ARTICLES

Randomized, Controlled Trial of an Intervention for

Toddlers With Autism: The Early Start Denver Model
AUTHORS: Geraldine Dawson, PhD,a,b,c Sally Rogers, WHAT’S KNOWN ON THIS SUBJECT: Previous studies on the
PhD,d Jeffrey Munson, PhD,e,f Milani Smith, PhD,e Jamie efficacy of early behavioral intervention for improving outcomes
Winter, PhD,e Jessica Greenson, PhD,e Amy Donaldson, for preschool-aged children with autism have yielded promising
PhD,g and Jennifer Varley, MSe results. However, no randomized clinical trials of early
aAutism Speaks; bDepartment of Psychiatry, University of North
developmental behavioral intervention designed for toddlers with
Carolina, Chapel Hill, North Carolina; cDepartment of Psychology,
eUniversity of Washington Autism Center, Center on Human
autism have been conducted to date.
Development and Disability, and fDepartment of Psychiatry and
Behavioral Sciences, University of Washington, Seattle, WHAT THIS STUDY ADDS: This study assessed the efficacy of the
Washington; dMIND Institute, Department of Psychiatry, Early Start Denver Model, a comprehensive developmental
University of California Davis, Sacramento, California; and behavioral intervention, for improving outcomes of toddlers with
gDepartment of Speech and Hearing Science, Portland State
ASD. The intervention, which was initiated when children were
University, Portland, Oregon
less than 21⁄2 years, resulted in significant improvements in IQ,
KEY WORDS language, adaptive behavior, and autism diagnosis.
autism, behavioral intervention, cognitive function,
developmental outcomes, early intervention
ABBREVIATIONS
ASD—autism spectrum disorder
ESDM—Early Start Denver Model
ABA—applied behavior analysis
A/M—assess and monitor
abstract
PDD—pervasive developmental disorder OBJECTIVE: To conduct a randomized, controlled trial to evaluate the
NOS—not otherwise specified efficacy of the Early Start Denver Model (ESDM), a comprehensive de-
MSEL—Mullen Scales of Early Learning velopmental behavioral intervention, for improving outcomes of tod-
ADOS—Autism Diagnostic Observation Schedule
VABS—Vineland Adaptive Behavior Scales dlers diagnosed with autism spectrum disorder (ASD).
RBS—Repetitive Behavior Scale METHODS: Forty-eight children diagnosed with ASD between 18 and 30
This trial has been registered at www.clinicaltrials.gov months of age were randomly assigned to 1 of 2 groups: (1) ESDM inter-
(identifier NCT00090415).
vention, which is based on developmental and applied behavioral analytic
www.pediatrics.org/cgi/doi/10.1542/peds.2009-0958 principles and delivered by trained therapists and parents for 2 years; or
doi:10.1542/peds.2009-0958 (2) referral to community providers for intervention commonly available in
Accepted for publication Jul 17, 2009 the community.
Address correspondence to Geraldine Dawson, PhD, Autism RESULTS: Compared with children who received community-intervention,
Speaks, 4120 Bioinformatics Building, University of North
Carolina, Chapel Hill, North Carolina. E-mail: gdawson@ children who received ESDM showed significant improvements in IQ, adap-
autismspeaks.org tive behavior, and autism diagnosis. Two years after entering intervention,
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). the ESDM group on average improved 17.6 standard score points (1 SD: 15
Copyright © 2009 by the American Academy of Pediatrics points) compared with 7.0 points in the comparison group relative to base-
FINANCIAL DISCLOSURE: Sally Rogers and Geraldine Dawson line scores. The ESDM group maintained its rate of growth in adaptive behav-
are authors of Early Start Denver Model for Young Children with ior compared with a normative sample of typically developing children. In con-
Autism from which they receive royalties.
trast, over the 2-year span, the comparison group showed greater delays in
adaptive behavior. Children who received ESDM also were more likely to expe-
rience a change in diagnosis from autism to pervasive developmental disor-
der, not otherwise specified, than the comparison group.
CONCLUSIONS: This is the first randomized, controlled trial to demon-
strate the efficacy of a comprehensive developmental behavioral inter-
vention for toddlers with ASD for improving cognitive and adaptive
behavior and reducing severity of ASD diagnosis. Results of this study
underscore the importance of early detection of and intervention in
autism. Pediatrics 2010;125:e17–e23

PEDIATRICS Volume 125, Number 1, January 2010 e17

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
Autism spectrum disorder (ASD) is ways. First, we maintained a high level vices the parents chose; and (2) the
characterized by impairments in so- of methodologic rigor, including gold- assess-and-monitor (A/M) group re-
cial reciprocity and communication, standard diagnostic criteria, random- ceived yearly assessments with inter-
and stereotyped and repetitive behav- ization, comprehensive outcome mea- vention recommendations and refer-
iors, with onset during early child- sures conducted by naive examiners, rals for intervention from commonly
hood. Intellectual disability is present high retention rates, and measures of available community providers in the
in a large proportion of individuals.1 fidelity of implementation of a manual- greater Seattle region.
With a prevalence of 1 per 150,2 autism ized intervention. Children were evaluated by experi-
costs the United States $35 billion per Second, to our knowledge, our study is enced examiners naive to intervention
year.3 The lifetime per-capita societal the first randomized, controlled trial of status at baseline (preintervention), 1
cost of autism is $3.2 million, with lost intervention for toddlers with autism; year after onset of the intervention,
productivity and adult care among the all children were younger than 30 and at either 2 years after onset of the
largest costs.4 Thus, early-intervention months at entry. Given the recent rec- intervention or at 48 months of age,
methods that can improve outcome ommendation by the American Acad- whichever yielded a longer time frame.
for individuals with ASD are of high emy of Pediatrics that 18-month-old
importance. children be screened for ASD,9 it is im- Participants
The 1987 report by Lovaas5 of an early perative that the efficacy of early- Participants were recruited through
behavioral intervention that resulted intervention models appropriate for
pediatric practices, Birth to Three cen-
in 49% of children in the study being toddlers with ASD be demonstrated.
ters, preschools, hospitals, and state
mainstreamed into regular class- Third, the intervention, the Early Start and local autism organizations. Exclu-
rooms and showing significant IQ Denver Model (ESDM),10 is a comprehen- sion criteria included (1) a neurodevel-
gains created a groundswell of inter- sive early behavioral intervention for in- opmental disorder of known etiology
est among parents and professionals fants to preschool-aged children with (eg, fragile X syndrome), (2) significant
in early intervention and raised ques- ASD that integrates applied behavior sensory or motor impairment, (3) ma-
tions about early plasticity in children analysis (ABA) with developmental and jor physical problems such as a
with autism.6 Although subsequent in- relationship-based approaches. The chronic serious health condition, (4)
tervention studies, including a ran- ESDM was designed to address the seizures at time of entry, (5) use of psy-
domized, controlled trial,7 have docu- needs of toddlers with ASD as young as choactive medications, (6) history of a
mented improvements in IQ for a 12 months. The intervention is provided serious head injury and/or neurologic
subgroup of children, questions re- in a toddler’s natural environment (the disease, (7) alcohol or drug exposure
garding the efficacy of early interven- home) and is delivered by trained thera- during the prenatal period, and (8) ra-
tion have remained. Many of the stud- pists and parents. In our study, children tio IQ below 35 as measured by mean
ies lacked methodologic rigor. Authors received structured intervention at high age equivalence score/chronological
of a recent meta-analysis of the effi- intensity, consistent with the National age on the visual reception and fine
cacy of early behavioral intervention Research Council’s recommendation.11 motor subscales of the Mullen Scales
argued that stronger evidence that of Early Learning (MSEL).12 Children
early behavioral intervention results PATIENTS AND METHODS
who developed seizures during the
in better outcomes than standard care Study Procedures course of the study were not excluded.
is still needed.8 Forty-eight children between 18 and 30 Inclusion criteria included age below
Our study was a randomized, con- months of age diagnosed with autistic 30 months at entry, meeting criteria
trolled trial of early intensive behav- disorder or pervasive developmental for autistic disorder on the Toddler Au-
ioral intervention for young children disorder (PDD), not otherwise speci- tism Diagnostic Interview,13 meeting
with ASD that was funded by the Na- fied (NOS), were randomly assigned to criteria for autism or ASD on the Au-
tional Institute of Mental Health (Dr 1 of 2 groups: (1) the ESDM group re- tism Diagnostic Observation Sched-
Dawson, principal investigator). It was ceived yearly assessments, 20 hours/ ule,14 and a clinical diagnosis based on
hypothesized that the early interven- week of the ESDM intervention from Diagnostic and Statistical Manual of
tion would result in significant im- University of Washington clinicians, Mental Disorders, Fourth Edition
provements in cognitive abilities of parent training, and parent delivery (DSM-IV) criteria15 using all available
young children with ASD. The study dif- for 5 or more hours/week of ESDM, in information, residing within 30 min-
fered from previous ones in several addition to whatever community ser- utes of the University of Washington,

e18 DAWSON et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
 ARTICLES

and willingness to participate in a ⱖ2- Assessed for eligibility

year intervention. At baseline, 18 chil- (n = 96)

dren in the A/M group and 21 in the Excluded (n = 45)

Enrollment
ESDM group received a DSM-IV diagno- Did not meet diagnostic
criteria (n = 38)
sis of autistic disorder. Six children in Met diagnostic criteria but
the A/M group and 3 in the ESDM group declined to participate in
randomization because of
received a diagnosis of PDD NOS. This demands of intervention
difference was not significant (Fisher’s (n = 7)

exact test, P ⫽ .231). The ethnicities Randomly assigned (n = 51)

Allocation
involved were Asian (12.5%), white
(72.9%), Latino (12.5%), and multira-
Excluded (n = 3) Allocated to ESDM Allocated to A/M
cial (14.6%). The male-to-female ratio Declined ESDM because of group (n = 24) group (n = 24)
reflected the expected ratio in ASD of intervention requirements
(n = 2)
3.5:1. Subsequently diagnosed
with Rett syndrome (n = 1) Time 1 assessment Time 1 assessment
Retention rates were 100% (1-year) (n = 24) (n = 23)
Followed (n = 24) Followed (n = 23)
and 100% (2-year) for the ESDM group Lost to follow-up Lost to follow-up
(n = 0) (n = 1)
Follow-up

and 96% (1-year) and 88% (2-year) for

the A/M group, which yielded a sample
size of 24 in the ESDM and 21 in the A/M Time 2 assessment Time 2 assessment
group at outcome. Figure 1 shows the (n = 24) (n = 21)
Followed (n = 24) Followed (n = 21)
participant flowchart. Lost to follow-up Lost to follow-up
(n = 0) (n = 3)

Measures
Analysis

Analyzed (n = 24) Analyzed (n = 21)

Autism Diagnostic Interview–Revised
The toddler version of the Autism Diag-
nostic Interview–Revised13 is a semi- FIGURE 1
Participant flowchart.
structured parent interview that as-
sesses autism symptoms across 3
domains: social relatedness; commu-
were administered: fine motor, visual re- sameness, self-injurious behavior)
nication; and repetitive, restricted
ception, expressive language, and recep- and a total score.
behaviors.
tive language. T scores for subscales
have a mean of 50 (SD: 10). The early- Randomization
Autism Diagnostic Observation
Schedule learning composite score is a standard Participants were stratified into 2
score with mean of 100 (SD: 15). groups on the basis of composite IQ at
The Autism Diagnostic Observation
entry (⬍55 and 55) and gender to en-
Schedule (ADOS), WPS version,14 is a Vineland Adaptive Behavior Scales
sure comparable IQ and gender ratios
semistructured standardized observa- The Vineland Adaptive Behavior Scales between groups. Within each of these
tion that measures autism symptoms (VABS)17 are a parent interview that as- strata, randomization was conducted
in social relatedness, communication, sesses social, communication, motor, by using random permuted blocks of 4.
play, and repetitive behaviors. A stan- and daily living skills. They provide age- The intervention groups did not differ
dardized severity score based on equivalent and standard scores for at baseline in severity of autism symp-
codes within these domains can be cal- several subscales, including expres- toms based on ADOS scores, chrono-
culated to compare autism symptoms sive and receptive language and social logical age, IQ, gender, or adaptive be-
across modules.16 adaptive functioning. haviors (see Table 1).
Mullen Scales of Early Learning Repetitive Behavior Scale
Intervention Groups
(MSEL) The Repetitive Behavior Scale (RBS)18
The MSEL12 are a standardized develop- is a parent questionnaire that charac- ESDM Group
mental test for children from birth to 68 terizes the severity of repetitive behav- The ESDM group was provided with in-
months of age. Four of the 5 subscales iors, yielding 6 subdomain scores (eg, tervention by trained therapists for

PEDIATRICS Volume 125, Number 1, January 2010 e19

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
TABLE 1 Baseline Measures for ESDM and A/M Groups therapist-delivered intervention hours
A/M Group ESDM Group F MS P per week were made available, the ac-
Mean SD N Mean SD N tual mean was 15.2 hours (SD: 1.4) be-
Age at study entry, mo 23.1 3.9 24 23.9 4.0 24 0.48 7.52 .490 cause of illnesses, vacations, and so
MSEL on. Parents reported spending an av-
Early-learning compositea 59.4 8.6 24 61.0 9.2 24 0.40 31.69 .530
erage of 16.3 hours/week (SD: 6.2) us-
Receptive languageb 21.2 3.8 24 21.1 4.7 24 0.01 0.19 .920
Expressive languageb 26.0 8.6 24 24.5 7.2 24 0.48 30.08 .492 ing ESDM strategies. The ESDM group
Visual receptionb 30.8 8.9 24 33.2 11.0 24 0.70 70.08 .406 reported an average of 5.2 hours/week
Fine motorb 30.6 10.7 24 33.9 11.9 24 1.02 130.02 .318 (SD: 2.1) in other therapies (eg, speech
VABS
Adaptive behavior compositea 69.9 7.3 24 69.5 5.7 24 0.04 1.69 .844 therapy, developmental preschool)
Communicationa 69.6 7.3 24 68.4 7.6 24 0.32 17.52 .577 over the study enrollment period.
Socializationa 72.4 9.4 24 73.8 7.7 24 0.29 21.33 .594 Other therapies were documented by
Daily living skillsa 86.8 10.0 24 87.3 11.4 24 0.03 3.52 .381
Motor skillsa 72.5 6.5 24 70.9 6.2 24 0.78 31.69 .862
using an intervention history interview
ADOS severity score 6.9 1.7 24 7.2 1.7 24 0.35 1.02 .557 administered every 6 months.
RBS total 21.5 19.2 24 15.2 10.8 24 1.93 468.75 .171
No significant differences among baseline measures were found (P ⬎ .10 on all measures). A/M Group
a Standard score (mean: 100 关SD: 15兴).

b T score (mean: 50 关SD: 10兴). Children who were randomly assigned

to the A/M group received comprehen-
sive diagnostic evaluations, interven-
2-hour sessions, twice per day, 5 days/ objectives from the curriculum that tion recommendations, and commu-
week, for 2 years. A detailed interven- they viewed as high priority. nity referrals at baseline and again at
tion manual and curriculum were Intervention programs were super- each of the 2 follow-up assessments.
used.19 One or both parents were pro- vised by a graduate-level, trained lead Families were given resource manuals
vided with parent training from the pri- therapist who had a minimum of 5 and reading materials at baseline and
mary therapist during semimonthly years’ experience providing early in- twice yearly throughout the study. The
meetings, during which the principles tervention to young children with au- A/M group reported an average of 9.1
and specific techniques of ESDM were tism, with ongoing consultation from a hours of individual therapy and an
taught. Parents were asked to use average of 9.3 hours/week of group in-
clinical psychologist, speech-language
ESDM strategies during daily activities terventions (eg, developmental pre-
pathologist, and developmental behav-
and to keep track of the number of school) across the 2-year period dur-
ioral pediatrician. An occupational
hours during which they used these ing which the intervention study was
therapist provided consultation as
strategies. ESDM uses teaching strate- conducted. In the greater Seattle area,
needed. Intervention objectives and
gies that involve interpersonal ex- there are a number of Birth to Three
strategies were reviewed, and the in-
change and positive affect, shared en- centers that provide interventions,
tervention was observed at least bi-
gagement with real-life materials and speech and language therapy, and
weekly by the lead therapist and every
activities, adult responsivity and sensi- occupational therapy. Developmen-
3 months by the speech-language pa-
tivity to child cues, and focus on verbal tal preschool programs vary but typ-
thologist. Intervention was delivered
and nonverbal communication, based ically include special education and
by therapists who typically held a bac- related services. There are a number
on a developmentally informed curric-
calaureate degree, received 2 months of private ABA providers in the
ulum that addresses all developmental
of training by the lead therapist, and community.
domains. Teaching strategies are con-
sistent with the principles of ABA, such met weekly with the lead therapist.
as the use of operant conditioning, Therapists were trained to compe- Data Analysis
shaping, and chaining. Each child’s tence, defined as completing course- The effect of ESDM intervention was as-
plan is individualized. There is a strong work, passing tests, mastering the in- sessed by using repeated-measures
parent-family role responsive to each tervention, demonstrating fidelity of analysis of variance, with a priori con-
family’s unique characteristics. Par- 85% of maximum scores on the fidelity trasts that compared baseline scores
ents are taught the basic ESDM strate- instrument, and maintaining ongoing with 1- and 2-year outcome scores. The
gies and asked to use them during ev- fidelity.20 primary outcome measures were the
eryday activities such as feeding, bath ESDM intervention hours were sys- MSEL composite standard score and
time, and play. Parents chose teaching tematically recorded. Although 20 the VABS composite standard score.

e20 DAWSON et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
 ARTICLES

TABLE 2 Child Outcomes After 1 and 2 Years of Study Participation

1-y Outcome 2-y Outcome Group ⫻ Time Group ⫻ Time
(Baseline vs (Baseline vs 2-y)
A/M (N ⫽ 23) ESDM (N ⫽ 24) A/M (N ⫽ 21) ESDM (N ⫽ 24) 1-y)
Mean SD ⌬ Mean SD ⌬ Mean SD ⌬ Mean SD ⌬ F MS P F MS P
Chronological age, mo 38.1 3.8 15.3 38.8 4.4 14.9 52.1 4.3 29.3 52.4 3.4 28.5 0.95 2.18 .334 1.27 6.91 .266
MSEL
Early-learning 64.0 13.8 4.4 76.4 23.4 15.4 66.3 15.3 7.0 78.6 24.2 17.6 5.99 1416.90 .018 4.31 1264.38 .044
composite
Receptive language 31.1 11.1 9.8 38.9 15.4 17.8 31.5 10.6 10.2 40.0 16.3 18.9 4.00 745.21 .051 4.14 843.56 .048
Expressive language 33.0 11.5 6.7 36.1 14.2 11.6 30.0 9.2 4.0 36.6 13.6 12.1 1.99 290.43 .165 4.88 748.07 .033
Visual reception 29.0 10.7 ⫺1.7 38.8 16.4 5.6 34.5 13.0 4.5 41.0 17.9 7.8 4.22 621.97 .046 0.63 126.23 .433
Fine motor 26.1 8.6 ⫺5.0 32.7 11.7 ⫺1.3 28.5 9.5 ⫺2.8 33.5 12.2 ⫺0.4 1.32 161.35 .256 0.46 63.81 .503
VABS 63.7 8.8 ⫺6.3 65.7 9.8 ⫺3.8 59.1 8.8 ⫺11.2 68.7 15.9 ⫺0.8 0.85 71.61 .360 7.05 1181.82 .011
Communication 71.0 13.0 1.2 73.5 11.7 5.0 69.4 15.8 ⫺0.7 82.1 21.8 13.7 1.38 175.69 .246 6.38 2300.98 .015
Socialization 68.9 12.1 ⫺3.5 70.0 9.9 ⫺3.8 63.1 9.3 ⫺8.9 69.2 11.6 ⫺4.6 0.01 1.15 .934 1.29 204.57 .263
Daily living skills 65.3 7.1 ⫺7.4 65.6 8.6 ⫺5.3 58.0 8.1 ⫺14.5 64.7 12.4 ⫺6.2 0.89 51.78 .350 6.73 773.34 .013
Motor skills 70.7 12.2 ⫺15.9 75.1 14.4 ⫺12.2 64.1 12.3 ⫺23.1 77.4 19.8 ⫺9.9 0.99 157.43 .326 7.40 1881.65 .009
ADOS severity score 7.3 2.1 0.4 6.5 1.5 ⫺0.7 7.3 1.8 0.3 7.0 1.9 ⫺0.2 3.38 13.15 .072 0.66 3.29 .422
RBS total 23.3 17.5 1.0 15.5 12.3 0.9 22.0 16.3 ⫺0.6 16.7 13.1 2.5 0.001 0.19 .976 0.37 92.50 .545
⌬ indicates mean change from baseline.

Secondary outcome measures were improvement in the A/M group, a dif- the ESDM group, whereas the A/M
the ADOS severity score,16 the RBS,18 ference that fell just short of statis- group improved 10.2 and 4.0 points,
MSEL, and VABS subscale scores, and tical significance. respectively.
changes in diagnostic status (autistic As a whole, children gained raw The ESDM and A/M groups signifi-
disorder, PDD NOS, and no diagnosis). score points in the daily living skills cantly differed in terms of their
RESULTS subscale of the VABS; however, adaptive behavior as measured by
progress was much slower in rela- the VABS composite standard scores
No serious adverse effects related to tion to the VABS normative sample at the 2-year outcome (see Fig 2). The
the intervention were reported during between baseline and 1 year. The ESDM group showed similar stan-
the 2-year period.
groups did not differ in terms of dard scores at the 1- and 2-year out-
1-Year Outcome adaptive behavior, measured by the comes, indicating a steady rate of de-
VABS composite standard score, af- velopment, whereas the A/M group,
Table 2 displays statistics for 1- and
ter 1 year (the ESDM group showed a on average, showed an 11.2-point av-
2-year outcomes, change scores rela-
3.8-point decline, and the A/M group erage decline. Thus, the A/M group’s
tive to baseline, and group compari-
showed a 6.3-point decline). The delays in overall adaptive behavior
sons for primary and secondary mea-
groups did not differ in terms of their became greater when compared
sures. Significant intervention effects
ADOS severity scores or RBS total with the normative sample. The A/M
were found for cognitive ability after 1
score after 1 year of intervention. group showed average declines in
year on the MSEL composite standard
scores. The ESDM group demonstrated standard scores that were twice as
2-Year Outcome great as those in the ESDM group in
an average IQ increase of 15.4 points
(⬎1 SD) compared with an increase of Two years after the baseline assess- the domains of socialization, daily liv-
4.4 points in the A/M group. The visual ment, the ESDM group showed signif- ing skills, and motor skills. The
reception subscale was the only indi- icantly improved cognitive ability, groups did not differ in terms of their
vidual subscale on the MSEL on measured by MSEL composite stan- ADOS severity scores or RBS total
which the groups significantly dif- dard scores, which increased 17.6 score after 2 years of intervention.
fered at the 1-year outcome. The points compared with 7.0 points in
ESDM group gained 5.6 T-score the A/M group. The bulk of this Diagnosis
points, whereas the A/M group de- change seems to have been a result At baseline, the diagnoses in each
clined 1.7 points. The ESDM group im- of receptive and expressive lan- group were not significantly differ-
proved 17.8 points on receptive lan- guage, which showed increases of ent (Fisher’s exact test, P ⫽ .461)
guage compared with a 9.8-point 18.9 and 12.1 points, respectively, for and were distributed as follows:

PEDIATRICS Volume 125, Number 1, January 2010 e21

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
 sample, whereas the community-
based intervention group continued
to fall farther behind in adaptive be-
havior. Given the importance of
adaptive behavior for everyday func-
tioning at home and school, the fact
that the ESDM group did not continue
to fall farther behind is likely to af-
fect ability to function in less-
restrictive environments. This dem-
onstrates that the ESDM intervention
accelerates overall development and
is generalizing to everyday life.
FIGURE 2 Whereas 71% of the children in the
Mean scores on the MSEL (left) and the VABS composite (right) for children in the ESDM and A/M group that received community-
groups 1 and 2 years after entering study. Error bars indicate ⫾1 SD.
based intervention retained their di-
agnosis of autistic disorder over the
2-year period, only 56% of children in
ESDM, 21 with autistic disorder and 3 DISCUSSION the ESDM group did so. The diagnosis
with PDD NOS; A/M, 18 with autistic Recommendations by the American of 7 children (30%) in the ESDM
disorder and 6 with PDD NOS. At the Academy of Pediatrics9 that all chil- group changed from autistic dis-
2-year outcome, 15 (62.5%) children dren be screened for autism at 18 order to PDD NOS, whereas this
in the ESDM group had the same di- months of age oblige the develop- only occurred for 1 child (5%) in
agnosis (14 with autistic disorder, 1 ment of interventions that are appro- the community-intervention group.
with PDD NOS) and 15 (71.4%) chil- priate for toddlers with ASD. To our These diagnostic assessments were
dren in the A/M group had the same knowledge, this study is the first to conducted by experienced clinicians
diagnosis (all 15 with autistic disor- demonstrate the efficacy of an inten- who were naive with respect to
der). Diagnosis improved (baseline sive intervention designed for tod- intervention-group status. However,
autistic disorder to PDD NOS at year dlers with ASD as young as 12 this change in diagnostic severity
2) for 7 (29.2%) children in the ESDM was not reflected in significant dif-
months of age. After 2 years of inter-
group but for only 1 (4.8%) child in ferences in the ADOS severity scores.
vention, children provided with the
the A/M group. However, the diagno- This lack of correspondence be-
ESDM19 showed significant improve-
sis changed from PDD NOS at base- tween measures is difficult to inter-
ments in IQ, adaptive behavior, and
line to autistic disorder at year 2 for pret, because the child’s perfor-
diagnostic status compared with
mance in the ADOS contributes to
2 (8.3%) children in the ESDM group children who received community in-
clinical diagnosis. However, other
and 5 (23.8%) children in the A/M terventions. Consistent evidence of
behaviors, including parental report,
group. Thus, children who received improvement in communicative abil- also contribute to overall clinical
ESDM were significantly more likely ities in the ESDM group was found, as diagnosis. The repetitive-behavior
to have improved diagnostic status demonstrated by gains in receptive scores also did not change over time
at the 2-year outcome compared with and expressive language scores on in either group.
children in the A/M group, as as- the MSEL subscales and the VABS
sessed by using Fisher’s exact test 2 communication subscale. Significant CONCLUSIONS
(intervention groups) ⫻ 2 (improved improvement for the ESDM group The outcomes of this study, which in-
versus worsened diagnosis) contin- was found for overall adaptive be- volve an increase in IQ scores of 17
gency table (P ⫽ .041). Fisher’s exact havior, communication, daily living points (⬎1 SD) and significant gains in
test for the 2 (intervention groups) skills, and motor skills. Specifically, language and adaptive behavior, com-
⫻ 3 (diagnostic change: no change the ESDM group, although still signif- pare favorably with other controlled
versus improved versus worsened icantly delayed in adaptive behavior, studies of intensive early intervention
diagnosis) contingency table was was able to keep pace with the rate (eg, Smith et al [2000],7 which deliv-
just short of significance (P ⫽ .060). of change of the VABS normative ered discrete trial intervention for ⬎2

e22 DAWSON et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
 ARTICLES

years for 25– 40 hours/week). The The results of this study suggest that ACKNOWLEDGMENTS
group differences reported from our the ESDM model,19 an intervention ap- This study was supported by National
study are larger than those produced proach that uses teaching strategies Institute of Mental Health grant
by other comparative trials of develop- of ABA that are delivered within an af- U54MH066399 (to Dr Dawson).
mental behavioral approaches, which fectively rich, relationship-focused We acknowledge the contributions of
were conducted for briefer periods of context, can be effective for improving the parents and children who took
time and with fewer hours of delivery outcomes of young children with au- part in this study and the support and
per week.21,22 Whether the children will tism. Parents’ use of these strategies effort of numerous undergraduate and
sustain their gains over a longer term at home during their daily activities graduate students and staff who are
is an important question that will re- likely was an important ingredient of part of the University of Washington
quire follow-up study. its success. Autism Center.
REFERENCES
1. Baird G, Simonoff E, Pickles A, et al. Preva- guage, and adaptive behavior: a systematic ders. 4th ed. Washington, DC: American Psy-
lence of disorders of the autism spectrum review and meta-analysis. J Pediatr. 2008; chiatric Association; 1994
in a population cohort of children in South In press 16. Gotham K, Pickles A, Lord C. Standardizing
Thames: the Special Needs and Autism 9. Johnson CP, Myers SM; American Academy ADOS scores for a measure of severity in
Project (SNAP). Lancet. 2006;368(9531): of Pediatrics, Council on Children With Dis- autism spectrum disorders. J Autism Dev
210 –215 abilities. Identification and evaluation of Disord. 2009;39(5):693–705
2. Yeargin-Allsopp M, Rice C, Karapurkar T, Do- children with autism spectrum disorders. 17. Sparrow S, Balla D, Cicchetti D. Vineland
emberg N, Coyle C, Murphy C. Prevalence of Pediatrics. 2007;120(5):1183–1215
Adaptive Behavior Scales: Interview Edition.
autism in a U.S. metropolitan area. JAMA. 10. Smith M, Rogers S, Dawson G. The Early Circle Pines, MN: American Guidance
2003;289(1):49 –55 Start Denver Model: a comprehensive early Service; 1984
3. Ganz ML. The costs of autism. In: Molden SO, intervention approach for toddlers with au-
18. Bodfish JW, Symons FJ, Lewis MH. The Re-
Rubenstein JLR, eds. Understanding tism. In: Handleman JS, Harris SL, eds. Pre-
petitive Behavior Scale: A Test Manual. Mor-
Autism: From Basic Neuroscience to Treat- school Education Programs for Children
ganton, NC: Western Carolina Center Re-
ment. Boca Raton, FL: Taylor and Francis With Autism. 3rd ed. Austin, TX: Pro-Ed Cor-
Group; 2006 poration, Inc; 2008:65–101 search Reports; 1998
4. Ganz ML. The lifetime distribution of the in- 11. National Research Council, Division of Be- 19. Rogers SJ, Dawson G. Early Start Denver
cremental societal costs of autism. Arch Pe- havioral and Social Sciences Education. Ed- Model for Young Children with Autism. New
diatr Adolesc Med. 2007;161(4):343–349 ucating Children With Autism. Washington, York, NY: Guilford Press; 2009
5. Lovaas OI. Behavioral treatment and normal DC: National Academy Press; 2001 20. Bond GR, Evans L, Salyers MP, Williams J,
educational and intellectual functioning in 12. Mullen EM. Mullen Scales of Early Learning: Kim HW. Measurement of fidelity in psychi-
young autistic children. J Consult Clin Psy- AGS Edition. Circle Pines, MN: American atric rehabilitation. Ment Health Serv Res.
chol. 1987;55(1):3–9 Guidance Service; 1995 2000;2(2):75– 87
6. Dawson G. Early behavioral intervention, 13. Lord C, Rutter M, Le Couteur A. Autism Diag- 21. Jocelyn LJ, Casiro OG, Beattie D, Bow J,
brain plasticity, and the prevention of au- nostic Interview-Revised: a revised version Kneisz J. Treatment of children with autism:
tism spectrum disorder. Dev Psychopathol. of a diagnostic interview for caregivers of a randomized controlled trial to evaluate a
2008;20(3):775– 803 individuals with possible pervasive develop- caregiver-based intervention program in
7. Smith T, Groen AD, Wynn JW. Randomized mental disorders. J Autism Dev Disord. community day-care centers. J Dev Behav
trial of intensive early intervention for chil- 1994;24(5):659 – 685 Pediatr. 1998;19(5):326 –334
dren with pervasive developmental disor- 14. Lord C, Rutter M, DiLavore PC, Risi S. Autism 22. Aldred C, Green J, Adams C. A new social
der. Am J Ment Retard. 2000;105(4): Diagnostic Observation Schedule. Los Ange- communication intervention for children
269 –285 les, CA: Western Psychological Services; with autism: pilot randomised controlled
8. Spreckley M, Boyd R. Efficacy of applied be- 1999 treatment study suggesting effectiveness.
havioral intervention in preschool children 15. American Psychiatric Association. Diagnos- J Child Psychol Psychiatry. 2004;45(8):
with autism for improving cognitive, lan- tic and Statistical Manual of Mental Disor- 1420 –1430

PEDIATRICS Volume 125, Number 1, January 2010 e23

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019
Randomized, Controlled Trial of an Intervention for Toddlers With Autism: The
Early Start Denver Model
Geraldine Dawson, Sally Rogers, Jeffrey Munson, Milani Smith, Jamie Winter,
Jessica Greenson, Amy Donaldson and Jennifer Varley
Pediatrics 2010;125;e17
DOI: 10.1542/peds.2009-0958 originally published online November 30, 2009;

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/125/1/e17
References This article cites 12 articles, 1 of which you can access for free at:
http://pediatrics.aappublications.org/content/125/1/e17#BIBL
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Developmental/Behavioral Pediatrics
http://www.aappublications.org/cgi/collection/development:behavior
al_issues_sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019

Randomized, Controlled Trial of an Intervention for Toddlers With Autism: The
Early Start Denver Model
Geraldine Dawson, Sally Rogers, Jeffrey Munson, Milani Smith, Jamie Winter,
Jessica Greenson, Amy Donaldson and Jennifer Varley
Pediatrics 2010;125;e17
DOI: 10.1542/peds.2009-0958 originally published online November 30, 2009;

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/125/1/e17

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it
has been published continuously since 1948. Pediatrics is owned, published, and trademarked by
the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois,
60007. Copyright © 2010 by the American Academy of Pediatrics. All rights reserved. Print ISSN:
1073-0397.

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on October 13, 2019