 607-612,

Stutsman, R. 0948}. Merrill Palmer Scale of Mental Tests(Preprints of Part III, Mental Measurement of Pre-school Children). Chicago: Stoelting.

Watson, L, R,, Baranek, G. T, & Di Lavore, P. C. (2003).Toddlers with autism, developmental perspectives. In-fants and Young Children, 16, 201-214.

Wechsler, D. (1990). Manual for Wechsler preschooland primary scale of intelligence—Revised (BritishAmendments). Kent, England: The PsychologicalCorporation,

Wing, L,, & Gould, J, (1978). Systematic recording of be-

haviours and skills of retarded and psychotic children.Joumal of Autism and Childhood Schizophrenia, 8,79-97.

Wing, L, Leekam, S. R,, Iibby, S.J,, Gould,J., & Larcombe,M. (2002). The diagnostic interview for social andcommunication disorders: Background, inter-rater re-liability and clinical use. Joumal of Child Psychologyand Psychiatry, 43, 307-325,

Wing, L,, & Potter, D. (2002). The epidemiology of autisticspectrum disorders: Is the prevalence rising? MentalRetardation & Developmental Disabilities ResearchReviews, 8O), 151-161,

World Health Organization, (1992). The ICD-IO classifi-cation of mental and behavioural disorders: Clini-cal descriptions and diagnostic guidelines. Geneva:Author,

vol. 18, no. 3, pp. 200-211 a model of early detection and...

Documents