1. HONOURS BACHELOR OF SCIENCE (2008)
MCMASTER UNIVERSITY
Hamilton, Ontario
TITLE: The Early Learning Measure as a predictor of outcome following Intensive Behavioural
Intervention in young children with severe autism
AUTHOR: Jacqueline Best
SUPERVISOR: Dr. Jo-Ann Reitzel
NUMBER OF PAGES: ix, 72
2. ABSTRACT
The Early Learning Measure as a predictor of outcome following Intensive Behavioural
Intervention in young children with severe autism
Jacqueline Best1
, Jo-Ann Reitzel1,2
, Jane Summers1,2
, Peter Szatmari1,2
, and Lonnie
Zwaigenbaum3
1
McMaster University Department, 2
Hamilton Health Sciences, 3
Capital Health, Alberta
Intensive Behavioural Intervention (IBI) for young children with autism is an extensively
researched intervention and has resulted in significant developmental gains in many (Lovaas,
1987; McEachin Smith, & Lovaas, 1993; Smith et al., 2000), but not all children (Sallows &
Graupner, 2005). To date, little attention has been paid to identifying predictors of good
outcome for young children with severe autism in a prospective rather than post-hoc manner.
Moreover, intensity of IBI and pre-treatment static measures such as intellectual functioning
(Smith et al., 2000) tend to be poor predictors of outcome and do not account for a child’s ability
to learn once in treatment. In Ontario, IBI is provided for children with severe autism, but
demand for treatment outweighs the limited resources, and many children remain on waitlists.
An understanding of the relationship between early learning characteristics and developmental
outcomes would be useful for maximizing available IBI resources and preventing long-term
exposure to ineffective treatment for children with autism if they are not benefiting from IBI.
The objective of this study is to determine if the Early Learning Measure (ELM) is a useful
predictor of 6- and 12-month outcomes on measures of cognitive functioning and adaptive
behaviour for children with severe autism following IBI, and if the ELM is a better predictor of
these outcomes than baseline characteristics, which include entry IQ, age at entry to IBI, and
autism severity.
3. A sample of 29 children with Autistic Disorder or PDD-NOS was administered the ELM
monthly for the first four months of IBI. The ELM is a measure of skill acquisition in four
domains: receptive instructions, expressive labels, non-verbal imitation, and verbal imitation.
Children were grouped into mastery or non-mastery groups, with mastery defined as a correct
response rate of 80% in all domains in the first 4 months of IBI. The Vineland Adaptive
Behaviour Scales (VABS) and a cognitive measure (either the Mullen Scales of Early Learning
or the Stanford-Binet Intelligence Scales-5th
Ed.) were used to assess adaptive behaviour and
cognitive functioning respectively at entry to IBI and at 6 and 12 months during IBI.
Results demonstrated that the ELM is a useful predictor of outcome following IBI.
Mastery of the ELM was correlated with significantly higher outcomes in adaptive behaviour
and cognitive functioning. The ELM domains that served as the best predictors of adaptive
behaviour outcomes were those involved in using or understanding language: Expressive Labels,
Verbal Imitation, and Receptive Instructions. Regression analyses completed to determine
whether mastery of the ELM served as a better predictor of outcome than baseline characteristics
demonstrated that, for some outcomes (12-month Vineland Adaptive Behaviour Composite, 12-
month Vineland Communication standard score), mastery of certain ELM domains or the total
ELM does predict outcome better than baseline characteristics. However, the ELM alone was
never the best predictor of outcome. The best predictor for each outcome was a combination of
baseline characteristics and mastery of the ELM or the Expressive Labels domain of the ELM.
This study demonstrates that while the ELM is a useful predictor of adaptive behaviour
and cognitive functioning outcomes, a better prediction of outcome can be obtained when
combining ELM mastery category with baseline characteristics. The study has also shown that
the ability to expressively label objects early in IBI may be an important skill that helps
4. distinguish children who are likely to benefit from IBI and those who are not. Determining
predictors of outcome in IBI has clinical implications and can help maximize IBI resources and
prevent exposure to IBI for children who do not stand to benefit from the treatment.