Autism in Young Children

SCREENING FOR AUTISM IN YOUNG CHILDREN:THE MODIFIED CHECKLIST FOR AUTISM IN

TODDLERS (M-CHAT) AND OTHER MEASURES

Thyde Dumont-Mathieu1,2* and Deborah Fein21Department of Pediatrics, University of Connecticut School of Medicine, Farmington, Connecticut

2Department of Psychology, University of Connecticut, Storrs, Connecticut

The literature on the importance of early identication and earlyintervention for children with developmental disabilities such as autismcontinues to grow. The increased prevalence of autistic spectrum disordershas fostered research efforts on the development and validation of autism-specic screening instruments for use with young children. There are cur-rently several such autism-specic screening tools meant to be used withyoung children in various stages of development. Data from a few of thesescreening instruments have been published, and they include the Checklistfor Autism in Toddlers (CHAT), Pervasive Developmental Disorders Screen-ing Test (PDDST), Screening Tool for Autism in Two year olds (STAT), Check-list for Autism in Toddlers-23 (CHAT-23), and the Modied Checklist forAutism in Toddlers (M-CHAT). In this review, these ve tools designed foruse with children under three years old will be highlighted. In particular, theModied Checklist for Autism in Toddlers (M-CHAT) will be discussed.

2005 Wiley-Liss, Inc.MRDD Research Reviews 2005;11:253262.

Key Words: autism; pediatric screening; early identication; early inter-vention

BACKGROUND

Autism is a neuro-developmental disorder rst describedby Leo Kanner in 1943. It is one of a group of disordersknown as Pervasive Developmental Disorders (PDDs),now more commonly referred to as Autistic Spectrum Disorders(ASDs). This group of disorders includes Autistic Disorder,Pervasive Developmental Disorder-Not Otherwise Specied(PDD-NOS), Aspergers syndrome, Rett syndrome, and Child-hood Disintegrative Disorder (CDD) (American Psychiatric As-sociation (APA), 1994). Autistic Spectrum Disorders are cur-rently estimated to affect somewhere between 1 in 166 and 1 in1000 children [Baird et al., 2000; Bertrand et al., 2001;Chakrabarti and Fombonne, 2001; Yeargin-Alsopp et al., 2003Volkmar et al., 2004; Barbaresi et al., 2005; Chakrabarti andFombonne, 2005].

The diagnostic criteria for autistic disorder require impair-ment in three areas: reciprocal social interaction, communica-tion, and specic patterns of behavior, interests, and activities[APA, 1994]. Research has documented areas of dysfunctionrelated to these core categories of symptoms in autism: (1) socialabilities (attachment, joint attention, social imitation, orientingto social stimuli, face and affect processing, expression of emo-tion, and symbolic play) [McEvoy et al., 1993; Dawson et al.,

1998; Klin et al., 2002; Klinger et al., 2003]; (2) communicativeabilities (use of language, both verbal and non-verbal, quality ofcommunication and play) [Travis et al., 2001; Klinger et al.,2003; Dawson et al., 2004] and (3) restricted or stereotypedbehaviors; interests, and activities (non-functional routines orritualistic behaviors, resistance to changes in the environment,repetitive motor mannerisms, unusual interests and preoccupa-tions; and visual fascinations) [Volkmar et al., 1986; Szatmari etal., 1989; Campbell et al., 1990; Turner, 1999; Charman andSwettenham, 2001; Klinger et al., 2003]. There are additionalsymptoms that are sometimes present with ASDs, these includeself injurious behaviors, functional disturbances such as difcul-ties with sleeping and eating, abnormal fears, and abnormalresponses to sensory stimuli [Klinger et al., 2003]. Cognitivefunctioning can range from severe mental retardation to aboveaverage intelligence quotient (IQ), with many children whohave an ASD showing some degree of cognitive impairment. Arecent study by Chakrabarti and Fombonne [2005] found that ina sample of four to six year olds diagnosed with an ASD, nearlya third had signicant cognitive impairment.

Previous studies have found that regression in ASD occursmost often between 18 and 24 months of age, with the fre-quency of the regressive type of autism being somewhere be-tween 10 and 50% [Tuchman and Rapin, 1997; Goldberg et al.,2003; Lord et al., 2004; Rogers, 2004]. Clinical experience ofthe present authors suggests that regression often occurs sooner,sometimes as early as 1218 months. It is possible that parentsmay recall 1824 months as the age of regression for childrenwho were diagnosed later, or that parents are experiencing recallbias [Siperstein and Volkmar, 2004]. It is possible that as the ageat diagnosis drops with earlier identication, so will the reportedage at regression.

Contract grant sponsor: National Institute of Child Health and Development(NICHD); Contract grant number: 5-R01 HD39961.*Correspondence to: Thyde Dumont-Mathieu, M.D., M.P.H., Department of Psy-chology, University of Connecticut, 406 Babbidge Rd., Storrs, CT 06269-1020.E-mail: [email protected] 11 July 2005; Accepted 12 July 2005Published online in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/mrdd.20072

MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIESRESEARCH REVIEWS 11: 253262 (2005)

2005 Wiley-Liss, Inc.

There are many important theoret-ical and clinical issues related to the sub-ject of regression in autistic spectrum dis-orders. Do parents and providers meanthe same thing when they discuss regres-sion? What constitutes regression as op-posed to developmental stagnation? Is theregression seen with autism similarmechanistically to that seen in other re-gressive disorders such as Landau-Klef-fner syndrome [Ballaban-Gil and Tuch-man, 2000; Robinson et al., 2001;Stefanatos et al., 2002 Trevathan, 2004],or to disorders within the autistic spec-trum, Childhood Disintegrative Disorder[Volkmar and Rutter, 1995; Malhotraand Gupta, 1999 Goldberg et al., 2003;Manning-Courtney et al., 2003], andRett syndrome [Glaze, 2004]? Is the out-come different for those children whohave had a normal developmental trajec-tory prior to the onset of regression, ascompared with those whose regression islimited only to language or those whowere developmentally delayed even be-fore the regression? Does the timing ofthe regression impact outcome? Does au-tism with regression differ in etiology,impact, and prognosis than autism with-out regression?

Various mechanisms of regressionhave been suggested, including illness,sub-clinical seizures, and variations inpostnatal brain development due eitherto the impact of genetic factors or post-natal events [Lainhart et al., 2002;Hrdlicka et al., 2004]. A discussion ofthese questions and issues is beyond thescope of this article and the readers arereferred to a review of developmentalregression published in this journal[Rogers, 2004]. However, there are twoissues pertinent to this discussion ofscreening young children. Screeningtools need to ask about regression in away that is compatible with parent andprovider perspectives. It is also importantto not only ask about regression, but lackof developmental progression. Secondly,it is important to remember that autismscreening will not capture regression ifthe screening precedes the regression.

Retrospective studies have shownthat children who are later diagnosedwith autism exhibit symptoms as early as812 months of age [Osterling and Daw-son, 1994 Baranek, 1999; Werner et al.,2000; Osterling et al., 2002]. In generalthese studies reviewed videotapes of thechildren prior to or at their rst birthdayand analyzed their behaviors, as com-pared with typically developing childrenor children with non-ASD diagnoses.Some children with ASDs seem to ex-hibit symptoms as early as 8 months, but

more consistently around 12 months.The identied symptoms are fairly con-sistent and include delay or absence inpointing, showing objects, looking atothers, and orienting to their name [Os-terling and Dawson, 1994]. In additionto the difculties with aspects of socialresponsiveness (poor visual orientation/attention, response to name) these chil-dren also seem to exhibit sensory-motordifculties (mouthing objects excessively,aversions to social touch) [Baranek,1999]. Other symptoms that seem to in-dicate possible ASD include stereotypedplay with objects, unusual posturing ofbody parts, looking at a camera less fre-quently, staring/xating on objects, andhaving less animated affective expressions[Baranek, 1999]. Lack of response toname seems to be the most consistentsingular symptom at this young age [Os-terling and Dawson, 1994; Baranek,1999; Werner et al., 2000 Osterling etal., 2002].

The ndings from these studies, al-though retrospective, document some ofthe behavioral signs of autism in veryyoung children, and suggest that earlyscreening might be feasible. Ongoing re-search is needed to uncover proles ofchildren with ASDs at young ages, sincethe signs may be different from thoseseen in older children. Prospective stud-ies are needed with infants and toddlersto further investigate the best means ofdifferentiating between variants of typicaldevelopment and the earliest symptomsof autism. In the meantime, it wouldseem prudent to verify that children un-der 12 months old being seen for theirwell child care visit are looking at othersand orienting to their name, since thesetwo skills may be two of the earliest redags for possible autism. Baranek [1999]states that the early markers for autismnoted on videotape review preceded thevoicing of concerns by parents. She alsomentions that the parents in her studycompensated for these markers, suggest-ing that these compensatory behaviors onthe parents part may somehow be reec-tive of the childs symptoms and a possi-ble avenue for developing measures tohelp with the early identication of thesechildren.

Despite this delay in parents con-cerns relative to the early markers iden-tied retrospectively, studies suggest thatparental concerns regarding their childsdevelopment were typically expressed topediatricians by the age of 1.5 years, andyet, a denitive diagnosis of autism is notmade until approximately 4 years old.[De Giacomo and Fombonne, 1998; Sie-gel et al., 1988; Flannagan and Nuallain,

2001]. Simultaneously, the evidence sug-gesting that young children as young as24 months old can be diagnosed withautism continues to grow [Dahlgren andGillberg, 1989; Lord, 1995; Stone et al.,1999; Chakrabati and Fombonne, 2001;Zwaigenbaum, 2001; Charman andBaird, 2002; Dawson et al., 2002]. Fur-thermore; the neurobiological abnormal-ities noted in the child with ASD in therst two years of life [Courchesne et al.,2004] support the idea that as early as 1year of age, children with autism mayshow signs of developing quite differ-ently from their typically developingpeers.

Although discussions continue asto whether and why autism seems to beon the upswing, awareness has certainlyincreased [Fombonne, 2003]. Severalstudies have demonstrated that early de-tection and early intervention do have apositive impact on outcomes for childrenwith autistic spectrum disorders [Hoysonet al., 1984; Lovaas, 1987; Rogers andLewis, 1989; Harris et al., 1991; Birn-brauer and Leach, 1993; McEachin et al.,1993; Lord, 1995; Dawson and Oster-ling, 1997; Smith et al., 1997; Jocelyn etal., 1998; Sheinkopf and Siegel, 1998;Smith and Lovaas, 1998; Harris andHandleman, 2000; Sonnander, 2000]. Infact research suggests that interventionbefore three or three and a half years ofage has the greatest impact. [Harris andHandleman, 2000; Woods and Weth-erby, 2003]. The ndings that early iden-tication and intervention lead to im-proved outcome suggest that autism is anappropriate disorder for which screeningshould be undertaken. This realizationhas prompted a focus on the develop-ment of autism-specic screening toolsappropriate for use with young children.

Several issues regarding the screen-ing for autism in young children needfurther clarication. What type ofscreening is best: should the screening bebroad-based developmental screening, orautism-specic? What is the best age atwhich to screen? What is the bestmethod for screening: should parent re-port be used, or is direct observation abetter approach?

WHAT TYPE OF SCREENINGIS BEST?

Since research suggests that the rstindication of a problem in children whoare subsequently diagnosed with an ASDmay occur as early as the rst year of life[Osterling and Dawson, 1994; Baranek,1999; Werner et al., 2000; Osterling etal., 2002] and that in the U.S., earlyintervention is not only an accepted but

254 MRDD RESEARCH REVIEWS SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES DUMONT-MATHIEU & FEIN

mandated means of addressing the needsidentied for children from birth to theage of three, screening all children onceor twice during that timeframe wouldfacilitate the timely identication of chil-dren who need early intervention and/ormonitoring.

There are generally two modelsused to screen for autism. The rst modelis consistent with practice parameters en-dorsed by the American Academy of Pe-diatrics (AAP), as well as the AmericanAcademy of Neurology (AAN) [Filipeket al., 1999; Filipek et al., 2000AAP,2001a, 2001b, 2001c]. It consists of on-going developmental surveillance of chil-dren within the primary care setting. InDworkin [1993], developmental surveil-lance is dened as:

a exible, continuous process whereby knowl-edgeable professionals perform skilled observationsof children during the provision of health care. Thecomponents of developmental surveillance includeeliciting and attending to parental concerns, obtain-ing a relevant developmental history, making accu-rate and informative observations of children, andsharing opinions and concerns with other relevantprofessionals (p. 533).

It is suggested that standardizedgeneral developmental screening tools beused as part of the developmental surveil-lance process. These tools can either be aparent-completed questionnaire or clini-cian-completed measure. If generalscreening raises concerns suggestive ofpossible autistic spectrum disorders, au-tism-specic screening could then be ad-ministered. The challenge of provider-implemented screening was highlightedby a recent study, which found that halfof the surveyed physicians used a formalscreening tool (Sices et al., 2003). How-ever, a primary benet of using thismodel to screen for autism is that it in-corporates autism screening into an al-ready accepted practice within the eldof pediatrics. As such, autism-specicscreening is less likely to represent anadditional item to be included in thealready overburdened well child care(WCC) visit. Another benet is that itallows the screening to be undertakenwithin the context of the childs medi-cal home.

The term medical home was rstused in 1967 [Sia et al., 2004]. TheAmerican Academy of Pediatrics (AAP)in a policy statement provided 7 dimen-sions (accessible, continuous, compre-hensive, family-centered, compassionate,culturally effective, and coordinated withspecialized services provided outside ofthe primary care setting) that should befound within a medical home [AAP,2002; Bethell et al., 2004]. As a result of

this type of care, the child is viewed asthe multi-faceted being that he or she is.The fact that any identied concerns willbe raised by a provider with whom thefamily and child already have a relation-ship may lend additional credibility to thescreening process and ndings. Thereare, however, potential limitations withthis model. It depends on the primarycare provider performing developmentalsurveillance/screening, recognizing thered ags for autism, and pursuing theappropriate next steps in a timely man-ner. This possibility is especially prob-lematic as the frequency of well-childcare visits begins to decrease in the tod-dler years. Also, a potential challengewith this approach is the limited amountof time available to most providers foraccomplishing a growing number of tasksduring the WCC visit. Perhaps due inpart to this, in some instances, the di-mensions of the medical home are notfully implemented [Strickland et al.,2004].

Although there are currently novalidated autism-specic screening in-struments designed for children less than18 months old, several groups are work-ing to better understand how autism pre-sents in that age group in a prospectivemanner. Zwaigenbaum et al. [2005] re-cently published data from their ongoingprospective study with siblings of chil-dren diagnosed with an ASD. They havedesigned an observational scale for assess-ing autism-specic behavior in children.This is a promising line of research whosendings may lead to validated measuresfor screening and diagnosing children lessthan 18 months of age, or even less than12 months of age. At the present time,providers should use validated generaldevelopmental surveillance/screeningwith their patients under 18 months old.

The second model consists of theroutine administration of autism-specicscreens at multiple high-risk ages (e.g., 18and 24 months of age) by primary careproviders, regardless of the presence ofsymptoms or concerns suggestive of anASD. This would be done in addition todevelopmental surveillance, includingthe use of general developmental screen-ing. The primary benet of screening atseveral ages is that the likelihood of notrecognizing a child with possible ASD isdiminished. There are three main limita-tions to this approach. First, it wouldrequire already busy primary care provid-ers to remember to administer an autism-specic screening instrument to all chil-dren, even those who they feel aredeveloping appropriately. Providers mayhesitate to raise the possibility of a disor-

der such as autism with every parent theyencounter at a preset interval due to con-cerns that the mere administration ofthese screens will be anxiety provokingfor some parents. Secondly, the questionarises as to which instrument primarycare providers should use. Screening in-struments for children younger than1618 months have not yet been vali-dated, although multiple research groupsare working to identify reliable signs ininfancy. Multiple versions of a screenerwould be needed for children of differentages. Lastly, as the number of screeninginstruments for different conditions in-crease, primary care providers may beunable to t them all into the limitedtime available for the well-child carevisit.

Research suggests that the age atdiagnosis seems to be inuenced by so-cioeconomic status and/or race/ethnicity[Mandell et al., 2002]. Furthermore,missed and misdiagnosed ASDs seem tobe more prevalent in certain populations[Mandell et al., 2002; Dyches et al.,2004; Pinto-Martin and Levy, 2004].The cause for this is not yet known, butmay be linked to several factors related toracial and ethnic disparities in the qualityof health care [Smedley et al., 2003],including screening and referral practices.There is also the possibility that thesymptoms of autism may differ acrosspopulations, not only in terms of whichsymptoms are present and the timing ofpresentation, but also which symptomsconcern parents, and whether they sharethese concerns with pediatric providers.Providers need to ensure that all childrenin their care are being screened. Towardsthat end, whichever model is used toscreen for autism should include prede-termined steps. Empirical data is neededto determine which model is most effec-tive. It is possible that different modelswill work best in different patient andprovider populations, and at differentpoints in development.

WHAT IS THE BEST AGE TOSCREEN?

There are both advantages and dis-advantages to early screening. On onehand, earlier screening may yield a diag-nosis and subsequent intervention at anearlier age, allowing the maximum timefor intervention. On the other hand, au-tism-specic screening at less than 1618months old has not yet been validated.Additionally, children who regress orthose with Aspergers syndrome and highfunctioning autism may be missed byearly screening, as they often present

255MRDD RESEARCH REVIEWS SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES DUMONT-MATHIEU & FEIN

later. Furthermore, health care providersmay be reluctant to screen for autism atyounger ages because of concerns thatsuch screening will lead to parental anx-iety. Perhaps contrary to this idea is thending that signicant numbers of par-ents have concerns about how theiryoung child is developing, but if they arenot asked, they may not share these con-cerns with their childs primary care pro-vider [Young et al., 1998; King and Glas-coe, 2003; Bethell et al., 2004]. A coreprinciple embedded in the concepts ofdevelopmental surveillance, the medicalhome, and early identication is elicitingand responding to parental concerns.Studies have demonstrated the impor-tance of parental report in the early de-tection of developmental problems[Stone and Lemanek, 1990; Stone et al.,1994; Glascoe and Dworkin, 1995;Young et al., 2003]. It is important thatproviders respond to parental concernsand yet, a recent study found that expres-sion of parental concerns did not lead toincreased likelihood of the child beingreferred for the purpose of diagnosis orservice provision [Sices et al., 2004].

Screening at later ages, such asscreening after 1824 months, may leadto missed opportunities for early inter-vention. However, this later screeningwill likely have better specicity and sen-sitivity. In addition, it may capture thechildren who develop autistic features atlater ages, those who regress, or arehigher functioning. Also, health careproviders may feel more comfortablewith the idea of screening older children.Another challenge to early diagnosis isthe possibility that the provider may beconcerned when the parents are not.Since participation in early interventionis voluntary, unless the parents agree toreceive the services, the child will nothave the opportunity to receive them.Some parents may even refuse to allowtheir child to undergo the evaluative pro-cess. In such instances, repeated screen-ings may increase the likelihood that theparents will be convinced that the con-cerns being identied are worth furtherinvestigation, and screening at an olderage may be less likely to encounter pa-rental denial of concern.

In deciding on what age is mostappropriate for autism-specic screening,it is important to remember that differentitems may be needed for screening atdifferent ages. For example, some studiessuggest that toddlers diagnosed with au-tism and those with developmental delaysboth have impaired pretend play [Baron-Cohen et al., 1996; Charman et al., 1998;Cox et al., 1999]. Yet, when preschool

aged children with autism are comparedwith those with developmental delays,children diagnosed with autism are foundto exhibit less pretend play skills thanthose with developmental disabilities[Lord et al., 1994; Cox et al., 1999; Not-erdaeme et al., 2000].

To further complicate the issue,many typically developing young tod-dlers may show behaviors consistent withthose seen in ASD, such as repetitivebehaviors (e.g., opening and closingdrawers and doors, turning lights on andoff), repetitive motor behaviors such ashand apping when excited, and insis-tence on routines (and indeed a certainsense of consistency and routine is crucialto minimizing frustration in that agegroup).

WHAT IS THE BEST METHODFOR SCREENING?

Several factors must be consideredin contemplating the best approach toautism screening. These include theavailability, cost, ease of administration ofinstruments with acceptable sensitivity/specicity, timing of well-child care visitsfor children, the amount of time availablefor these visits, the general approach todevelopmental screening and use of non-physician personnel (e.g., nurse practitio-ners, physical assistants, home visitors),cost-effectiveness, reimbursement, andthe cultural expectations of the popula-tion served regarding screening and diag-nosis both in terms of the process and thecontent.

Two methods often used in devel-opmental screening are parent reportsand observations by trained clinicians;each has advantages and drawbacks. Pa-rental report is encouraged by many pri-marily because parents know their chil-dren best. They spend the most timewith them in a variety of settings. This,therefore, provides them with many op-portunities to observe the child acrossmultiple settings, interacting with variousindividuals while they are in a variety ofmoods. This method, then, does not relyon a brief and what may be atypical sam-ple of behaviors, and allows for observa-tion of behaviors that may not be observ-able in most primary care ofces, such asinterest in peers.

Well-developed parent checklistsare easy to administer, and in the shortterm may be the most cost effective op-tion available to primary pediatric pro-viders. Developmental surveillance,which includes clinical judgment, isaided by the systematic elicitation of pa-rental concerns across developmental do-mains, using validated parent question-

naires.[Glascoe et al., 1989; Glascoe,1994; Glascoe and Dworkin, 1995; Glas-coe and Sandler, 1995; Glascoe, 1997].Finally, parental concerns regarding theemotional, behavioral, cognitive, nemotor, and language development oftheir children are predictive of the exis-tence of a true problem [Glascoe andDworkin, 1995].

Efforts to increase public awarenessof developmental disorders such as autisminclude the Centers for Disease Controland Prevention (CDC)s Learn the Signscampaign (www.cdc.gov/actearly; 1800-CDC-INFO), as well as a comprehensiveparent-developed site (www.rstsigns.org).These campaigns are aimed at educatingparents about child development and thewarning signs of developmental disorderssuch as autism. As part of the campaign,both the CDC and First Signs have devel-oped kits with information about screeningand early development for pediatric pro-viders. It is always possible that parents oftypically developing children may unnec-essarily become concerned as a result ofpublic awareness campaigns such as Learnthe Signs. In such instances, reassurance andparent education may be all that is required.More research is needed to ascertainwhether the nature of the concerns raisedby parents of children found to have de-velopmental disorders differ from thoseraised by the parents of typically developingchildren.

The advantages of professional ob-servation include knowledge of the fullrange of what constitutes typical devel-opment. Clinicians are more objectivethan parents and may be less likely toover- or under-estimate a given childsskills or problem behaviors. The signi-cance of a given behavior may not be asclear to a parent as it might be to a trainedobserver. Finally, in many instances, thesetting used for undertaking the observa-tions can be standardized.

GENERAL DEVELOPMENTALSCREENING TOOLS

There are many broad develop-mental screening tools that may have arole in the early identication process.Discussion of these many tools is beyondthe scope of this article, and the reader isreferred to a recent review article on thesubject [Glascoe, 2000]. This review fo-cuses briey on two of the tools widelyused by pediatric providers and one thatfocuses on communication and symbolicbehavior. The three measures are theParents Evaluation of DevelopmentalStatus (PEDS), the Ages and StagesQuestionnaire (ASQ), and the Commu-nication and Symbolic Behavior Scales


Developmental Prole (CSBS DP). Asscreens of general development some ofthese tools may not target the core signsand symptoms of autism, for example,the PEDS does not include questionsabout play and imitation.

The PEDS is a ten item instrumentdesigned to be self-administered by par-ents of children from birth to 8 years ofage, and elicits parent concern in devel-opmental domains such as ne and grossmotor skills, receptive and expressive lan-guage, and self-help skills. The type ofparental concerns leads to categorizationof risk for developmental difculties:high, moderate, and low with corre-sponding responses such as refer, reassure,and offer developmental promotion andparent education. The PEDS has beenstandardized and validated and meets therecommended psychometric propertiesfor a screening instrument (sensitivity andspecicity greater than 0.70) [Glascoe,2001; Glascoe, 2003; www.pedstest.com].

The Ages and Stages Question-naire (ASQ) is developed for use withchildren 460 months old, parents areasked to respond to descriptions of a listof skills with the response options: notyet, sometimes, or yes [Bricker andSquires, 1999]. Each form is intended foruse with a certain age group (e.g., 4 yearsold) and consists of thirty developmentalitems divided into ve areas: communi-cation, gross motor, ne motor, problemsolving, and personal-social. There is alsoa section for general parental concerns.Many of the items are accompanied byillustrations, and the reading level is sixthgrade or below. The items were selectedbased on the likelihood of being ob-served or elicited by parents in the homesetting. The ASQ reportedly has a rangeof classication agreement from 76 to91% [Bricker and Squires, 1999], withdevelopmental assessment tools such asthe Bayley Scales of Infant Development[Bayley, 1993].

The Communication and Sym-bolic Behavior Scales DevelopmentalProle (CSBS DP, Wetherby and Pri-zant, 2002] includes three measures: theInfant-Toddler checklist, an expandedCaregiver Questionnaire, and a BehaviorSample. The CSBS DP is based on a toolpreviously designed by the same authorsthe CSBS [Wetherby and Prizant, 1993].The CSBS DP is a screening and evalu-ation instrument designed to measure thecommunicative and symbolic abilities ofchildren aged 1224 months. The mea-sured skills form three composites: social(emotion, eye gaze, and communica-tion), speech (sounds and words), and

symbolic (understanding and object use)[Wetherby et al., 2004]. The 24 itemInfant-Toddler checklist asks about de-velopmental milestones (e.g., when youlook at and point to a toy across theroom; does your child look at it?; doesyour child point to objects?). Nineteen ofthe twenty four items have the answerchoices: not yet, sometimes, and often,the remaining ve are how many ques-tions. At the bottom of the form parentsare also asked a yes/no question: do youhave any concerns about your childs de-velopment? They are asked to describeany existing concerns they may have. It isdesigned to be completed by the parentor caregiver. The Behavior Sample is avideotaped evaluation of the child duringan interaction with his/her parent andthe clinician. Wetherby et al., [2004]propose the use of the Infant-ToddlerChecklist as a rst level screen, with theBehavior Sample serving as a second levelevaluation tool. The tool was not de-signed to be an autism-specic screeningtool, but rather a tool to screen for delaysin communication, including prelinguis-tic skills; as such, it may be sensitive to arange of social communicative delays, in-cluding autistic spectrum disorder.

The PEDS, ASQ, and CSBS DPare not designed to selectively screen forautism, but they may be effective in de-tecting children whose developmentalproblems are consistent with autism.Screening a large sample of children froman unselected population will be neededto demonstrate their usefulness as pri-mary screens for children with autism.

SCREENING TOOLS FORAUTISTIC SPECTRUMDISORDERS (ASD)

There are several autism-specictools currently under development, orrevision, and as such it would be prema-ture to recommend one. Many of thesepromising tools still need to undertakecontinued follow up of the original sam-ples used in their development, as well ascross-validation with new samples.

In addition to the ModiedChecklist for Autism in Toddlers (M-CHAT), which will be discussed in moredetail later, four other autism-specicscreening tools are the Checklist for Au-tism in Toddlers (CHAT), Pervasive De-velopmental Disorders Screening Test-II(PDDST-II), Screening Tool for Autismin Two year olds (STAT), and Checklistfor Autism in Toddlers-23 (CHAT-23).

CHATThe CHAT is a screening tool in-

tended for use in the general population

at 18 months of age. It consists of nineparent report items (A19) and ve ob-servation items (Bi-v) [Baron-Cohen etal., 2000; 1996; 1992]. When looked attogether ve of the fourteen items havebeen found to be key indicators of pos-sible autism based on the Diagnostic andStatistical Manual (DSM) criteria for au-tistic disorder: gaze monitoring (Bii);proto-declarative pointing (pointing toindicate interest rather than to request)(A7 Biv), and pretend play (A5 Biii). Baron-Cohen et al., [1996] foundthat if at 18 months the items gaze mon-itoring, proto-declarative pointing, andpretend play were failed, there was an83.3% risk of being subsequently diag-nosed with autism via a diagnostic eval-uation. A follow up study [Baird et al.,2000] with 16,235 eighteen month oldchildren using the CHAT found, how-ever, that although the specicity of theCHAT was excellent, the sensitivity(2038% depending on criteria used) wasunacceptably low. This suggests that theCHAT may not be an appropriate tool touse as an exclusive screening tool foridentifying children who may have anASD, and is currently under revision bythe authors.

PDDST-IIPDD ST-II [Siegel, 2004] consists

of 3 stages designed to be used in threedifferent clinical settings. The PDDSTStage 1 consists of 22 items with theresponse choices yes, usually true orno, usually not true. Its intended use isin the primary care setting with 1248month olds. A positive screen signals theneed for further evaluation and is denedas 5 or more items being answered yes,usually true. The author reports a sensi-tivity of 0.92 and 0.91 based on a sampleof 681 children at risk for ASD and 256children with mild-to-moderate otherdevelopmental disorders.

The other two stages of thePDDST are intended to be used in set-tings other than the primary care provid-ers ofce. Stage 2 is a 14-item screeningtool for use in developmental clinics. Thecut-off is still 5 items with a reportedsensitivity and specicity of 0.73 and 0.49based on 490 children with conrmedASD (Autism, PDD-NOS, or Aspergerssyndrome) and 194 children who wereevaluated for an ASD, but who did notreceive a diagnosis on the autistic spec-trum [Siegel, 2004]. Stage 3 is a 12 itemscreening tool designed to be adminis-tered in autism clinics with a cut-offscore of 8. The reported sensitivity andspecicity is 0.58 and 0.60, respectively[Siegel, 2004]. The sensitivity and spec-


icity reported for stage 1 are good.Those for stage 2 and 3 are not as good,particularly, the specicity for stage 2.The major statistical shortcoming of thisinstrument is that no sensitivity/specic-ity data are reported for large-scalescreening of an unselected sample.

STATThe STAT [Stone et al., 2000;

Stone et al., 2004] is designed as a secondlevel screening tool for autism. It is in-tended to be used by professionals incommunities to identify children withpossible autism, as opposed to other de-velopmental disabilities. The items areadministered within a twenty-minutelong play-based interactive session. Theten scored items include play (2), imita-tion (4), directing attention (4), and donot require language comprehension.Two additional requesting items are in-cluded as a means of promoting interac-tion between the evaluator and child.Two samples were used to assess the va-lidity of the STAT as a second levelscreening instrument. The sample used inthe development phase consisted of 40children (3 with autism, 33 without), thesample used for validation had 33 chil-dren (12 with autism, 21 without au-tism). The children in the two samplesranged in age from 24 to 35 months old.The children with autism had lower de-velopmental ages than the children with-out autism. Data from a subsample ofchildren with (12) and without (12) au-tism matched for developmental age wasalso analyzed. The two examiners admin-istering the STAT were blind to the re-ferral questions and the results of the di-agnostic assessments. The parents of theparticipating children were blind to theresults of the diagnostic evaluation untilafter the STAT screening was completed.The diagnoses were made based onDSM-IV criteria (APA, 1994) and theChildhood Autism Rating Scale (CARS;Schopler et al., 1988]. The sensitivity andspecicity of the subsample matched forDA were each 0.83. The sensitivity andspecicity for the validation sample were0.83 and 0.86, respectively.

The authors of the STAT highlightthree areas needing additional research.Firstly, replication studies with largersamples from many different settings areneeded before the ndings can be gener-alized. Secondly, future studies need touse standardized measures such as theAutism Diagnostic Observation ScheduleGeneric (ADOS-G) [Lord et al., 1999],in addition to clinical judgment. Finally,follow up studies with larger sample sizesneed to match those children diagnosed

with autism and those without for mentalages and assess the sensitivity and speci-city of the STAT in that instance. Nev-ertheless, the STAT is a promising Level2 screening tool for autism in childrenbetween two and three years old.

The M-CHATThe M-CHAT [Robins et al.,

2001] is a 23-item (yes/no) parent reportchecklist developed to screen childrenaged 1630 months. As suggested by thename; the M-CHAT is a modication ofCHAT [Baron-Cohen et al., 1992], de-scribed above. The M-CHAT expandsupon the parent report portion of theCHAT, and seeks to identify childrenwith a possible autistic spectrum disorder,including PDD-NOS, and not juststrictly dened Autistic Disorder.

The initial population screened bythe M-CHAT consisted of 1,293 chil-dren aged 1630 months [Robins et al.,2001]. Of the 1293 children screened 58were evaluated and 39 were diagnosedwith an autistic spectrum disorder. Noneof the children evaluated at that timewere found to be typically developing.Although absolute sensitivity and speci-city are pending completion of followup of the initial sample, based on thediscriminant function analysis (DFA)classication at the time of rst screeningand evaluation, the M-CHAT has a sen-sitivity of 0.87, specicity of 0.99, posi-tive predictive power of 0.80, and nega-tive predictive power of 0.99 [Robins etal., 2001].

Excluded from the ongoing valida-tion study are children who already havean ASD diagnosis prior to completion ofthe M-CHAT, and children youngerthan 16 months or older than 30 monthsat the time of M-CHAT completion.The sample is drawn from primary carepractices, as well as early interventionsites. Children who fail the M-CHATreceive a conrmatory follow up tele-phone call at which time the completedM-CHAT is reviewed to clarify the se-lected answers and ascertain whether thechild has in fact screened positive on thechecklist. The completed forms arescored and the child is considered to havefailed the initial screening if he/she failsany three of the twenty three items ortwo of the six critical items. The criticalitems on the M-CHAT are as follows:item 2 (interest in other children), item 7(proto-declarative pointing), item 9(bringing objects to show the parent),item 13 (imitating), item 14 (respondingto name), and item 15 (following apoint).

Those who fail in both the initialscreening and the telephone follow upare evaluated by a team of evaluators,using the Autism Diagnostic Interview-Revised (ADI-R) [Lord et al., 1994];ADOS-G [Lord et al., 2000; 1999];CARS [Schopler et al., 1988]; MullenScales of Early Learning [Mullen, 1995];Vineland Adaptive Behavior Scales[Sparrow et al., 1984], and the DSM-IVcriteria [APA, 1994]. The ADOS is asemistructured instrument, which allowsassessment of a childs communicationand social interactions. The ADI is a par-ent report measure with questions on thechilds communication, social related-ness, including play and interests and be-haviors. The ADI and ADOS are cur-rently considered the gold standard fordiagnosing autism.

The number screened at Time 1(completed form between 16 and 30months old) has been increased to 4,200children. 236 children who failed the ini-tial screening have been evaluated at amean age of 27.63 months: 165 werefound to have an ASD, 67 have beendiagnosed with a non-ASD developmen-tal disorder, and 4 have been assessed astypically developing. To date 1,937 chil-dren evaluated at Time 1 have reachedeligibility for rescreening at age 4. Datahas been collected from 940 (mean age55.35 months). At rescreening, parentsare asked to complete the M-CHAT, andare also asked whether in the interim thechild has been referred for or diagnosedwith an ASD. Six children have beenidentied as possible misses (failed therescreening or referred for possible ASD):2 of the 6 have been conrmed as misses,two are not misses, and two are classiedas possible misses because they have notbeen able to be evaluated. Sixty threechildren evaluated at Time 1 have beenreevaluated at Time 2. At Time 2, 38 ofthe 63 children have been diagnosedwith an ASD, 17 have been diagnosedwith a non-ASD developmental disor-der, and 8 have been assessed as typicallydeveloping. Based on the M-CHATscore at Time 1 and the evaluation out-come at Time 2, the sensitivity and spec-icity are 0.85 and 0.93, respectively.(The sensitivity is as high as 0.95 if thepossible misses are not included asmisses). It is possible that the psychomet-ric properties of the M-CHAT will bedifferent when larger numbers from anunselected population are obtained.

Although this is still a work inprogress, based on the data collected todate, certain items appear to be helpful indistinguishing between children withASD as compared to those with a non-


ASD developmental disorder [Dixon etal., in press]. In particular, item 7 pointfor interest seem to be the most potentdiscriminator between the two groups ofchildren.

The diagnoses made at 2 years oldalso appear to be relatively stable basedon the available follow up data. Althoughin a few cases, children moved from anon-ASD diagnosis to the typical devel-opment category, most of the changes indiagnosis were from an ASD to a non-ASD developmental disorder. Whetherthis change in diagnosis is due to mis-diagnosis at Time 1 or is the result ofearly intervention and/or increasing ma-turity remains unclear at this time.

Part of the Early Detection study isthe screening of the siblings of childrenwith ASDs. Recent analyses of data col-lected from younger siblings of childrendiagnosed with an ASD found an initialscreen positive rate (before phone inter-view) of 43%. As the M-CHAT is aparent-report measure, it is possible thatthese gures are partially inated becauseof increased parental concern as a resultof the previous ASD diagnosis of theolder sibling. After the phone interview,30% remained a positive screen. Twentythree percent of the younger siblingswere evaluated and diagnosed with anASD. This is a higher rate than what hasbeen reported in the literature for sib-lings, which may be due to an ascertain-ment bias, the parents of the siblings inthis study may have sought participationin the study as a result of concerning signsthat they observed.

The study also involves screeningof children within low risk settings (pri-mary care provider sites, Peds) and highrisk settings (early intervention sites, EI).Based on 20 children screened throughpediatric ofces, the items that discrimi-nate the children with Peds-referredASD from the EI-referred ASD appear tobe highly similar.

CHAT-23Wong et al. [2004] used the M-

CHAT and CHAT to develop a screen-ing tool for use with a sample of Chinesechildren, the CHAT-23. The changesmade to the original M-CHAT were thetranslation of the questions into tradi-tional chinese and the addition of agraded scoring system [never (0%), sel-dom (25%), usually (2550%), or often(50%)] for 22 of the 23 M-CHATquestions instead of the yes/no responsesof the original M-CHAT. The gradedsystem was subsequently collapsed intoyes (usually/often) and no (never/sel-dom) grouped to dene fail/pass. In ad-

dition to the 23 item-questionnaire, theCHAT-23 has an observational sectionconsisting of 5 direct observational itemsfrom the CHAT. When used with 212children (87 with ASD and 125 non-ASD) with the mental age range of1824 months, Wong et al. found thatfailing any 2 of 7 key questions (imitate,pretend, point for interest, check reac-tion of parent, bring objects to show,follow a point, and take interest in otherchildren) or any 6 of the overall 23 ques-tions in the parent administered ques-tionnaire yielded a sensitivity of 0.931and 0.839, respectively. The specicitywas of 0.768 or 0.848, respectively. Theauthors suggest a two-level screening sys-tem beginning with the M-CHAT andproceeding to the use of the observa-tional portion of the CHAT for thosewho failed the M-CHAT. If the obser-vation is failed, the child would then bereferred for an autism evaluation. Thefact that the psychometric properties be-ing found when the CHAT-23 is usedwith a sample of Chinese children soclosely resemble what is being foundwith the M-CHAT supports the validityof these instruments. In Wong et al.sstudy, the chronologic ages of the en-rolled children whose mental ages were1824 months old were 1686 monthsold. It is important to note that whenused with children whose mental ages arenot known to be 1824 months, theefcacy of the tool may be different. Inaddition, this range of chronologic ages isquite wide, future research should exam-ine whether the screener is equally validacross this age range.

CONCLUSIONAt this time, early intervention is

the best response to Autistic SpectrumDisorders. Given this fact, the develop-ment of reliable tools for screening chil-dren under three years old is imperative.Early screening will lead to early diagno-sis, early intervention, and the ultimategoalimproved outcomes. There remainsthe need for further research to elucidatemany issues. When is the best age toscreen? Are the diagnoses made in thetoddler years stable? If the children sub-sequently improve, is it due to the effectsof early intervention or was the earlydiagnosis incorrect? What is the best wayto screen for autism? Is preliminary gen-eral developmental surveillance/screen-ing best or is autism-specic screeningneeded for all children? Are there popu-lations, such as the siblings of childrenwith ASDs, for whom autism-specictools are best?

Since many of the autism-specictools for screening in young children arestill under development, close attentionneeds to be paid to their psychometricproperties in different SES and ethno-cultural groups. The M-CHAT is one ofa few tools, which show promise as ascreening tool in different populations ofunselected children. It has been translatedinto Turkish, Chinese, Japanese, andSpanish. It will be important to evaluatethe ndings of studies using theM-CHAT in these different languagesand cultures for similarities and differ-ences in their ndings. Beyond the lan-guages being used, different cultural andsocio-economic groups may interpret thequestions differently, yielding differentresults. Not only may the specic signs/symptoms of autism vary across culturalgroups, the meaning attributed to themby parents may as well. For example, thefact that a child does not point or makeeye contact may be considered appropri-ate and acceptable if to point and makeeye contact with adults is considered asign of disrespect. Therefore, more workfocused on understanding what concernscompatible with ASDs may be more uni-versal than others is needed.

Despite these unresolved issues, afew facts remain clear. First, primary careproviders are the appropriate people toscreen children between birth and theage of 5 years old, since they are thequalied professionals with whom par-ents have ongoing contact. Second, thewell-child care visit provides only a lim-ited amount of time for the screening ofchildren, with each of the many disorderspotentially requiring that their ownscreening tool and process be used. Thetools that will be used by providers arelikely to be those that accurately ag chil-dren for many developmental disorders.Third, in light of the data supporting theusefulness of eliciting parental concerns,any tool used will probably need to in-corporate the parents perspectives.Fourth, data suggests that children of par-ents from minority racial/ethnic back-grounds regardless of SES and those withlower socio-economic status may not bereceiving quality health care. This dispar-ate care includes a lack of screening, eval-uation/diagnosis, and service provision.This must be addressed promptly, sincethe literature supports the notion that theearlier the identication of and interven-tion for developmental disorders, thebetter the outcomes.

ACKNOWLEDGMENTSWe thank the Early Detection Team

at the University of Connecticut (Mari-


anne Barton, Sarah Hodgson, James Green,Gail Marshia, Pam Dixon, Jamie Klein-man, Juhi Pandey, Hilary Boorstein,Emma Esser, Leandra Wilson, MichaelRosenthal, Saasha Sutera, and Alyssa Ver-balis); our collaborators at the Yale Univer-sity Child Study Center (Fred Volkmar,Ami Klin, Katarzyna Chawarska, TammyBabbitz), the University of Washington(Geraldine Dawson, Karen Toth), and Di-ana Robins at Georgia State University.Most importantly, we thank the familiesand children who are participating in theongoing Early Detection Study. Specialthanks to the library staff at St. FrancisHospital in Hartford, CT for their ongoingassistance.

REFERENCESAmerican Academy of Pediatrics. 2001a. Technical

report: the pediatricians role in the diagnosisand management of autistic spectrum disorderin children. Pediatrics 107(5).

American Academy of Pediatrics, Committee onChildren with Disabilities. 2001b. Develop-mental surveillance and screening of infantsand young children. Pediatrics 108:192196.

American Academy of Pediatrics, Committee onChildren with Disabilities. 2001c. The pedi-atricians role in the diagnosis and manage-ment of autistic spectrum disorder in children.Pediatrics 107:12211226.

American Academy of Pediatrics, Medical HomeInitiatives for Children with Special NeedsProject Advisory Committee. 2002. TheMedical home. Pediatrics 110:184186.

American Psychiatric Association (APA). 1994. Di-agnostic and Statistical Manual of Mental dis-orders, 4th edition (DSM-IV). Washington,DC: American Psychiatric Association.

Baird G, Charman T, Baron-Cohen S, et al. 2000.A screening instrument for autism at 18months of age: a six-year follow-up study.J Am Acad Child Adolesc Psychiatry 39:694702.

Ballaban-Gil K, Tuchman R. 2000. Epilepsy andepileptiform EEG: association with autismand language disorders. Ment Retard DevDisabil Res Rev 6:300308.

Baranek GT. 1999. Autism during infancy: a ret-rospective video analysis of sensory-motorand social behaviors. J Autism Dev Disord29:213224.

Barbaresi WJ, Katusic SK, Colligan RC, et al.2005. The incidence of autism in Olmstedcounty, Minnesota, 19761997: results froma population-based study. Arch Pediatr Ado-lesc Med 159:3744.

Baron-Cohen S, Allen J, Gillberg C. 1992. Canautism be detected at 18 months? The needle,the haystack, and the CHAT. Br J Psychiatry161:839843.

Baron-Cohen S, Cox A, Baird G, et al. 1996.Psychological markers in the detection of au-tism in infancy in a large population. Br JPsychiatry 168:158163.

Baron-Cohen S, Wheelwright S, Cox A, et al.2000. Early identication of autism by theChecklist for Autism in Toddlers (CHAT). JR Soc Med 93:521525.

Bayley N. 1993. The Bayley Scales of Infant De-velopment, 2nd edition. San Antonio, TX:Harcourt Brace.

Bertrand J, Mars A, Boyle C, et al. 2001. Preva-lence of autism in a United States population:

the Brick Township, New Jersey, investiga-tion. Pediatrics 108:11551161.

Bethell C, Reuland CH, Halfon N, et al. 2004.Measuring the quality of preventive and de-velopmental services for young children: na-tional estimates and patterns of clinicians per-formance. Pediatrics 113:19731983.

Birnbrauer JS, Leach DJ. 1993. The Murdoch earlyintervention program after 2 years. BehavChange 10: 6374.

Bricker D, Squires J. 1999. Ages and Stages Ques-tionnaires: A parent-completed childmoni-toring system. 2nd ed. Baltimore: Paul H.Brookes Publishing Co.

Campbell M, Locascio JJ, Choroco MC, et al.1990. Stereotypies and tardive dyskinesia: ab-normal movements in autistic children. Psy-chopharmacol Bull 26:260266.

CDCs Learn the Signs. Available at www.cdc.gov/actearly; 1800-CDC-INFO

Chakrabarti S, Fombonne E. 2001. Pervasive de-velopmental disorders in preschool children.JAMA 285:30933099.

Chakrabarti S, Fombonne E. 2005. Pervasive de-velopmental disorders in preschool children:conrmation of high prevalence. Am J Psy-chiatry 162:11331141.

Charman T, Baird G. 2002. Practitioner review:diagnosis of autism spectrum disorder in 2-and 3-year-old children. J Child Psychol Psy-chiatry 43:289305.

Charman T, Swettenham J. 2001. Repetitive be-haviors and social-communicative impair-ments in autism: implications for develop-mental theory and diagnosis. In: Burack JA,Charman T, Yirmiya N, Zelazo PR, editors.The Development of Autism: Perspectivesfrom Theory and Research. Mahwah, NJ:Erlbaum. p 325345.

Charman T, Swettenham J, Baron-Cohen S, et al.1998. An experimental investigation of social-cognitive abilities in infants with autism: clin-ical implications. Infant Ment Health J 192:260275.

Courchesne E, Redcay E, Kennedy D. 2004. Theautistic brain: birth through adulthood. CurrOpin Neurol 17:489496.

Cox A, Klein K, Charman T, et al. 1999. Autismspectrum disorders at 20 and 42 months ofage: stability of clinical and ADI-R diagnosis.J Child Psychol Psychiatry 40:719732.

Dahlgren SO, Gillberg C. 1989. Symptoms in therst two years of life: a preliminary populationstudy of infantile autism. Eur Arch PsychiatryNeurol Sci 238:169174.

Dawson G, Osterling J. 1997. Early intervention inautism. In: Guralnick MJ, editor. The Effec-tiveness of Early Intervention. Baltimore,MD: Paul H. Brookes. p 307326.

Dawson G, Meltzoff AN, Osterling J, et al. 1998.Children with autism fail to orient to natu-rally occurring social stimuli. J Autism DevDisord 28:479485.

Dawson G, Munson J, Estes A, et al. 2002. Neu-rocognitive function and joint attention abil-ity in young children with autism spectrumdisorder versus developmental delay. ChildDev 73:345358.

Dawson G, Toth K, Abbott R, et al. 2004. Earlysocial attention impairments in autism: socialorienting, joint attention, and attention todistress. Dev Psychol 40:271283.

DeGiacomo A, Fombonne E. 1998. Parental rec-ognition of developmental abnormalities inautism. Eur J Child Adolesc Psychiatr 7:131136.

Dixon P, Kleinman J, Pandey J, et al. (in press).Differentiating between autism spectrum dis-orders and other developmental disabilities in

children who failed a screening instrument forASD. J Autism Dev Disord.

Dworkin PH. 1993. Detection of behavioral, de-velopmental and psychosocial problems in pe-diatric primary care practice. Curr Opin Pe-diatr 5:531536.

Dyches TT, Wilder LK, Sudweeks RR, et al. 2004.Multicultural issues in autism. J Autism DevDisord 34:211222.

Filipek P, Accardo P, Ashwal S, et al. 2000. Practiceparameter: screening and diagnosis of autism:a report of the quality standards committee ofthe American Academy of Neurology and theChild Neurology Society. Neurology55:468479.

Filipek P, Accardo P, Baranek G, et al. 1999. Thescreening and diagnosis of autistic spectrumdisorders. J Autism Dev Disord 29:439484.

First Signs. Available at www.rstsigns.orgFlanagan O, Nuallain S. 2001. A study looking at

the effectiveness of developmental screeningin identifying learning disabilities in earlychildhood. Ir Med J 94:148150.

Fombonne E. 2003. Epidemiological surveys ofautism and other pervasive developmentaldisorders: an update. J Autism Dev Disord33:365382.

Glascoe F. 1994. Its not what it seems. The rela-tionship between parents concerns and chil-dren with global delays. Clin Pediatr 33:292296.

Glascoe F. 2001. Parents Evaluation of Develop-mental Status. Nashville, TN: Ellsworth andVandermeer Press, LLC. www.pedstest.com

Glascoe F. 2003. Parents evaluation of develop-mental status: how well do parents concernsidentify children with behavioral and emo-tional problems? Clin Pediatr 42:133138.

Glascoe FP. 1997. Parents concerns about chil-drens development: prescreening techniqueor screening test? Pediatrics 99:522528.

Glascoe FP. 2000. Early detection of developmentaland behavioral problems. Pediatr Rev21:272280.

Glascoe FP, Dworkin PH. 1995. The role of par-ents in the detection of developmental andbehavioral problems. Pediatrics 95:829836.

Glascoe FP, Sandler H. 1995. Value of parentsestimates of childrens developmental ages.Disabil Rehabil 127:831835.

Glascoe FP, Altemeier WA, MacLean WE. 1989.The importance of parents concerns abouttheir childs development. Am J Dis Child143:955958.

Glaze D. 2004. Rett syndrome: of girls and micelessons for regression in autism. Ment RetardDev Disabil Res Rev 10:154158.

Goldberg WA, Osann K, Filipek PA, et al. 2003.Language and other regression: assessmentand timing. J Autism Dev Disord 33:607616.

Harris S, Handleman J. 2000. Age and IQ at intakeas predictors of placement for young childrenwith autism: a four to six-year follow-up. JAutism Dev Disord 30:137142.

Harris SL, Handleman JS, Gordon R, et al. 1991.Changes in cognitive and language function-ing of preschool children with autism. J Au-tism Dev Disord 21:281290.

Hoyson M, Jamisson B, Strain P. 1984. Individu-alized group instruction of normally develop-ing and autistic-like children: the LEAP cur-riculum model. J Div Early Child 8:157172.

Hrdlicka M, Komarek V, Propper L, et al. 2004.Not EEG abnormalities but epilepsy is asso-ciated with autistic regression and mental


functioning in childhood autism. Eur ChildAdolesc Psychiatry 13:209213.

Jocelyn LJ, Casiro OG, Beattie D, et al. 1998.Treatment of children with autism: a ran-domized control trial to evaluate a caregiver-based intervention program in communityday-care centers. J Dev Behav Pediatr 19:326334.

King T, Glascoe F. 2003. Developmental surveil-lance of infants and young children in pedi-atric primary care. Curr Opin Pediatr 15:624629.

Klin A, Jones W, Schultz R, et al. 2002. Deningand quantifying the social phenotype in au-tism. Am J Psychiatry 1596:895908.

Klinger, Dawson & Renner. 2003. Autistic disor-der. In: Mash EJ, Barkley RA, editors. ChildPsychopathology. 2nd ed. New York: TheGuilford Press. p 409454.

Lainhart JE, Ozonoff S, Coon H, et al. 2002. Au-tism, regression, and the broader autism phe-notype. Am J Med Genet 113:231237.

Lord C. 1995. Follow-up of two-year-olds referredfor possible autism. J Child Adolesc Psychiatr36:13651382.

Lord C, Risi S, Lambrecht L, et al. 2000. Theautism diagnostic observation schedulege-neric: a standard measure of social and com-munication decits associated with the spec-trum of autism. J Autism Dev Disord 30:205223.

Lord C, Rutter M, DiLavore P, et al. 1999. AutismDiagnostic Observation Schedule-WPS edi-tion. Los Angeles, CA: Western Psychologi-cal Services.

Lord C, Rutter M, Le Couteur A. 1994. Autismdiagnostic interview-revised: a revised versionof a diagnostic interview for caregivers ofindividuals with possible pervasive develop-mental disorders. J Autism Dev Disord 24:659685.

Lord C, Shulman C, DiLavore P. 2004. Regressionand word loss in autistic spectrum disorders.J Child Psychol Psychiatr 45:936955.

Lovaas O. 1987. Behavioral treatment and normaleducational and intellectual functioning inyoung autistic children. J Consult Clin Psy-chol 55:39.

Malhotra S, Gupta N. 1999. Childhood disintegra-tive disorder. J Autism Dev Disord 29:491498.

Mandell DS, Listerud J, Levy SE, et al. 2002. Racedifferences in the age at diagnosis amongMedicaid-eligible children with autism. J AmAcad Child Adolesc Psychiatr 41:14471453.

Manning-Courtney P, Brown J, Molloy CA, et al.2003. Diagnosis and treatment of autism spec-trum disorders. Curr Probl Pediatr AdolescHealth Care 33:283304.

McEachin J, Smith T, Lovaas O. 1993. Long-termoutcome for children with autism who re-ceived early intensive behavioral treatment.Am J Ment Retard 97:359372.

McEvoy R, Rogers S, Pennington B. 1993. Exec-utive function and social communication def-icits in young autistic children. J Child Psy-chol Psychiatr 34:563578.

Mullen E. 1995. Mullen Scales of Early Learning.Circle Pines, MN: American Guidance.

Noterdaeme M, Sitter S, Mildenberger K, et al.2000. Diagnostic assessment of communica-tive and interactive behaviours in childrenwith autism and receptive language disorder.Eur Child Adolesc Psychiatr 9:295300.

Osterling J, Dawson G. 1994. Early recognition ofchildren with autism: a study of rst birthdayhome videotapes. J Autism Dev Disord 24:247257.

Osterling J, Dawson G, Munson J. 2002. Earlyrecognition of 1-year-old infants with autismspectrum disorder versus mental retardation.Dev Psychopathol 14:239251.

Pinto-Martin J, Levy S. 2004. Early diagnosis ofautism spectrum disorders. Curr Treat Op-tions Neurol 391400.

Robins DL, Fein D, Barton ML, et al. 2001. Themodied checklist for autism in toddlers: aninitial study investigating the early detectionof autism and pervasive developmental disor-ders. J Autism Dev Disord 31:131144.

Robinson R, Baird G, Robinson G, et al. 2001.Landau-Kleffner syndrome: course and corre-lates with outcome. Dev Med Child Neurol43:243247.

Rogers S. 2004. Developmental regression in au-tism spectrum disorders. Ment Retard DevDisabil Res Rev 10:139143.

Rogers S, Lewis H. 1989. An effective day treat-ment model for young children with perva-sive developmental disorders. J Am AcadChild Adolesc Psychiatr 28:207214.

Schopler E, Reichler RJ, Renner BR. 1988. TheChildhood Autism Rating Scale. Los Angeles:Western Psychological Services.

Sheinkopf SH, Siegel B. 1998. Home-based behav-ioral treatment of young children with au-tism. J Dev Disord 28:1523.

Sia C, Tonniges TF, Osterhus E, et al. 2004. His-tory of the medical home concept. Pediatrics113:14731478.

Sices L, Feudtner C, McLaughlin J, et al. 2003.How do primary care physicians identifyyoung children with developmental delays? Anational survey. Dev Behav Pediatr 24:409417.

Sices L, Feudtner C, McLaughlin J, et al. 2004.How do primary care physicians manage chil-dren with possible developmental delays? Anational survey with an experimental design.Pediatrics 113:274281.

Siegel B. 2004. Pervasive Developmental DisordersScreening Test II (PDDST-II): Early Child-hood Screeners for Autistic Spectrum Disor-ders. Harcourt Assessment, Inc.

Siegel B, Pliner C, Eschler J, et al. 1988. Howchildren with autism are diagnosed: difcul-ties in identication of children with multipledevelopmental delays. J Dev Behav Pediatr9:199204.

Siperstein R, Volkmar F. 2004. Brief report: pa-rental reporting of regression in children withpervasive developmental disorders. J AutismDev Disord 34:731734.

Smedley BD, Stith AY, Nelson AR. 2003. UnequalTreatment: Confronting Racial and EthnicDisparities in Health Care. National Acade-mies Press.

Smith T, Eikeseth S, Klevstrand M, et al. 1997.Intensive behavioral treatment for preschool-ers with severe mental retardation and mentalretardation and pervasive developmental dis-order. Am J Ment Retard 102:238249.

Smith T, Lovaas O. 1998. Intensive and early be-havioral intervention with autism: the UCLAyoung autism project. Infants Young Child10:6778.

Sonnander K. 2000. Early identication of childrenwith developmental disabilities. Acta PaediatrSuppl 89:1723.

Sparrow S, Balla D, Cichetti D. 1984. The Vine-land Adaptive Behavior Scales. Circles Pines,MN: American Guidance Services.

Stefanatos G, Kinsbourne M, Wasserstein J. 2002.Acquired epileptiform aphasia: a dimensionalview of Lansu-Kleffner syndrome and the re-lation to regressive autistic spectrum disoders.

Neuropsychol Dev Cogn C Child Neuropsy-chol 8:195228.

Stone W, Coonrod E, Ousley O. 2000. Brief re-port: screening tool for autism in two-year-olds (STAT): development and preliminarydata. J Autism Dev Disord 30:607612.

Stone WL, Coonrod EE, Turner LM, et al. 2004.Psychometric properties of the STAT forearly autism screening. J Autism Dev Disord34:691701.

Stone W, Hoffman E, Lewis S, et al. 1994. Earlyrecognition of autism: parental reports vs.clinical observation. Arch Pediatr AdolescMed 148:74179.

Stone W, Lee E, Ashford L, et al. 1999. Can autismbe diagnosed accurately in children under 3years? J Child Psychol Psychiatry 40:219226.

Stone W, Lemanek K. 1990. Parental report ofsocial behaviors in autistic preschoolers. J Au-tism Dev Disord 20:513522.

Strickland B, McPherson M, Weissman G, et al.2004. Access to the medical home: results ofthe National Survey of Children with specialhealth care needs. Pediatrics 113:14851492.

Szatmari P, Bartolucci G, Bremmer R. 1989. As-pergers syndrome and autism: comparison ofearly history and outcome. Dev Med ChildNeurol 31:709720.

Travis L, Sigman M, Ruskin E. 2001. Links be-tween social understanding and social behav-ior in verbally able children with autism. JAutism Dev Disord 31:119130.

Trevathan E. 2004. Seizures and epilepsy amongchildren with language regression and autisticspectrum disorders. J Child Neurol 19:S49S57.

Tuchman R, Rapin I. 1997. Regression in perva-sive developmental disorders: seizures and ep-ileptiform electroencephalogram correlates.Pediatrics 99:560566.

Turner M. 1999. Annotation: repetitive behaviorin autism: a review of psychological research.J Child Psychol Psychiatr 40:839849.

Volkmar F, Cohen D, Paul R. 1986. An evaluationof DSM-III criteria for infantile autism. J AmAcad Child Psychiatr 25:190197.

Volkmar F, Rutter M. 1995. Childhood disintegra-tive disorder: results of the DSM-IV autismFIELD trial. J Am Acad Child Adolesc Psy-chiatr 34:10921095.

Volkmar FR, Lord C, Bailey A, et.al. 2004. Autismand pervasive developmental disorders.J Child Psychol Psychiatr 45(1): 135170.

Werner E, Dawson G, Osterling J, et al. 2000. Briefreport: recognition of autism spectrum disor-der before one year of age: a retrospectivestudy based on home videos. J Autism DevDisord 30:157162.

Wetherby A, Prizant B. 1993. The Communica-tion and Symbolic Behavior Scale. NewYork, NY: Riverside.

Wetherby A, Prizant B. 2002. Communication andSymbolic Behavior Scales DevelopmentalProleFirst Normed Edition. Baltimore,MD: Paul H. Brookes.

Wetherby AM, Woods J, Allen L, et al. 2004. Earlyindicators of autism spectrum disorders in thesecond year of life. J Autism Dev Disord34:473493.

Wong V, Hui L, Lee W, et al. 2004. A ModiedScreening Tool for Autism (Checklist for Au-tism in Toddlers [CHAT-23] for Chinesechildren. Pediatrics 114(2):e166e176.

Woods J, Wetherby A. 2003. Early identicationand intervention for infants and toddlers whoare at risk for autism spectrum disorder. Lan-guage Speech Hearing Services Schools 34:180193.


Yeargin-Allsopp M, Rice C, Karapurkar T, etal. 2003. Prevalence of autism in a USmetropolitan area. JAMA. 289(1):4955.

Young KT, Davis K, Schoen C, et al. 1998. Lis-tening to parents. A national survey of parents

with young children. Arch Pediatr AdolescMed 152:255262.

Young R, Brewer N, Pattison C. 2003. Parentalidentication of early behavioural abnormal-ities in children with autistic disorder. Autism2003 7:125143.

Zwaigenbaum L. 2001. Autistic spectrum disordersin preschool children. Can Fam Physician 47:20372042.

Zwaigenbaum L, Bryson S, Rogers T, et al. 2005.Behavioral manifestations of autism in the rstyear of life. Int J Dev Neurosci 23:143152.

APPENDIX: M-CHAT*Please ll out the following about how your child usually is. Please try to answer every question. If the behavior is rare (e.g., youhave seen it once or twice), please answer as if the child does not do it.

1. Does your child enjoy being swung, bounced on your knee, etc.? Yes No2. Does your child take an interest in other children? Yes No3. Does your child like climbing on things, such as up stairs? Yes No4. Does your child enjoy playing peek-a-boo/hide-and-seek? Yes No5. Does your child ever pretend, for example, to talk on the phone or take care of dolls, or pretend other things? Yes No6. Does your child ever use his/her index nger to point, to ask for something? Yes No7. Does your child ever use his/her index nger to point, to indicate interest in something? Yes No8. Can your child play properly with small toys (e.g., cars or bricks) without just mouthing, ddling, or dropping them? Yes No9. Does your child ever bring objects over to you (parent) to show you something? Yes No

10. Does your child look you in the eye for more than a second or two? Yes No11. Does your child ever seem oversensitive to noise? (e.g., plugging ears) Yes No12. Does your child smile in response to your face or your smile? Yes No13. Does your child imitate you? (e.g., you make a facewill your child imitate it?) Yes No14. Does your child respond to his/her name when you call? Yes No15. If you point at a toy across the room, does your child look at it? Yes No16. Does your child walk? Yes No17. Does your child look at things you are looking at? Yes No18. Does your child make unusual nger movements near his/her face? Yes No19. Does your child try to attract your attention to his/her own activity? Yes No20. Have you ever wondered if your child is deaf? Yes No21. Does your child understand what people say? Yes No22. Does your child sometimes stare at nothing or wander with no purpose? Yes No23. Does your child look at your face to check your reaction when faced with something unfamiliar? Yes No

*1999 Diana Robins, Deborah Fein, and Marianne Barton


Documents

Autism in Young Children