MENTAL RETARDATION AND DEVELOPMENTAL DISABILITIES

RESEARCH REVIEWS 11: 253–262 (2005)

SCREENING FOR AUTISM IN YOUNG CHILDREN:

THE MODIFIED CHECKLIST FOR AUTISM IN
TODDLERS (M-CHAT) AND OTHER MEASURES
Thyde Dumont-Mathieu1,2* and Deborah Fein2
1
Department of Pediatrics, University of Connecticut School of Medicine, Farmington, Connecticut
2
Department of Psychology, University of Connecticut, Storrs, Connecticut

The literature on the importance of early identification and early 1998; Klin et al., 2002; Klinger et al., 2003]; (2) communicative
intervention for children with developmental disabilities such as autism abilities (use of language, both verbal and non-verbal, quality of
continues to grow. The increased prevalence of autistic spectrum disorders
has fostered research efforts on the development and validation of autism- communication and play) [Travis et al., 2001; Klinger et al.,
specific screening instruments for use with young children. There are cur- 2003; Dawson et al., 2004] and (3) restricted or stereotyped
rently several such autism-specific screening tools meant to be used with behaviors; interests, and activities (non-functional routines or
young children in various stages of development. Data from a few of these
screening instruments have been published, and they include the Checklist
ritualistic behaviors, resistance to changes in the environment,
for Autism in Toddlers (CHAT), Pervasive Developmental Disorders Screen- repetitive motor mannerisms, unusual interests and preoccupa-
ing Test (PDDST), Screening Tool for Autism in Two year olds (STAT), Check- tions; and visual fascinations) [Volkmar et al., 1986; Szatmari et
list for Autism in Toddlers-23 (CHAT-23), and the Modified Checklist for al., 1989; Campbell et al., 1990; Turner, 1999; Charman and
Autism in Toddlers (M-CHAT). In this review, these five tools designed for
use with children under three years old will be highlighted. In particular, the Swettenham, 2001; Klinger et al., 2003]. There are additional
Modified Checklist for Autism in Toddlers (M-CHAT) will be discussed. symptoms that are sometimes present with ASDs, these include
© 2005 Wiley-Liss, Inc. self injurious behaviors, functional disturbances such as difficul-
MRDD Research Reviews 2005;11:253–262.
ties with sleeping and eating, abnormal fears, and abnormal
responses to sensory stimuli [Klinger et al., 2003]. Cognitive
Key Words: autism; pediatric screening; early identification; early inter- functioning can range from severe mental retardation to above
vention average intelligence quotient (IQ), with many children who
have an ASD showing some degree of cognitive impairment. A
recent study by Chakrabarti and Fombonne [2005] found that in
BACKGROUND a sample of four to six year olds diagnosed with an ASD, nearly

A
utism is a neuro-developmental disorder first described a third had significant cognitive impairment.
by Leo Kanner in 1943. It is one of a group of disorders Previous studies have found that regression in ASD occurs
known as Pervasive Developmental Disorders (PDDs), most often between 18 and 24 months of age, with the fre-
now more commonly referred to as Autistic Spectrum Disorders quency of the regressive type of autism being somewhere be-
(ASDs). This group of disorders includes Autistic Disorder, tween 10 and 50% [Tuchman and Rapin, 1997; Goldberg et al.,
Pervasive Developmental Disorder-Not Otherwise Specified 2003; Lord et al., 2004; Rogers, 2004]. Clinical experience of
(PDD-NOS), Asperger’s syndrome, Rett syndrome, and Child- the present authors suggests that regression often occurs sooner,
hood Disintegrative Disorder (CDD) (American Psychiatric As- sometimes as early as 12–18 months. It is possible that parents
sociation (APA), 1994). Autistic Spectrum Disorders are cur- may recall 18 –24 months as the age of regression for children
rently estimated to affect somewhere between 1 in 166 and 1 in
who were diagnosed later, or that parents are experiencing recall
1000 children [Baird et al., 2000; Bertrand et al., 2001;
bias [Siperstein and Volkmar, 2004]. It is possible that as the age
Chakrabarti and Fombonne, 2001; Yeargin-Alsopp et al., 2003
Volkmar et al., 2004; Barbaresi et al., 2005; Chakrabarti and at diagnosis drops with earlier identification, so will the reported
Fombonne, 2005]. age at regression.
The diagnostic criteria for autistic disorder require impair-
ment in three areas: reciprocal social interaction, communica- Contract grant sponsor: National Institute of Child Health and Development
tion, and specific patterns of behavior, interests, and activities (NICHD); Contract grant number: 5-R01 HD39961.
*Correspondence to: Thyde Dumont-Mathieu, M.D., M.P.H., Department of Psy-
[APA, 1994]. Research has documented areas of dysfunction chology, University of Connecticut, 406 Babbidge Rd., Storrs, CT 06269-1020.
related to these core categories of symptoms in autism: (1) social E-mail: tdumont@stfranciscare.org
abilities (attachment, joint attention, social imitation, orienting Received 11 July 2005; Accepted 12 July 2005
Published online in Wiley InterScience (www.interscience.wiley.com).
to social stimuli, face and affect processing, expression of emo- DOI: 10.1002/mrdd.20072
tion, and symbolic play) [McEvoy et al., 1993; Dawson et al.,
© 2005 Wiley-Liss, Inc.
 There are many important theoret- more consistently around 12 months. 2001]. Simultaneously, the evidence sug-
ical and clinical issues related to the sub- The identified symptoms are fairly con- gesting that young children as young as
ject of regression in autistic spectrum dis- sistent and include delay or absence in 24 months old can be diagnosed with
orders. Do parents and providers mean pointing, showing objects, looking at autism continues to grow [Dahlgren and
the same thing when they discuss regres- others, and orienting to their name [Os- Gillberg, 1989; Lord, 1995; Stone et al.,
sion? What constitutes regression as op- terling and Dawson, 1994]. In addition 1999; Chakrabati and Fombonne, 2001;
posed to developmental stagnation? Is the to the difficulties with aspects of social Zwaigenbaum, 2001; Charman and
regression seen with autism similar responsiveness (poor visual orientation/ Baird, 2002; Dawson et al., 2002]. Fur-
mechanistically to that seen in other re- attention, response to name) these chil- thermore; the neurobiological abnormal-
gressive disorders such as Landau-Klef- dren also seem to exhibit sensory-motor ities noted in the child with ASD in the
fner syndrome [Ballaban-Gil and Tuch- difficulties (mouthing objects excessively, first two years of life [Courchesne et al.,
man, 2000; Robinson et al., 2001; aversions to social touch) [Baranek, 2004] support the idea that as early as 1
Stefanatos et al., 2002 Trevathan, 2004], 1999]. Other symptoms that seem to in- year of age, children with autism may
or to disorders within the autistic spec- dicate possible ASD include stereotyped show signs of developing quite differ-
trum, Childhood Disintegrative Disorder play with objects, unusual posturing of ently from their typically developing
[Volkmar and Rutter, 1995; Malhotra body parts, looking at a camera less fre- peers.
and Gupta, 1999 Goldberg et al., 2003; quently, staring/fixating on objects, and Although discussions continue as
Manning-Courtney et al., 2003], and having less animated affective expressions to whether and why autism seems to be
Rett syndrome [Glaze, 2004]? Is the out- [Baranek, 1999]. Lack of response to on the upswing, awareness has certainly
come different for those children who name seems to be the most consistent increased [Fombonne, 2003]. Several
have had a normal developmental trajec- singular symptom at this young age [Os- studies have demonstrated that early de-
tory prior to the onset of regression, as terling and Dawson, 1994; Baranek, tection and early intervention do have a
compared with those whose regression is 1999; Werner et al., 2000 Osterling et positive impact on outcomes for children
limited only to language or those who al., 2002]. with autistic spectrum disorders [Hoyson
were developmentally delayed even be- The findings from these studies, al- et al., 1984; Lovaas, 1987; Rogers and
fore the regression? Does the timing of though retrospective, document some of Lewis, 1989; Harris et al., 1991; Birn-
the regression impact outcome? Does au- the behavioral signs of autism in very brauer and Leach, 1993; McEachin et al.,
tism with regression differ in etiology, young children, and suggest that early 1993; Lord, 1995; Dawson and Oster-
impact, and prognosis than autism with- screening might be feasible. Ongoing re- ling, 1997; Smith et al., 1997; Jocelyn et
out regression? search is needed to uncover profiles of al., 1998; Sheinkopf and Siegel, 1998;
Various mechanisms of regression children with ASDs at young ages, since Smith and Lovaas, 1998; Harris and
have been suggested, including illness, the signs may be different from those Handleman, 2000; Sonnander, 2000]. In
sub-clinical seizures, and variations in seen in older children. Prospective stud- fact research suggests that intervention
postnatal brain development due either ies are needed with infants and toddlers before three or three and a half years of
to the impact of genetic factors or post- to further investigate the best means of age has the greatest impact. [Harris and
natal events [Lainhart et al., 2002; differentiating between variants of typical Handleman, 2000; Woods and Weth-
Hrdlicka et al., 2004]. A discussion of development and the earliest symptoms erby, 2003]. The findings that early iden-
these questions and issues is beyond the of autism. In the meantime, it would tification and intervention lead to im-
scope of this article and the readers are seem prudent to verify that children un- proved outcome suggest that autism is an
referred to a review of developmental der 12 months old being seen for their appropriate disorder for which screening
regression published in this journal well child care visit are looking at others should be undertaken. This realization
[Rogers, 2004]. However, there are two and orienting to their name, since these has prompted a focus on the develop-
issues pertinent to this discussion of two skills may be two of the earliest red ment of autism-specific screening tools
screening young children. Screening flags for possible autism. Baranek [1999] appropriate for use with young children.
tools need to ask about regression in a states that the early markers for autism Several issues regarding the screen-
way that is compatible with parent and noted on videotape review preceded the ing for autism in young children need
provider perspectives. It is also important voicing of concerns by parents. She also further clarification. What type of
to not only ask about regression, but lack mentions that the parents in her study screening is best: should the screening be
of developmental progression. Secondly, compensated for these markers, suggest- broad-based developmental screening, or
it is important to remember that autism ing that these compensatory behaviors on autism-specific? What is the best age at
screening will not capture regression if the parents’ part may somehow be reflec- which to screen? What is the best
the screening precedes the regression. tive of the child’s symptoms and a possi- method for screening: should parent re-
Retrospective studies have shown ble avenue for developing measures to port be used, or is direct observation a
that children who are later diagnosed help with the early identification of these better approach?
with autism exhibit symptoms as early as children.
8 –12 months of age [Osterling and Daw- Despite this delay in parents’ con- WHAT TYPE OF SCREENING
son, 1994 Baranek, 1999; Werner et al., cerns relative to the early markers iden- IS BEST?
2000; Osterling et al., 2002]. In general tified retrospectively, studies suggest that Since research suggests that the first
these studies reviewed videotapes of the parental concerns regarding their child’s indication of a problem in children who
children prior to or at their first birthday development were typically expressed to are subsequently diagnosed with an ASD
and analyzed their behaviors, as com- pediatricians by the age of 1.5 years, and may occur as early as the first year of life
pared with typically developing children yet, a definitive diagnosis of autism is not [Osterling and Dawson, 1994; Baranek,
or children with non-ASD diagnoses. made until approximately 4 years old. 1999; Werner et al., 2000; Osterling et
Some children with ASDs seem to ex- [De Giacomo and Fombonne, 1998; Sie- al., 2002] and that in the U.S., early
hibit symptoms as early as 8 months, but gel et al., 1988; Flannagan and Nuallain, intervention 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 needs this type of care, the child is viewed as der such as autism with every parent they
identified for children from birth to the the multi-faceted being that he or she is. encounter at a preset interval due to con-
age of three, screening all children once The fact that any identified concerns will cerns that the mere administration of
or twice during that timeframe would be raised by a provider with whom the these screens will be anxiety provoking
facilitate the timely identification of chil- family and child already have a relation- for some parents. Secondly, the question
dren who need early intervention and/or ship may lend additional credibility to the arises as to which instrument primary
monitoring. screening process and findings. There care providers should use. Screening in-
There are generally two models are, however, potential limitations with struments for children younger than
used to screen for autism. The first model this model. It depends on the primary 16 –18 months have not yet been vali-
is consistent with practice parameters en- care provider performing developmental dated, although multiple research groups
dorsed by the American Academy of Pe- surveillance/screening, recognizing the are working to identify reliable signs in
diatrics (AAP), as well as the American red flags for autism, and pursuing the infancy. Multiple versions of a screener
Academy of Neurology (AAN) [Filipek appropriate next steps in a timely man- would be needed for children of different
et al., 1999; Filipek et al., 2000AAP, ner. This possibility is especially prob- ages. Lastly, as the number of screening
2001a, 2001b, 2001c]. It consists of on- lematic as the frequency of well-child instruments for different conditions in-
going developmental surveillance of chil- care visits begins to decrease in the tod- crease, primary care providers may be
dren within the primary care setting. In dler years. Also, a potential challenge unable to fit them all into the limited
Dworkin [1993], developmental surveil- with this approach is the limited amount time available for the well-child care
lance is defined as: of time available to most providers for visit.
accomplishing a growing number of tasks Research suggests that the age at
“a flexible, continuous process whereby knowl- during the WCC visit. Perhaps due in
edgeable professionals perform skilled observations diagnosis seems to be influenced by so-
of children during the provision of health care. The part to this, in some instances, the di- cioeconomic status and/or race/ethnicity
components of developmental surveillance include mensions of the medical home are not [Mandell et al., 2002]. Furthermore,
eliciting and attending to parental concerns, obtain- fully implemented [Strickland et al., missed and misdiagnosed ASDs seem to
ing a relevant developmental history, making accu- 2004].
rate and informative observations of children, and
be more prevalent in certain populations
sharing opinions and concerns with other relevant
Although there are currently no [Mandell et al., 2002; Dyches et al.,
professionals (p. 533).” validated autism-specific screening in- 2004; Pinto-Martin and Levy, 2004].
struments designed for children less than The cause for this is not yet known, but
It is suggested that standardized 18 months old, several groups are work- may be linked to several factors related to
general developmental screening tools be ing to better understand how autism pre- racial and ethnic disparities in the quality
used as part of the developmental surveil- sents in that age group in a prospective of health care [Smedley et al., 2003],
lance process. These tools can either be a manner. Zwaigenbaum et al. [2005] re- including screening and referral practices.
parent-completed questionnaire or clini- cently published data from their ongoing There is also the possibility that the
cian-completed measure. If general prospective study with siblings of chil- symptoms of autism may differ across
screening raises concerns suggestive of dren diagnosed with an ASD. They have populations, not only in terms of which
possible autistic spectrum disorders, au- designed an observational scale for assess- symptoms are present and the timing of
tism-specific screening could then be ad- ing autism-specific behavior in children. presentation, but also which symptoms
ministered. The challenge of provider- This is a promising line of research whose concern parents, and whether they share
implemented screening was highlighted findings may lead to validated measures these concerns with pediatric providers.
by a recent study, which found that half for screening and diagnosing children less Providers need to ensure that all children
of the surveyed physicians used a formal than 18 months of age, or even less than in their care are being screened. Towards
screening tool (Sices et al., 2003). How- 12 months of age. At the present time, that end, whichever model is used to
ever, a primary benefit of using this providers should use validated general screen for autism should include prede-
model to screen for autism is that it in- developmental surveillance/screening termined steps. Empirical data is needed
corporates autism screening into an al- with their patients under 18 months old. to determine which model is most effec-
ready accepted practice within the field The second model consists of the tive. It is possible that different models
of pediatrics. As such, autism-specific routine administration of autism-specific will work best in different patient and
screening is less likely to represent an screens at multiple high-risk ages (e.g., 18 provider populations, and at different
additional item to be included in the and 24 months of age) by primary care points in development.
already overburdened well child care providers, regardless of the presence of
(WCC) visit. Another benefit is that it symptoms or concerns suggestive of an
allows the screening to be undertaken ASD. This would be done in addition to WHAT IS THE BEST AGE TO
within the context of the child’s “medi- developmental surveillance, including SCREEN?
cal home.” the use of general developmental screen- There are both advantages and dis-
The term “medical home” was first ing. The primary benefit of screening at advantages to early screening. On one
used in 1967 [Sia et al., 2004]. The several ages is that the likelihood of not hand, earlier screening may yield a diag-
American Academy of Pediatrics (AAP) recognizing a child with possible ASD is nosis and subsequent intervention at an
in a policy statement provided 7 dimen- diminished. There are three main limita- earlier age, allowing the maximum time
sions (accessible, continuous, compre- tions to this approach. First, it would for intervention. On the other hand, au-
hensive, family-centered, compassionate, require already busy primary care provid- tism-specific screening at less than 16 –18
culturally effective, and coordinated with ers to remember to administer an autism- months old has not yet been validated.
specialized services provided outside of specific screening instrument to all chil- Additionally, children who regress or
the primary care setting) that should be dren, even those who they feel are those with Asperger’s syndrome and high
found within a medical home [AAP, developing appropriately. Providers may functioning autism may be missed by
2002; Bethell et al., 2004]. As a result of hesitate to raise the possibility of a disor- early screening, as they often present
MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN 255
later. Furthermore, health care providers aged children with autism are compared naires.[Glascoe et al., 1989; Glascoe,
may be reluctant to screen for autism at with those with developmental delays, 1994; Glascoe and Dworkin, 1995; Glas-
younger ages because of concerns that children diagnosed with autism are found coe and Sandler, 1995; Glascoe, 1997].
such screening will lead to parental anx- to exhibit less pretend play skills than Finally, parental concerns regarding the
iety. Perhaps contrary to this idea is the those with developmental disabilities emotional, behavioral, cognitive, fine
finding that significant numbers of par- [Lord et al., 1994; Cox et al., 1999; Not- motor, and language development of
ents have concerns about how their erdaeme et al., 2000]. their children are predictive of the exis-
young child is developing, but if they are To further complicate the issue, tence of a true problem [Glascoe and
not asked, they may not share these con- many typically developing young tod- Dworkin, 1995].
cerns with their child’s primary care pro- dlers may show behaviors consistent with Efforts to increase public awareness
vider [Young et al., 1998; King and Glas- those seen in ASD, such as repetitive of developmental disorders such as autism
coe, 2003; Bethell et al., 2004]. A core behaviors (e.g., opening and closing include the Centers for Disease Control
principle embedded in the concepts of drawers and doors, turning lights on and and Prevention (CDC)’s Learn the Signs
developmental surveillance, the medical off), repetitive motor behaviors such as campaign (www.cdc.gov/actearly; 1– 800-
home, and early identification is eliciting hand flapping when excited, and insis- CDC-INFO), as well as a comprehensive
and responding to parental concerns. tence on routines (and indeed a certain parent-developed site (www.firstsigns.org).
Studies have demonstrated the impor- sense of consistency and routine is crucial These campaigns are aimed at educating
tance of parental report in the early de- to minimizing frustration in that age parents about child development and the
tection of developmental problems group). warning signs of developmental disorders
[Stone and Lemanek, 1990; Stone et al., such as autism. As part of the campaign,
1994; Glascoe and Dworkin, 1995; WHAT IS THE BEST METHOD both the CDC and First Signs have devel-
Young et al., 2003]. It is important that FOR SCREENING? oped kits with information about screening
providers respond to parental concerns Several factors must be considered and early development for pediatric pro-
and yet, a recent study found that expres- in contemplating the best approach to viders. It is always possible that parents of
sion of parental concerns did not lead to autism screening. These include the typically developing children may unnec-
increased likelihood of the child being availability, cost, ease of administration of essarily become concerned as a result of
referred for the purpose of diagnosis or instruments with acceptable sensitivity/ public awareness campaigns such as Learn
service provision [Sices et al., 2004]. specificity, timing of well-child care visits the Signs. In such instances, reassurance and
Screening at later ages, such as for children, the amount of time available parent education may be all that is required.
screening after 18 –24 months, may lead for these visits, the general approach to More research is needed to ascertain
to missed opportunities for early inter- developmental screening and use of non- whether the nature of the concerns raised
vention. However, this later screening physician personnel (e.g., nurse practitio- by parents of children found to have de-
will likely have better specificity and sen- ners, physical assistants, home visitors), velopmental disorders differ from those
sitivity. In addition, it may capture the cost-effectiveness, reimbursement, and raised by the parents of typically developing
children who develop autistic features at the cultural expectations of the popula- children.
later ages, those who regress, or are tion served regarding screening and diag- The advantages of professional ob-
higher functioning. Also, health care nosis both in terms of the process and the servation include knowledge of the full
providers may feel more comfortable content. range of what constitutes typical devel-
with the idea of screening older children. Two methods often used in devel- opment. Clinicians are more objective
Another challenge to early diagnosis is opmental screening are parent reports than parents and may be less likely to
the possibility that the provider may be and observations by trained clinicians; over- or under-estimate a given child’s
concerned when the parents are not. each has advantages and drawbacks. Pa- skills or problem behaviors. The signifi-
Since participation in early intervention rental report is encouraged by many pri- cance of a given behavior may not be as
is voluntary, unless the parents agree to marily because parents know their chil- clear to a parent as it might be to a trained
receive the services, the child will not dren best. They spend the most time observer. Finally, in many instances, the
have the opportunity to receive them. with them in a variety of settings. This, setting used for undertaking the observa-
Some parents may even refuse to allow therefore, provides them with many op- tions can be standardized.
their child to undergo the evaluative pro- portunities to observe the child across
cess. In such instances, repeated screen- multiple settings, interacting with various GENERAL DEVELOPMENTAL
ings may increase the likelihood that the individuals while they are in a variety of SCREENING TOOLS
parents will be convinced that the con- moods. This method, then, does not rely There are many broad develop-
cerns being identified are worth further on a brief and what may be atypical sam- mental screening tools that may have a
investigation, and screening at an older ple of behaviors, and allows for observa- role in the early identification process.
age may be less likely to encounter pa- tion of behaviors that may not be observ- Discussion of these many tools is beyond
rental denial of concern. able in most primary care offices, such as the scope of this article, and the reader is
In deciding on what age is most interest in peers. referred to a recent review article on the
appropriate for autism-specific screening, Well-developed parent checklists subject [Glascoe, 2000]. This review fo-
it is important to remember that different are easy to administer, and in the short cuses briefly on two of the tools widely
items may be needed for screening at term may be the most cost effective op- used by pediatric providers and one that
different ages. For example, some studies tion available to primary pediatric pro- focuses on communication and symbolic
suggest that toddlers diagnosed with au- viders. Developmental surveillance, behavior. The three measures are the
tism and those with developmental delays which includes clinical judgment, is Parents’ Evaluation of Developmental
both have impaired pretend play [Baron- aided by the systematic elicitation of pa- Status (PEDS), the Ages and Stages
Cohen et al., 1996; Charman et al., 1998; rental concerns across developmental do- Questionnaire (ASQ), and the Commu-
Cox et al., 1999]. Yet, when preschool mains, using validated parent question- nication and Symbolic Behavior Scales
256 MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN
Developmental Profile (CSBS DP). As symbolic (understanding and object use) at 18 months of age. It consists of nine
screens of general development some of [Wetherby et al., 2004]. The 24 item parent report items (A1–9) and five ob-
these tools may not target the core signs Infant-Toddler checklist asks about de- servation items (Bi-v) [Baron-Cohen et
and symptoms of autism, for example, velopmental milestones (e.g., when you al., 2000; 1996; 1992]. When looked at
the PEDS does not include questions look at and point to a toy across the together five of the fourteen items have
about play and imitation. room; does your child look at it?; does been found to be key indicators of pos-
The PEDS is a ten item instrument your child point to objects?). Nineteen of sible autism based on the Diagnostic and
designed to be self-administered by par- the twenty four items have the answer Statistical Manual (DSM) criteria for au-
ents of children from birth to 8 years of choices: not yet, sometimes, and often, tistic disorder: gaze monitoring (Bii);
age, and elicits parent concern in devel- the remaining five are ‘how many’ ques- proto-declarative pointing (pointing to
opmental domains such as fine and gross tions. At the bottom of the form parents indicate interest rather than to request)
motor skills, receptive and expressive lan- are also asked a yes/no question: do you (A7 ⫹ Biv), and pretend play (A5 ⫹
guage, and self-help skills. The type of have any concerns about your child’s de- Biii). Baron-Cohen et al., [1996] found
parental concerns leads to categorization velopment? They are asked to describe that if at 18 months the items gaze mon-
of risk for developmental difficulties: any existing concerns they may have. It is itoring, proto-declarative pointing, and
high, moderate, and low with corre- designed to be completed by the parent pretend play were failed, there was an
sponding responses such as refer, reassure, or caregiver. The Behavior Sample is a 83.3% risk of being subsequently diag-
and offer developmental promotion and videotaped evaluation of the child during nosed with autism via a diagnostic eval-
parent education. The PEDS has been an interaction with his/her parent and uation. A follow up study [Baird et al.,
standardized and validated and meets the the clinician. Wetherby et al., [2004] 2000] with 16,235 eighteen month old
recommended psychometric properties propose the use of the Infant-Toddler children using the CHAT found, how-
for a screening instrument (sensitivity and Checklist as a first level screen, with the ever, that although the specificity of the
specificity greater than 0.70) [Glascoe, Behavior Sample serving as a second level CHAT was excellent, the sensitivity
2001; Glascoe, 2003; www.pedstest. evaluation tool. The tool was not de- (20 –38% depending on criteria used) was
com]. signed to be an autism-specific screening unacceptably low. This suggests that the
The Ages and Stages Question- tool, but rather a tool to screen for delays CHAT may not be an appropriate tool to
naire (ASQ) is developed for use with in communication, including prelinguis- use as an exclusive screening tool for
children 4 – 60 months old, parents are tic skills; as such, it may be sensitive to a identifying children who may have an
asked to respond to descriptions of a list range of social communicative delays, in- ASD, and is currently under revision by
of skills with the response options: not cluding autistic spectrum disorder. the authors.
yet, sometimes, or yes [Bricker and The PEDS, ASQ, and CSBS DP
Squires, 1999]. Each form is intended for are not designed to selectively screen for PDDST-II
use with a certain age group (e.g., 4 years autism, but they may be effective in de- PDD ST-II [Siegel, 2004] consists
old) and consists of thirty developmental tecting children whose developmental of 3 stages designed to be used in three
items divided into five areas: communi- problems are consistent with autism. different clinical settings. The PDDST
cation, gross motor, fine motor, problem Screening a large sample of children from Stage 1 consists of 22 items with the
solving, and personal-social. There is also an unselected population will be needed response choices ‘yes, usually true’ or
a section for general parental concerns. to demonstrate their usefulness as pri- ‘no, usually not true.’ Its intended use is
Many of the items are accompanied by mary screens for children with autism. in the primary care setting with 12– 48
illustrations, and the reading level is sixth month olds. A positive screen signals the
grade or below. The items were selected SCREENING TOOLS FOR need for further evaluation and is defined
based on the likelihood of being ob- AUTISTIC SPECTRUM as 5 or more items being answered ‘yes,
served or elicited by parents in the home DISORDERS (ASD) usually true.’ The author reports a sensi-
setting. The ASQ reportedly has a range There are several autism-specific tivity of 0.92 and 0.91 based on a sample
of classification agreement from 76 to tools currently under development, or of 681 children ‘at risk for ASD’ and 256
91% [Bricker and Squires, 1999], with revision, and as such it would be prema- children with ‘mild-to-moderate other
developmental assessment tools such as ture to recommend one. Many of these developmental disorders’.
the Bayley Scales of Infant Development promising tools still need to undertake The other two stages of the
[Bayley, 1993]. continued follow up of the original sam- PDDST are intended to be used in set-
The Communication and Sym- ples used in their development, as well as tings other than the primary care provid-
bolic Behavior Scales Developmental cross-validation with new samples. er’s office. Stage 2 is a 14-item screening
Profile (CSBS DP, Wetherby and Pri- In addition to the Modified tool for use in developmental clinics. The
zant, 2002] includes three measures: the Checklist for Autism in Toddlers (M- cut-off is still 5 items with a reported
Infant-Toddler checklist, an expanded CHAT), which will be discussed in more sensitivity and specificity of 0.73 and 0.49
Caregiver Questionnaire, and a Behavior detail later, four other autism-specific based on 490 children with confirmed
Sample. The CSBS DP is based on a tool screening tools are the Checklist for Au- ASD (Autism, PDD-NOS, or Asperger’s
previously designed by the same authors tism in Toddlers (CHAT), Pervasive De- syndrome) and 194 children who were
the CSBS [Wetherby and Prizant, 1993]. velopmental Disorders Screening Test-II evaluated for an ASD, but who did not
The CSBS DP is a screening and evalu- (PDDST-II), Screening Tool for Autism receive a diagnosis on the autistic spec-
ation instrument designed to measure the in Two year olds (STAT), and Checklist trum [Siegel, 2004]. Stage 3 is a 12 item
communicative and symbolic abilities of for Autism in Toddlers-23 (CHAT-23). screening tool designed to be adminis-
children aged 12–24 months. The mea- tered in autism clinics with a cut-off
sured skills form three composites: social CHAT score of 8. The reported sensitivity and
(emotion, eye gaze, and communica- The CHAT is a screening tool in- specificity is 0.58 and 0.60, respectively
tion), speech (sounds and words), and tended for use in the general population [Siegel, 2004]. The sensitivity and spec-
MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN 257
ificity reported for stage 1 are good. with autism and those without for mental Those who fail in both the initial
Those for stage 2 and 3 are not as good, ages and assess the sensitivity and speci- screening and the telephone follow up
particularly, the specificity for stage 2. ficity of the STAT in that instance. Nev- are evaluated by a team of evaluators,
The major statistical shortcoming of this ertheless, the STAT is a promising Level using the Autism Diagnostic Interview-
instrument is that no sensitivity/specific- 2 screening tool for autism in children Revised (ADI-R) [Lord et al., 1994];
ity data are reported for large-scale between two and three years old. ADOS-G [Lord et al., 2000; 1999];
screening of an unselected sample. CARS [Schopler et al., 1988]; Mullen
Scales of Early Learning [Mullen, 1995];
STAT The M-CHAT Vineland Adaptive Behavior Scales
The STAT [Stone et al., 2000; The M-CHAT [Robins et al., [Sparrow et al., 1984], and the DSM-IV
Stone et al., 2004] is designed as a second 2001] is a 23-item (yes/no) parent report criteria [APA, 1994]. The ADOS is a
level screening tool for autism. It is in- checklist developed to screen children semistructured instrument, which allows
tended to be used by professionals in aged 16 –30 months. As suggested by the assessment of a child’s communication
communities to identify children with name; the M-CHAT is a modification of and social interactions. The ADI is a par-
possible autism, as opposed to other de- CHAT [Baron-Cohen et al., 1992], de- ent report measure with questions on the
velopmental disabilities. The items are scribed above. The M-CHAT expands child’s communication, social related-
administered within a twenty-minute upon the parent report portion of the ness, including play and interests and be-
long play-based interactive session. The CHAT, and seeks to identify children haviors. The ADI and ADOS are cur-
ten scored items include play (2), imita- with a possible autistic spectrum disorder, rently considered the gold standard for
tion (4), directing attention (4), and do including PDD-NOS, and not just diagnosing autism.
not require language comprehension. strictly defined Autistic Disorder. The number screened at Time 1
Two additional requesting items are in- The initial population screened by (completed form between 16 and 30
cluded as a means of promoting interac- the M-CHAT consisted of 1,293 chil- months old) has been increased to 4,200
tion between the evaluator and child. dren aged 16 –30 months [Robins et al., children. 236 children who failed the ini-
Two samples were used to assess the va- 2001]. Of the 1293 children screened 58 tial screening have been evaluated at a
lidity of the STAT as a second level were evaluated and 39 were diagnosed mean age of 27.63 months: 165 were
screening instrument. The sample used in with an autistic spectrum disorder. None found to have an ASD, 67 have been
the development phase consisted of 40 of the children evaluated at that time diagnosed with a non-ASD developmen-
children (3 with autism, 33 without), the were found to be typically developing. tal disorder, and 4 have been assessed as
sample used for validation had 33 chil- Although absolute sensitivity and speci- typically developing. To date 1,937 chil-
dren (12 with autism, 21 without au- ficity are pending completion of follow dren evaluated at Time 1 have reached
tism). The children in the two samples up of the initial sample, based on the eligibility for rescreening at age 4. Data
ranged in age from 24 to 35 months old. discriminant function analysis (DFA) has been collected from 940 (mean age ⫽
The children with autism had lower de- classification at the time of first screening 55.35 months). At rescreening, parents
velopmental ages than the children with- and evaluation, the M-CHAT has a sen- are asked to complete the M-CHAT, and
out autism. Data from a subsample of sitivity of 0.87, specificity of 0.99, posi- are also asked whether in the interim the
children with (12) and without (12) au- tive predictive power of 0.80, and nega- child has been referred for or diagnosed
tism matched for developmental age was tive predictive power of 0.99 [Robins et with an ASD. Six children have been
also analyzed. The two examiners admin- al., 2001]. identified as possible misses (failed the
istering the STAT were blind to the re- Excluded from the ongoing valida- rescreening or referred for possible ASD):
ferral questions and the results of the di- tion study are children who already have 2 of the 6 have been confirmed as misses,
agnostic assessments. The parents of the an ASD diagnosis prior to completion of two are not misses, and two are classified
participating children were blind to the the M-CHAT, and children younger as possible misses because they have not
results of the diagnostic evaluation until than 16 months or older than 30 months been able to be evaluated. Sixty three
after the STAT screening was completed. at the time of M-CHAT completion. children evaluated at Time 1 have been
The diagnoses were made based on The sample is drawn from primary care reevaluated at Time 2. At Time 2, 38 of
DSM-IV criteria (APA, 1994) and the practices, as well as early intervention the 63 children have been diagnosed
Childhood Autism Rating Scale (CARS; sites. Children who fail the M-CHAT with an ASD, 17 have been diagnosed
Schopler et al., 1988]. The sensitivity and receive a confirmatory follow up tele- with a non-ASD developmental disor-
specificity of the subsample matched for phone call at which time the completed der, and 8 have been assessed as typically
DA were each 0.83. The sensitivity and M-CHAT is reviewed to clarify the se- developing. Based on the M-CHAT
specificity for the validation sample were lected answers and ascertain whether the score at Time 1 and the evaluation out-
0.83 and 0.86, respectively. child has in fact screened positive on the come at Time 2, the sensitivity and spec-
The authors of the STAT highlight checklist. The completed forms are ificity are 0.85 and 0.93, respectively.
three areas needing additional research. scored and the child is considered to have (The sensitivity is as high as 0.95 if the
Firstly, replication studies with larger failed the initial screening if he/she fails possible misses are not included as
samples from many different settings are any three of the twenty three items or misses). It is possible that the psychomet-
needed before the findings can be gener- two of the six critical items. The critical ric properties of the M-CHAT will be
alized. Secondly, future studies need to items on the M-CHAT are as follows: different when larger numbers from an
use standardized measures such as the item 2 (interest in other children), item 7 unselected population are obtained.
Autism Diagnostic Observation Schedule (proto-declarative pointing), item 9 Although this is still a work in
–Generic (ADOS-G) [Lord et al., 1999], (bringing objects to show the parent), progress, based on the data collected to
in addition to clinical judgment. Finally, item 13 (imitating), item 14 (responding date, certain items appear to be helpful in
follow up studies with larger sample sizes to name), and item 15 (following a distinguishing between children with
need to match those children diagnosed point). ASD as compared to those with a non-
258 MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN
ASD developmental disorder [Dixon et dition to the 23 item-questionnaire, the Since many of the autism-specific
al., in press]. In particular, item 7 ‘point CHAT-23 has an observational section tools for screening in young children are
for interest’ seem to be the most potent consisting of 5 direct observational items still under development, close attention
discriminator between the two groups of from the CHAT. When used with 212 needs to be paid to their psychometric
children. children (87 with ASD and 125 non- properties in different SES and ethno-
The diagnoses made at 2 years old ASD) with the mental age range of cultural groups. The M-CHAT is one of
also appear to be relatively stable based 18 –24 months, Wong et al. found that a few tools, which show promise as a
on the available follow up data. Although failing any 2 of 7 key questions (imitate, screening tool in different populations of
in a few cases, children moved from a pretend, point for interest, check reac- unselected children. It has been translated
non-ASD diagnosis to the typical devel- tion of parent, bring objects to show, into Turkish, Chinese, Japanese, and
opment category, most of the changes in follow a point, and take interest in other Spanish. It will be important to evaluate
diagnosis were from an ASD to a non- children) or any 6 of the overall 23 ques- the findings of studies using the
ASD developmental disorder. Whether tions in the parent administered ques- M-CHAT in these different languages
this change in diagnosis is due to mis- tionnaire yielded a sensitivity of 0.931 and cultures for similarities and differ-
diagnosis at Time 1 or is the result of and 0.839, respectively. The specificity ences in their findings. Beyond the lan-
early intervention and/or increasing ma- was of 0.768 or 0.848, respectively. The guages being used, different cultural and
turity remains unclear at this time. authors suggest a two-level screening sys- socio-economic groups may interpret the
Part of the Early Detection study is tem beginning with the M-CHAT and questions differently, yielding different
the screening of the siblings of children proceeding to the use of the observa- results. Not only may the specific signs/
with ASDs. Recent analyses of data col- tional portion of the CHAT for those symptoms of autism vary across cultural
lected from younger siblings of children who failed the M-CHAT. If the obser- groups, the meaning attributed to them
diagnosed with an ASD found an initial vation is failed, the child would then be by parents may as well. For example, the
screen positive rate (before phone inter- referred for an autism evaluation. The fact that a child does not point or make
view) of 43%. As the M-CHAT is a fact that the psychometric properties be- eye contact may be considered appropri-
parent-report measure, it is possible that ing found when the CHAT-23 is used ate and acceptable if to point and make
these figures are partially inflated because with a sample of Chinese children so eye contact with adults is considered a
of increased parental concern as a result closely resemble what is being found sign of disrespect. Therefore, more work
of the previous ASD diagnosis of the with the M-CHAT supports the validity focused on understanding what concerns
older sibling. After the phone interview, of these instruments. In Wong et al.’s compatible with ASDs may be more uni-
30% remained a positive screen. Twenty study, the chronologic ages of the en- versal than others is needed.
three percent of the younger siblings rolled children whose mental ages were Despite these unresolved issues, a
were evaluated and diagnosed with an 18 –24 months old were 16 – 86 months few facts remain clear. First, primary care
ASD. This is a higher rate than what has old. It is important to note that when providers are the appropriate people to
been reported in the literature for sib- used with children whose mental ages are screen children between birth and the
lings, which may be due to an ascertain- not known to be 18 –24 months, the age of 5 years old, since they are the
ment bias, the parents of the siblings in efficacy of the tool may be different. In qualified professionals with whom par-
this study may have sought participation addition, this range of chronologic ages is ents have ongoing contact. Second, the
in the study as a result of concerning signs quite wide, future research should exam- well-child care visit provides only a lim-
that they observed. ine whether the screener is equally valid ited amount of time for the screening of
The study also involves screening across this age range. children, with each of the many disorders
of children within low risk settings (pri- potentially requiring that their own
mary care provider sites, Peds) and high screening tool and process be used. The
risk settings (early intervention sites, EI). CONCLUSION tools that will be used by providers are
Based on 20 children screened through At this time, early intervention is likely to be those that accurately flag chil-
pediatric offices, the items that discrimi- the best response to Autistic Spectrum dren for many developmental disorders.
nate the children with Peds-referred Disorders. Given this fact, the develop- Third, in light of the data supporting the
ASD from the EI-referred ASD appear to ment of reliable tools for screening chil- usefulness of eliciting parental concerns,
be highly similar. dren under three years old is imperative. any tool used will probably need to in-
Early screening will lead to early diagno- corporate the parent’s perspectives.
CHAT-23 sis, early intervention, and the ultimate Fourth, data suggests that children of par-
Wong et al. [2004] used the M- goal–improved outcomes. There remains ents from minority racial/ethnic back-
CHAT and CHAT to develop a screen- the need for further research to elucidate grounds regardless of SES and those with
ing tool for use with a sample of Chinese many issues. When is the best age to lower socio-economic status may not be
children, the CHAT-23. The changes screen? Are the diagnoses made in the receiving quality health care. This dispar-
made to the original M-CHAT were the toddler years stable? If the children sub- ate care includes a lack of screening, eval-
translation of the questions into “tradi- sequently improve, is it due to the effects uation/diagnosis, and service provision.
tional chinese” and the addition of a of early intervention or was the early This must be addressed promptly, since
graded scoring system [never (0%), sel- diagnosis incorrect? What is the best way the literature supports the notion that the
dom (⬍25%), usually (25–50%), or often to screen for autism? Is preliminary gen- earlier the identification of and interven-
(⬎50%)] for 22 of the 23 M-CHAT eral developmental surveillance/screen- tion for developmental disorders, the
questions instead of the yes/no responses ing best or is autism-specific screening better the outcomes.
of the original M-CHAT. The graded needed for all children? Are there popu-
system was subsequently collapsed into lations, such as the siblings of children ACKNOWLEDGMENTS
yes (usually/often) and no (never/sel- with ASDs, for whom autism-specific We thank the Early Detection Team
dom) grouped to define fail/pass. In ad- tools are best? at the University of Connecticut (Mari-
MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN 259
anne Barton, Sarah Hodgson, James Green, the Brick Township, New Jersey, investiga- children who failed a screening instrument for
Gail Marshia, Pam Dixon, Jamie Klein- tion. Pediatrics 108:1155–1161. ASD. J Autism Dev Disord.
Bethell C, Reuland CH, Halfon N, et al. 2004. Dworkin PH. 1993. Detection of behavioral, de-
man, Juhi Pandey, Hilary Boorstein, Measuring the quality of preventive and de- velopmental and psychosocial problems in pe-
Emma Esser, Leandra Wilson, Michael velopmental services for young children: na- diatric primary care practice. Curr Opin Pe-
Rosenthal, Saasha Sutera, and Alyssa Ver- tional estimates and patterns of clinicians’ per- diatr 5:531–536.
balis); our collaborators at the Yale Univer- formance. Pediatrics 113:1973–1983. Dyches TT, Wilder LK, Sudweeks RR, et al. 2004.
Birnbrauer JS, Leach DJ. 1993. The Murdoch early Multicultural issues in autism. J Autism Dev
sity Child Study Center (Fred Volkmar, intervention program after 2 years. Behav
Ami Klin, Katarzyna Chawarska, Tammy Disord 34:211–222.
Change 10: 63–74. Filipek P, Accardo P, Ashwal S, et al. 2000. Practice
Babbitz), the University of Washington Bricker D, Squires J. 1999. Ages and Stages Ques- parameter: screening and diagnosis of autism:
(Geraldine Dawson, Karen Toth), and Di- tionnaires: A parent-completed child–moni-
a report of the quality standards committee of
ana Robins at Georgia State University. toring system. 2nd ed. Baltimore: Paul H.
the American Academy of Neurology and the
Brookes Publishing Co.
Most importantly, we thank the families Campbell M, Locascio JJ, Choroco MC, et al.
Child Neurology Society. Neurology
and children who are participating in the 1990. Stereotypies and tardive dyskinesia: ab- 55:468 – 479.
ongoing Early Detection Study. Special normal movements in autistic children. Psy- Filipek P, Accardo P, Baranek G, et al. 1999. The
chopharmacol Bull 26:260 –266. screening and diagnosis of autistic spectrum
thanks to the library staff at St. Francis disorders. J Autism Dev Disord 29:439 – 484.
Hospital in Hartford, CT for their ongoing CDC’s Learn the Signs. Available at www.cdc.gov/
actearly; 1– 800-CDC-INFO First Signs. Available at www.firstsigns.org
assistance. Chakrabarti S, Fombonne E. 2001. Pervasive de- Flanagan O, Nuallain S. 2001. A study looking at
velopmental disorders in preschool children. the effectiveness of developmental screening
REFERENCES JAMA 285:3093–3099. in identifying learning disabilities in early
American Academy of Pediatrics. 2001a. Technical Chakrabarti S, Fombonne E. 2005. Pervasive de- childhood. Ir Med J 94:148 –150.
report: the pediatrician’s role in the diagnosis velopmental disorders in preschool children: Fombonne E. 2003. Epidemiological surveys of
and management of autistic spectrum disorder confirmation of high prevalence. Am J Psy- autism and other pervasive developmental
in children. Pediatrics 107(5). chiatry 162:1133–1141. disorders: an update. J Autism Dev Disord
American Academy of Pediatrics, Committee on Charman T, Baird G. 2002. Practitioner review: 33:365–382.
Children with Disabilities. 2001b. Develop- diagnosis of autism spectrum disorder in 2- Glascoe F. 1994. It’s not what it seems. The rela-
mental surveillance and screening of infants and 3-year-old children. J Child Psychol Psy- tionship between parents’ concerns and chil-
and young children. Pediatrics 108:192–196. chiatry 43:289 –305. dren with global delays. Clin Pediatr 33:292–
American Academy of Pediatrics, Committee on Charman T, Swettenham J. 2001. Repetitive be- 296.
Children with Disabilities. 2001c. The pedi- haviors and social-communicative impair- Glascoe F. 2001. Parents’ Evaluation of Develop-
atrician’s role in the diagnosis and manage- ments in autism: implications for develop- mental Status. Nashville, TN: Ellsworth and
ment of autistic spectrum disorder in children. mental theory and diagnosis. In: Burack JA, Vandermeer Press, LLC. www.pedstest.com
Pediatrics 107:1221–1226. Charman T, Yirmiya N, Zelazo PR, editors. Glascoe F. 2003. Parents’ evaluation of develop-
American Academy of Pediatrics, Medical Home The Development of Autism: Perspectives mental status: how well do parents’ concerns
Initiatives for Children with Special Needs from Theory and Research. Mahwah, NJ:
identify children with behavioral and emo-
Project Advisory Committee. 2002. The Erlbaum. p 325–345.
tional problems? Clin Pediatr 42:133–138.
Medical home. Pediatrics 110:184 –186. Charman T, Swettenham J, Baron-Cohen S, et al.
Glascoe FP. 1997. Parents’ concerns about chil-
American Psychiatric Association (APA). 1994. Di- 1998. An experimental investigation of social-
cognitive abilities in infants with autism: clin- dren’s development: prescreening technique
agnostic and Statistical Manual of Mental dis- or screening test? Pediatrics 99:522–528.
orders, 4th edition (DSM-IV). Washington, ical implications. Infant Ment Health J 192:
260 –275. Glascoe FP. 2000. Early detection of developmental
DC: American Psychiatric Association.
Courchesne E, Redcay E, Kennedy D. 2004. The and behavioral problems. Pediatr Rev
Baird G, Charman T, Baron-Cohen S, et al. 2000.
A screening instrument for autism at 18 autistic brain: birth through adulthood. Curr 21:272–280.
months of age: a six-year follow-up study. Opin Neurol 17:489 – 496. Glascoe FP, Dworkin PH. 1995. The role of par-
J Am Acad Child Adolesc Psychiatry 39:694 – Cox A, Klein K, Charman T, et al. 1999. Autism ents in the detection of developmental and
702. spectrum disorders at 20 and 42 months of behavioral problems. Pediatrics 95:829 – 836.
Ballaban-Gil K, Tuchman R. 2000. Epilepsy and age: stability of clinical and ADI-R diagnosis. Glascoe FP, Sandler H. 1995. Value of parents’
epileptiform EEG: association with autism J Child Psychol Psychiatry 40:719 –732. estimates of children’s developmental ages.
and language disorders. Ment Retard Dev Dahlgren SO, Gillberg C. 1989. Symptoms in the Disabil Rehabil 127:831– 835.
Disabil Res Rev 6:300 –308. first two years of life: a preliminary population Glascoe FP, Altemeier WA, MacLean WE. 1989.
Baranek GT. 1999. Autism during infancy: a ret- study of infantile autism. Eur Arch Psychiatry The importance of parent’s concerns about
rospective video analysis of sensory-motor Neurol Sci 238:169 –174. their child’s development. Am J Dis Child
and social behaviors. J Autism Dev Disord Dawson G, Osterling J. 1997. Early intervention in 143:955–958.
29:213–224. autism. In: Guralnick MJ, editor. The Effec- Glaze D. 2004. Rett syndrome: of girls and mice–
Barbaresi WJ, Katusic SK, Colligan RC, et al. tiveness of Early Intervention. Baltimore, lessons for regression in autism. Ment Retard
2005. The incidence of autism in Olmsted MD: Paul H. Brookes. p 307–326. Dev Disabil Res Rev 10:154 –158.
county, Minnesota, 1976 –1997: results from Dawson G, Meltzoff AN, Osterling J, et al. 1998. Goldberg WA, Osann K, Filipek PA, et al. 2003.
a population-based study. Arch Pediatr Ado- Children with autism fail to orient to natu- Language and other regression: assessment
lesc Med 159:37– 44. rally occurring social stimuli. J Autism Dev and timing. J Autism Dev Disord 33:607–
Baron-Cohen S, Allen J, Gillberg C. 1992. Can Disord 28:479 – 485. 616.
autism be detected at 18 months? The needle, Dawson G, Munson J, Estes A, et al. 2002. Neu-
Harris S, Handleman J. 2000. Age and IQ at intake
the haystack, and the CHAT. Br J Psychiatry rocognitive function and joint attention abil-
as predictors of placement for young children
161:839 – 843. ity in young children with autism spectrum
with autism: a four to six-year follow-up. J
Baron-Cohen S, Cox A, Baird G, et al. 1996. disorder versus developmental delay. Child
Dev 73:345–358. Autism Dev Disord 30:137–142.
Psychological markers in the detection of au-
tism in infancy in a large population. Br J Dawson G, Toth K, Abbott R, et al. 2004. Early Harris SL, Handleman JS, Gordon R, et al. 1991.
Psychiatry 168:158 –163. social attention impairments in autism: social Changes in cognitive and language function-
Baron-Cohen S, Wheelwright S, Cox A, et al. orienting, joint attention, and attention to ing of preschool children with autism. J Au-
2000. Early identification of autism by the distress. Dev Psychol 40:271–283. tism Dev Disord 21:281–290.
Checklist for Autism in Toddlers (CHAT). J DeGiacomo A, Fombonne E. 1998. Parental rec- Hoyson M, Jamisson B, Strain P. 1984. Individu-
R Soc Med 93:521–525. ognition of developmental abnormalities in alized group instruction of normally develop-
Bayley N. 1993. The Bayley Scales of Infant De- autism. Eur J Child Adolesc Psychiatr 7:131– ing and autistic-like children: the LEAP cur-
velopment, 2nd edition. San Antonio, TX: 136. riculum model. J Div Early Child 8:157–172.
Harcourt Brace. Dixon P, Kleinman J, Pandey J, et al. (in press). Hrdlicka M, Komarek V, Propper L, et al. 2004.
Bertrand J, Mars A, Boyle C, et al. 2001. Preva- Differentiating between autism spectrum dis- Not EEG abnormalities but epilepsy is asso-
lence of autism in a United States population: orders and other developmental disabilities in ciated with autistic regression and mental

260 MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN
 functioning in childhood autism. Eur Child Osterling J, Dawson G, Munson J. 2002. Early Neuropsychol Dev Cogn C Child Neuropsy-
Adolesc Psychiatry 13:209 –213. recognition of 1-year-old infants with autism chol 8:195–228.
Jocelyn LJ, Casiro OG, Beattie D, et al. 1998. spectrum disorder versus mental retardation. Stone W, Coonrod E, Ousley O. 2000. Brief re-
Treatment of children with autism: a ran- Dev Psychopathol 14:239 –251. port: screening tool for autism in two-year-
domized control trial to evaluate a caregiver- Pinto-Martin J, Levy S. 2004. Early diagnosis of olds (STAT): development and preliminary
based intervention program in community autism spectrum disorders. Curr Treat Op- data. J Autism Dev Disord 30:607– 612.
day-care centers. J Dev Behav Pediatr 19: tions Neurol 391– 400. Stone WL, Coonrod EE, Turner LM, et al. 2004.
326 –334. Robins DL, Fein D, Barton ML, et al. 2001. The Psychometric properties of the STAT for
King T, Glascoe F. 2003. Developmental surveil- modified checklist for autism in toddlers: an early autism screening. J Autism Dev Disord
lance of infants and young children in pedi- initial study investigating the early detection 34:691–701.
atric primary care. Curr Opin Pediatr 15: of autism and pervasive developmental disor- Stone W, Hoffman E, Lewis S, et al. 1994. Early
624 – 629. ders. J Autism Dev Disord 31:131–144. recognition of autism: parental reports vs.
Klin A, Jones W, Schultz R, et al. 2002. Defining Robinson R, Baird G, Robinson G, et al. 2001. clinical observation. Arch Pediatr Adolesc
and quantifying the social phenotype in au- Landau-Kleffner syndrome: course and corre- Med 148:74 –179.
tism. Am J Psychiatry 1596:895–908. lates with outcome. Dev Med Child Neurol Stone W, Lee E, Ashford L, et al. 1999. Can autism
Klinger, Dawson & Renner. 2003. Autistic disor- 43:243–247. be diagnosed accurately in children under 3
der. In: Mash EJ, Barkley RA, editors. Child Rogers S. 2004. Developmental regression in au- years? J Child Psychol Psychiatry 40:219 –
Psychopathology. 2nd ed. New York: The tism spectrum disorders. Ment Retard Dev 226.
Guilford Press. p 409 – 454. Disabil Res Rev 10:139 –143. Stone W, Lemanek K. 1990. Parental report of
Lainhart JE, Ozonoff S, Coon H, et al. 2002. Au- Rogers S, Lewis H. 1989. An effective day treat- social behaviors in autistic preschoolers. J Au-
tism, regression, and the broader autism phe- ment model for young children with perva- tism Dev Disord 20:513–522.
notype. Am J Med Genet 113:231–237. sive developmental disorders. J Am Acad Strickland B, McPherson M, Weissman G, et al.
Lord C. 1995. Follow-up of two-year-olds referred Child Adolesc Psychiatr 28:207–214. 2004. Access to the medical home: results of
for possible autism. J Child Adolesc Psychiatr Schopler E, Reichler RJ, Renner BR. 1988. The the National Survey of Children with special
36:1365–1382. Childhood Autism Rating Scale. Los Angeles: health care needs. Pediatrics 113:1485–1492.
Lord C, Risi S, Lambrecht L, et al. 2000. The Western Psychological Services. Szatmari P, Bartolucci G, Bremmer R. 1989. As-
autism diagnostic observation schedule– ge- Sheinkopf SH, Siegel B. 1998. Home-based behav- perger’s syndrome and autism: comparison of
neric: a standard measure of social and com- ioral treatment of young children with au- early history and outcome. Dev Med Child
munication deficits associated with the spec- tism. J Dev Disord 28:15–23. Neurol 31:709 –720.
trum of autism. J Autism Dev Disord 30:205– Sia C, Tonniges TF, Osterhus E, et al. 2004. His- Travis L, Sigman M, Ruskin E. 2001. Links be-
223. tory of the medical home concept. Pediatrics tween social understanding and social behav-
Lord C, Rutter M, DiLavore P, et al. 1999. Autism 113:1473–1478. ior in verbally able children with autism. J
Diagnostic Observation Schedule-WPS edi- Sices L, Feudtner C, McLaughlin J, et al. 2003. Autism Dev Disord 31:119 –130.
tion. Los Angeles, CA: Western Psychologi- How do primary care physicians identify Trevathan E. 2004. Seizures and epilepsy among
children with language regression and autistic
cal Services. young children with developmental delays? A
spectrum disorders. J Child Neurol 19:S49 –
Lord C, Rutter M, Le Couteur A. 1994. Autism national survey. Dev Behav Pediatr 24:409 –
S57.
diagnostic interview-revised: a revised version 417.
Tuchman R, Rapin I. 1997. Regression in perva-
of a diagnostic interview for caregivers of Sices L, Feudtner C, McLaughlin J, et al. 2004.
sive developmental disorders: seizures and ep-
individuals with possible pervasive develop- How do primary care physicians manage chil-
ileptiform electroencephalogram correlates.
mental disorders. J Autism Dev Disord 24: dren with possible developmental delays? A
Pediatrics 99:560 –566.
659 – 685. national survey with an experimental design.
Turner M. 1999. Annotation: repetitive behavior
Lord C, Shulman C, DiLavore P. 2004. Regression Pediatrics 113:274 –281.
in autism: a review of psychological research.
and word loss in autistic spectrum disorders. Siegel B. 2004. Pervasive Developmental Disorders
J Child Psychol Psychiatr 40:839 – 849.
J Child Psychol Psychiatr 45:936 –955. Screening Test –II (PDDST-II): Early Child- Volkmar F, Cohen D, Paul R. 1986. An evaluation
Lovaas O. 1987. Behavioral treatment and normal hood Screeners for Autistic Spectrum Disor- of DSM-III criteria for infantile autism. J Am
educational and intellectual functioning in ders. Harcourt Assessment, Inc. Acad Child Psychiatr 25:190 –197.
young autistic children. J Consult Clin Psy- Siegel B, Pliner C, Eschler J, et al. 1988. How Volkmar F, Rutter M. 1995. Childhood disintegra-
chol 55:3–9. children with autism are diagnosed: difficul- tive disorder: results of the DSM-IV autism
Malhotra S, Gupta N. 1999. Childhood disintegra- ties in identification of children with multiple FIELD trial. J Am Acad Child Adolesc Psy-
tive disorder. J Autism Dev Disord 29:491– developmental delays. J Dev Behav Pediatr chiatr 34:1092–1095.
498. 9:199 –204. Volkmar FR, Lord C, Bailey A, et.al. 2004. Autism
Mandell DS, Listerud J, Levy SE, et al. 2002. Race Siperstein R, Volkmar F. 2004. Brief report: pa- and pervasive developmental disorders.
differences in the age at diagnosis among rental reporting of regression in children with J Child Psychol Psychiatr 45(1): 135–170.
Medicaid-eligible children with autism. J Am pervasive developmental disorders. J Autism Werner E, Dawson G, Osterling J, et al. 2000. Brief
Acad Child Adolesc Psychiatr 41:1447–1453. Dev Disord 34:731–734. report: recognition of autism spectrum disor-
Manning-Courtney P, Brown J, Molloy CA, et al. Smedley BD, Stith AY, Nelson AR. 2003. Unequal der before one year of age: a retrospective
2003. Diagnosis and treatment of autism spec- Treatment: Confronting Racial and Ethnic study based on home videos. J Autism Dev
trum disorders. Curr Probl Pediatr Adolesc Disparities in Health Care. National Acade- Disord 30:157–162.
Health Care 33:283–304. mies Press. Wetherby A, Prizant B. 1993. The Communica-
McEachin J, Smith T, Lovaas O. 1993. Long-term Smith T, Eikeseth S, Klevstrand M, et al. 1997. tion and Symbolic Behavior Scale. New
outcome for children with autism who re- Intensive behavioral treatment for preschool- York, NY: Riverside.
ceived early intensive behavioral treatment. ers with severe mental retardation and mental Wetherby A, Prizant B. 2002. Communication and
Am J Ment Retard 97:359 –372. retardation and pervasive developmental dis- Symbolic Behavior Scales Developmental
McEvoy R, Rogers S, Pennington B. 1993. Exec- order. Am J Ment Retard 102:238 –249. Profile–First Normed Edition. Baltimore,
utive function and social communication def- Smith T, Lovaas O. 1998. Intensive and early be- MD: Paul H. Brookes.
icits in young autistic children. J Child Psy- havioral intervention with autism: the UCLA Wetherby AM, Woods J, Allen L, et al. 2004. Early
chol Psychiatr 34:563–578. young autism project. Infants Young Child indicators of autism spectrum disorders in the
Mullen E. 1995. Mullen Scales of Early Learning. 10:67–78. second year of life. J Autism Dev Disord
Circle Pines, MN: American Guidance. Sonnander K. 2000. Early identification of children 34:473– 493.
Noterdaeme M, Sitter S, Mildenberger K, et al. with developmental disabilities. Acta Paediatr Wong V, Hui L, Lee W, et al. 2004. A Modified
2000. Diagnostic assessment of communica- Suppl 89:17–23. Screening Tool for Autism (Checklist for Au-
tive and interactive behaviours in children Sparrow S, Balla D, Cichetti D. 1984. The Vine- tism in Toddlers [CHAT-23] for Chinese
with autism and receptive language disorder. land Adaptive Behavior Scales. Circles Pines, children. Pediatrics 114(2):e166 – e176.
Eur Child Adolesc Psychiatr 9:295–300. MN: American Guidance Services. Woods J, Wetherby A. 2003. Early identification
Osterling J, Dawson G. 1994. Early recognition of Stefanatos G, Kinsbourne M, Wasserstein J. 2002. and intervention for infants and toddlers who
children with autism: a study of first birthday Acquired epileptiform aphasia: a dimensional are at risk for autism spectrum disorder. Lan-
home videotapes. J Autism Dev Disord 24: view of Lansu-Kleffner syndrome and the re- guage Speech Hearing Services Schools 34:
247–257. lation to regressive autistic spectrum disoders. 180 –193.

MRDD RESEARCH REVIEWS ● SCREENING FOR AUTISM: THE M-CHAT AND OTHER MEASURES ● DUMONT-MATHIEU & FEIN 261
Yeargin-Allsopp M, Rice C, Karapurkar T, et with young children. Arch Pediatr Adolesc Zwaigenbaum L. 2001. Autistic spectrum disorders
al. 2003. Prevalence of autism in a US Med 152:255–262. in preschool children. Can Fam Physician 47:
metropolitan area. JAMA. 289(1):49 – Young R, Brewer N, Pattison C. 2003. Parental 2037–2042.
55. identification of early behavioural abnormal- Zwaigenbaum L, Bryson S, Rogers T, et al. 2005.
Young KT, Davis K, Schoen C, et al. 1998. Lis- ities in children with autistic disorder. Autism Behavioral manifestations of autism in the first
tening to parents. A national survey of parents 2003 7:125–143. year of life. Int J Dev Neurosci 23:143–152.

APPENDIX: M-CHAT*
Please fill out the following about how your child usually is. Please try to answer every question. If the behavior is rare (e.g., you
have 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 No
2. Does your child take an interest in other children? Yes No
3. Does your child like climbing on things, such as up stairs? Yes No
4. Does your child enjoy playing peek-a-boo/hide-and-seek? Yes No
5. Does your child ever pretend, for example, to talk on the phone or take care of dolls, or pretend other things? Yes No
6. Does your child ever use his/her index finger to point, to ask for something? Yes No
7. Does your child ever use his/her index finger to point, to indicate interest in something? Yes No
8. Can your child play properly with small toys (e.g., cars or bricks) without just mouthing, fiddling, or dropping them? Yes No
9. 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 No
11. Does your child ever seem oversensitive to noise? (e.g., plugging ears) Yes No
12. Does your child smile in response to your face or your smile? Yes No
13. Does your child imitate you? (e.g., you make a face—will your child imitate it?) Yes No
14. Does your child respond to his/her name when you call? Yes No
15. If you point at a toy across the room, does your child look at it? Yes No
16. Does your child walk? Yes No
17. Does your child look at things you are looking at? Yes No
18. Does your child make unusual finger movements near his/her face? Yes No
19. Does your child try to attract your attention to his/her own activity? Yes No
20. Have you ever wondered if your child is deaf? Yes No
21. Does your child understand what people say? Yes No
22. Does your child sometimes stare at nothing or wander with no purpose? Yes No
23. 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

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