This document discusses research on early intervention for autism spectrum disorder (ASD). It finds that high-intensity, comprehensive early intervention models can significantly improve outcomes for young children with ASD. Interventions like the Early Start Denver Model (ESDM) have been shown to substantially increase IQ scores, language abilities, and adaptive behaviors in children with ASD. Lower intensity parent-implemented interventions have also demonstrated effectiveness but produce more modest gains and have less long-term data available. The document emphasizes the need to identify children with ASD as early as possible, provide accurate early diagnoses, and ensure access to evidence-based early intervention services.

1. How malleable is autism?
Outcome studies of the youngest
children with ASD
Sally J. Rogers
University of California Davis
MIND Institute

2. Acknowledgements
NICHD U19 HD 85468 (Rogers)
NIMH R01 MH 081757 (Rogers)
NIDCD DC R03 05574 (Rogers)
NIMH U54 MH66399 (Dawson)
OSERS #G008100247 (Rogers)
NICHD R21 065275 (Rogers)
Autism Speaks, NAAR, CAN
OAR, MIND (Vismara)
OSERS (Rogers)
MIND Institute
Coleman Institute
Children’s Miracle Network
John & Marcia Goldman
Foundation
Geraldine Dawson
Laurie Vismara
Diane Osaki
Annette Estes
Cathy Lord
Jeff Munson
Nick Lange
Giacomo Vivanti
Greg Young
Maria Rocha
Jamie Winter
Children and families

3. 3
Conflict of interest
royalties
honoraria

4. 4

5. MselELC
20
30
40
50
60
70
80
90
2 yr 9 yr
Lord et al., 2006
20
30
40
50
60
70
80
90
2-3 yr 10-12 yr
ASD Other ID
Sigman & Ruskin, 1999
IQ scores in ASD
across childhood
n 79
n 42
n 84
n 50

6. Life Magazine, May 7, 1965
“Screams, Slaps, and Love”

7. The game-changer
Lovaas,1987; McEachin et al 1993
40
50
60
70
80
90
32 mon age 7-8 13 yr
Exp Comp
IQ scores
40
50
60
70
80
90
Adaptive Behavior at 13

8. Meta-analysis of DTT studies
Reichow & Wohler, 2009
.69

9. Remington 2007 Lovaas Early Intensive
Behavioral Intervention
Wetherby et al, 2006, 2015 Early Social Interaction
Dawson, Rogers et al 2010 Early Start
Denver Model (ESDM)
High intensity, global
intervention models. How much
can impairment be reduced?
(Rogers & Vismara, 2013)

10. Wetherby ESI 2014: change scores
*
***
ESI-1:1
ESI-Group
**

11.
Early Start Denver Model
ASD specific
Comprehensive
Interdisciplinary
Integrates developmental and
learning science
Evidence-based teaching
Data-based

12. 2010 outcomes from RCT of ESDM
versus community care (Dawson et al., 2010)
MullenCompositeStandardScore
40
50
60
70
80
90
100
Baseline 1 year 2 years
Intervention
Community
p < .05
IQ Language
p < .05
(Dawson NIMH U54 MH66399)
p < .05
p < .05
15
22
29
36
43
50
Baseline 2 years
p < .05

13. Replication: ESDM in 1:4 groups
Vivanti, Dissanayake et al., 2014
40
50
60
70
80
90
IQ 1 IQ2
ESDM n=27
ASD n=30
40
50
60
70
80
90
Adapt 1 Adapt 2
*p=.05
30
40
50
60
70
80
90
Rec Lang1 Rec Lang2
*p=.05
*p=.05

14. +
COMM Tx
ESDM Tx
Gains last for years
(Estes et al., 2015)

15. 15
• Hard to isolate
effective elements
• Cost
• Complex to learn and
deliver
Strengths Weaknesses
• Significant change
• In multiple areas
• Unified approach
High intensity,
comprehensive
interventions

16. Low intensity models:
target a specific change
with a specific procedure
Kasari et al, 2015 JASPER: Joint attention,
engagement, symbolic play
Ingersoll et al, 2012, 2012 Reciprocal Imitation: RIT
Green, et al, 2015 VIPP: Video Interaction for
Producing Positive Parenting
Steiner et al, 2013 PRT: Pivotal Response Training
Baranek et al, 2015 Responsive Teaching +
social commun, sensory reg

17.
Effective with other groups
Allow for intensity, generalization
Implementable using distance technology
More than 5 with demonstrated efficacy
Parent
Implemented
Interventions

18. Process: How many
interventionists think about it
Professional ChildParent

19. How it really goes
Professional
Child
Parent
Other family and
Care providers

20. New:
Parent-ESDM
from a distance;
effects
on child
communication
Vismara et al,
in press

21. Parallel
Processes
Procedural & declarative
Hands on, guided learning
Goal focused,self-assessed
Much practice
Mastery

22. 22
Effective
Low cost
Low intensity
Brief
Parent learning
No long term data
Addresses a few needs
Not meant to stand
alone
Requires variety of plans
to address all needs
Strengths Weaknesses
Low intensity models

23. Biology
of autism
Child does not
adequately
engage in
social learning
Increasing
social deprivation
due to isolating
effects of autism
Lack of social
learning alters
course of
neural and
psychological
development
Model: Interactional Effects of Autism
Mundy 1995, Dawson et al.1998

24. Smiles per minGaze per min
Directed voc per min Quality soc engage
25
25

25. DQ scores of infant sibs (n=48) who develop
ASD and comparisons (n=92) (Ozonoff et al)MselELC
60
75
90
105
120
6m 9m 12m 18 m 24 m 36 m
ASD
TD

26. Can we begin sooner?
Rogers & Vismara, NICHD R21 HD065275; Autism Speaks
Decreased gaze, social interest
Little intentional communication
Little coordinated voice, gaze, gesture
Delayed phonemic development
Visual fixations on objects
Atypical repetitive behaviors

28. Treated Group/Refused Group

29. Five comparison groups: autism onset, declined
enrollment, high risk (sibs), low risk, treated

30. The most effective
interventions
30
Majority have skills in normal range by 6-7.
90% (ESDM) spontaneous, generative,
phrase or sentence speech, by 4.
How is this possible?

33. Phenotype is emerging

34. Phenotype is emerging
Neural readiness

35. Phenotype is emerging
Neural readiness
Positive emotion

36. Phenotype is emerging
Neural readiness
Positive emotion
Social attention - resonance

37. Phenotype is emerging
Neural readiness
Positive emotion
Social attention - resonance
Targeted learning, enrichment

38. Phenotype is emerging
Neural readiness
Positive emotion
Social attention - resonance
Targeted learning, enrichment
Self-righting

39. Phenotype is emerging
Neural readiness
Positive emotion
Social attention - resonance
Targeted learning, enrichment
Self-righting
Growth or compensation?

41. How do we move forward?
to find them?
to diagnose them?
to treat them?

42. 30–50% concerns by 12 months
Age of first concern = 18 – 19 months
Age of U.S. diagnosis 48 – 53 months (CDC, 2012)
Find them

43. 43
Find them
pediatric contacts
news articles
parent advocacy groups
autism treatment groups
child care educational groups

44. Age 3 Clinical best estimate outcomes
Of 294 infant sibs Ozonoff et al, 2014
Monitor sibs
Find them

45. Diagnose them
Use formal tools: screeners and tests
Good screeners
Infant-Toddler Checklist (ITC) 6 – 24 months
http://firstwords.fsu.edu/pdf/checklist.pdf
Modified Checklist for Autism in Toddlers 16 – 30
months https://www.m-chat.org
Good tests
Autism Observation Scale for Infants (AOSI)
ADOS-Toddler Module

46. Early diagnosis provisional, descriptive
Goal of early diagnosis is early treatment
Cannot predict outcomes from infant behavior
If parents did not want to know, they would not
be here
Infant mental health mindset

47. Act for earlier evidence-based services
Learn, provide a parent-implemented tx
Find those already seeing infants and network
Use your influence in public children’s services
We need to help families get more treatment
availability
Treat them

48. Prevent
Reverse
Improve