Content uploaded by Trevor R Parmenter
Author content
All content in this area was uploaded by Trevor R Parmenter on Jan 15, 2019
Content may be subject to copyright.
A randomised controlled trial of two early intervention programs
for young children with autism: Centre-based with parent program
and home-based
§
Jacqueline Roberts
a,
*, Katrina Williams
b,1
, Mark Carter
c
, David Evans
d
, Trevor Parmenter
d
,
Natalie Silove
e
, Trevor Clark
f
, Anthony Warren
f
a
University of Sydney, Australia
b
University of New South Wales, Australia
c
Macquarie University, North Ryde, NSW 2113, Australia
d
University of Sydney, Faculty of Education and Social Work, Camperdown, NSW 2050, Australia
e
Child Development Unit, Children‘s Hospital, Westmead, NSW 2145, Australia
f
Autism Spectrum Australia, Aspect, French‘s Forest, NSW 2086, Australia
Autism is a severe developmental disability characterised by severe impairments in language and social development,
repetitive behaviours and restricted interests. Between 2003 and 2005 the identification of autism in Australia has increased
from 46 in 10,000 children to 63 in 10,000 children (Williams et al., 2005) with many of these children being identified in the
preschool year. At the same time as the increase in diagnosis of autism, early intervention has been widely recommended as
Research in Autism Spectrum Disorders 5 (2011) 1553–1566
ARTICLE INFO
Article history:
Received 22 February 2011
Received in revised form 1 March 2011
Accepted 1 March 2011
Available online 12 April 2011
Keywords:
Autism spectrum disorders
Early-intervention
Home-based
Centre-based
Parent programs
ABSTRACT
This study compares outcomes of early intervention programs for young children with
autism; an individualised home-based program (HB), a small group centre-based program
for children combined with a parent training and support group (CB) and a non-treatment
comparison group (WL). Outcome measures of interest include social and communication
skill development in children, quality of life and stress for parents. Eligible (diagnosed ASD,
preschool age) participants were randomised into 2 groups HB (n= 28) and CB (n= 28). A
WL (n= 29) comparison group was also recruited. HB and CB groups had 12-month
programs. Pre and post assessments were conducted using formal and informal measures.
Children in the CB group improved significantly more than HB and WL groups on some
social and communication measures. On parent measures outcomes varied with CB group
parents making the most gains in perception of competence and quality of life. The small
group centre-based program combined with parent pro gram resulted in the best most cost
effective outcomes for children and families however this is not an option for all. The CB
program did not suit some children and families. The heterogeneity of the autism
population indicates that a range of intervention options is necessary to meet the needs of
children with autism and their families.
Crown Copyright ß2011 Published by Elsevier Ltd. All rights reserved.
§
This research was funded by an Australian Research Council Linkage Projects grant (No. LP0562663) in conjunction with Autism Spectrum Australia
(Aspect). The Australian Research Council and Autism Spectrum Australia had no other involvement in this research.
* Corresponding author at: Present address: Child Development Unit, Children‘s Hospital Westmead, Sydney, NSW 2145, Australia. Tel.: +61 2 96870377;
fax: +61 2 98452088.
E-mail addresses: jacquir@chw.edu.au,jacqueliner@sydney.edu.au (J. Roberts), Katrina.Williams@rch.org.au (K. Williams), mark.carter.mq@gmail.com
(M. Carter), david.evans@sydney.edu.au (D. Evans), Trevor.parmenter@sydney.edu.au (T. Parmenter), NatalieS@chw.edu.au (N. Silove),
tclark@autismspectrum.org.au (T. Clark), awarren@autismspectrum.org.au (A. Warren).
1
Present address: University of Melbourne, Melbourne, Victoria 3000, Australia.
Contents lists available at ScienceDirect
Research in Autism Spectrum Disorders
Journal homepage: http://ees.elsevier.com/RASD/default.asp
1750-9467/$ – see front matter. Crown Copyright ß2011 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.rasd.2011.03.001
early as possible for children with developmental disabilities including autism (Carpenter & Egerton, 2005) at levels of
sufficient intensity (Dawson & Osterling, 1997; NRC, 2001). The involvement of parents as providers of intervention is also
increasingly highlighted with commensurate interest in programs for parents to facilitate their capacity to provide
intervention (McConachie & Diggle, 2007).
There is considerable interest in and debate about the relative merits of different types of intervention, including parent
mediated interventions, in terms of outcomes for children and, to a lesser extent, outcomes for families. There is general
agreement that there is insufficient research evidence to support one type of intervention over others and that, due to
individual variation among children with autism and families, there is unlikely to be one treatment or intervention that suits
all (Howlin, Magiati, & Charman, 2009; National Standards Report, 2009; Ospina et al., 2008; Roberts & Prior, 2006)oran
optimum level of intensity for all children with autism. These reviews of research of outcomes of early intervention for
children with autism at best demonstrate that some form of treatment is preferable to none. Summarising outcomes of any
particular intervention is challenging because of focus on different outcomes, different measures, and the heterogeneity of
the autism population (Krebs Seida et al., 2009). While there is consensus that more research is needed to demonstrate that
positive outcomes result from interventions and that Randomised Controlled Trials (RCT) are the most robust way to do this,
there are significant methodological and ethical difficulties associated with the conduct of RCT trials for learning-based
interventions. Difficulties include the heterogeneity of the autism population, recruitment of sufficient numbers of
participants, practical and ethical issues associated with the random assignment of participants to a ‘no treatment’ control
group, difficulty with ‘blinding’, limited availability of valid and reliable outcome measures, complexity and expense of
administering the intervention and outcome assessment and determining treatment fidelity (Krebs Seida et al., 2009).
The evaluation of a parent mediated communication focused intervention for children with autism (PACT) in the UK
indicates that large, high quality RCTs are possible. In this trial the characteristics of autism decreased for both the treatment
as usual and the treatment as usual plus PACT group with no statistically significant differences in Autism Diagnostic
Observation System (ADOS) scores. The authors suggest that ADOS scores may not be able to detect change, and as such
question their suitability as a primary outcome measure. However, objective measures are required especially when
blinding of parents and carers and outcome assessors to treatment group is not possible. On one observation measure used in
the PACT, measures of parent–child dyadic social communication, for which the assessors were ‘‘blinded’’, study participants
did significantly better. Positive change in parent–child dyadic social communication is associated with improved long-term
outcomes (Green et al., 2010).
In reviews of interventions for children with autism (Dawson & Osterling, 1997; NRC, 2001; Roberts & Prior, 2006) there is
considerable variation of the optimal amount of intervention (intensity) that is recommended, ranging from 15 to 25 h per
week. There is also the issue of intervention quality, in that it is essential to consider the nature of the intervention in
addition to the number of hours (Roberts & Prior, 2006), but this is seldom done. With research indicating generally that
effective programs are likely to be more intensive (at least 15 h per week) it is not surprising that intervention service
providers report an increase in the demand from families with children with autism for services that provide intensive early
intervention programs. Quantity, however, does not necessarily equate to quality for a particular child/family. Further, 15 or
more hours of intervention per week is costly to deliver and may not be available or accessible for families.
For programs with a focus on teaching parents to both understand the nature of their child’s difficulties and deliver
effective intervention outside the programs, it is plausible that intervention will carryover well beyond the direct contact
hours. In addition, there is some evidence that children may make some progress even with relatively low intensity
intervention (Vismara, Colombi, & Rogers, 2009).
Current research does not provide adequate information about the comparative effectiveness of home-based and centre-
based models of early intervention service delivery in terms of outcomes for children and their families and in relation to cost
benefit. For children with autism, Rickards, Walstab, Wright-Rossi, Simpson, and Reddihough, (2009) showed that a weekly
home visiting program (1.5 h per week for 12 months) in addition to a centre-based program (5 h per week for 12 months)
resulted in sustained gains in aspects of child cognitive development for the home visit plus centre-based group in
comparison to the centre-based only group. Home visits were most beneficial for families reporting high levels of stress.
Tonge et al. (2006) showed that a parent education and behaviour management skills program delivered to parents in
addition to treatment as usual for their children was associated with significant improvements in parental mental health and
adjustment for the treatment group compared to the matched control group. Green et al. (2010) showed that the addition of
a parent training program targeting enhanced social communication between the parent and child with autism resulted in
improved timing of parent child interaction and increased the amount of communication from the children. Behavioural
training for parents in combination with medication was more effective than medication alone in reducing complex severe
behaviour problems in children with autism (Aman et al., 2009). For service providers and for families, this type of reliable
empirically based information about program outcomes is useful to guide decisions about service provision. Knowledge of
the child and family, family priorities and clinical experience and judgement are also important.
Autism Spectrum Australia (Aspect) is a not for profit organisation which provides services for people with autism in
Australia. Aspect provides the Building Blocks
1
early intervention program for young children with autism and their
families. The Building Blocks model draws on a variety of approaches, including: naturalistic play-based intervention,
behavioural and developmental theory, structured teaching, the development of functional communication skills, positive
behaviour support, assessment of sensory processing issues, and extensive use of visual supports (Dawson & Osterling, 1997;
Roberts & Prior, 2006). The Building Blocks program provides both an individualised home-based service and a small group
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1554
centre-based service. The home-based service involves individualised 1:1 work with child and parent on a needs basis while
the manualised (Autism Association of NSW, 2004a) centre-based program involves intervention delivered to a small group
of 5–6 children by two staff and a simultaneous parent group, providing a manualised (Autism Association of NSW, 2004b),
structured, parent information, training and support program. Families in the Building Blocks program receive an average of
1–2 h of direct service per week. This falls well short of the recommended minimum 15 h per week, but is what was available
given the limited available resources at the time.
The current randomised trial was designed to evaluate the effectiveness of early intervention delivered to preschool
children with an ASD (Building Blocks program) by comparing two variations of the program; a home-based service and a
centre-based small group program. In addition, a wait list (WL) control group was recruited. Outcomes for children,
outcomes for families and staff cost of the home-based and centre-based service are assessed. The research set out to
determine whether there was a significant difference in terms of outcomes for children with autism and for their parents
between individual home-based intervention and small group centre-based intervention, and whether there was a
significant difference in the child and parent outcomes of home-based or centre-based programs compared with the
outcomes for a WL comparison group.
1. Institutional consent
Ethics approval was granted by The University of Sydney, Macquarie University, The Children’s Hospital at Westmead and
Autism Spectrum Australia (Aspect). Participant consent was obtained where appropriate.
2. Research design
A randomised controlled design was used to compare the centre-based and home-based Building Blocks programs over
two consecutive offerings of the programs (2006 & 2007). A wait-list comparison group was also included in each year.
3. Recruitment
There were two recruitment periods for this study, late 2005 and late 2006. The programs being investigated were one
year (40 weeks) in duration and 2 program cycles (2006 & 2007) were required to recruit sufficient participants. There was a
target enrolment of 30 children in each arm of the trial. The total number of children enrolled in the home-based and centre-
based programs during the study was 127 in 2006 and 141 in 2007 but not all met criteria for inclusion. Criteria for inclusion
included pre school age at the start of the program, a diagnosis of Autistic Disorder, Asperger or PDD-NOS according to DSM-
IV (American Psychiatric Association, 1994) made by a referring clinician (medical practitioner and/or psychologist),
domicile within reasonable distance of a centre-based group and consent to be involved. Parents and staff also judged the
child’s readiness for a centre-based program; in a few cases it was agreed that the child was not ready for a centre-based
program and these children were therefore not eligible for inclusion in the study. Some eligible families elected not to be
involved in the trial. Following consent, children and their families were randomised to either home-based or centre-based
intervention. Randomisation was conducted by a statistician using computer generated random number tables and the
project co-ordinator was informed of the randomisation result. Families who were eligible to receive a service but did not
meet the study inclusion criteria or did not agree to participate in the study, were allocated to the service of their choice as
per the usual procedure for the Building Blocks program. Families from the early intervention program waiting list were
recruited for the WL comparison group in both years. Children recruited to the study as the WL group for 2006 were then not
eligible for recruitment to the study for 2007. WL control families were offered places in the program of their choice in the
following year. Allocation could not be concealed from the child or family. Building Blocks staff running the home-based and
centre-based programs were not aware of which children and families in their groups were involved in the research.
Allocation concealment for randomised and the WL groups was maintained for the research investigators, data entry team
and baseline and outcome assessors until data had been entered and analysed.
4. Participants
Ninety-five eligible children were recruited to this trial, 67 of these were randomised (Fig. 1). Post assessments were
completed for 27 participants in the home-based and 29 in the centre-based programs and 28 on the wait-list comparison
group, resulting in a total of 84 participants completing the study. Eleven children withdrew from the study, 10 of these
withdrew prior to the commencement of intervention. One withdrew as a result of a staff decision that he was too stressed in
the small group setting. Twenty-eight families agreed to participate in the WL group with no attrition. Only data for
participants who completed the study are included in the analysis.
5. Baseline measures
The Autism Diagnostic Observation Schedule (ADOS) (Lord et al., 2000; Lord, Rutter, DiLavore, & Risi, 1999) was used as a
supplementary diagnostic measure. The Griffiths Mental Development Scales – Extended Revised (GMDS, Luiz et al., 2006)
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1555
was administered to provide the developmental quotient for each child. Baseline information was also collected about each
child’s age, gender, ethnicity, parental education status and first language at home. Parents were interviewed about current
strategies and services used to meet the needs of their child, and medical history of their child at baseline and at the end of
the treatment year.
5.1. Families
Family background information was obtained through pre-test administration of the Beach Centre Family Quality of Life
Scale (Summers et al., 2005), which was returned by post.
6. Outcome assessments
Primary dependant variables for this study for children with autism were; communication, social skills, adaptive
functioning and psychopathology; for the parents, dependant variables are levels of stress, perception of competence in
managing their child and quality of life. Table 1 shows the tests used to assess the primary dependent variables for this study
for children and adults. All tests to assess key domains were administered at baseline and treatment completion by
personnel experienced in the assessment of children with ASD.
Parent responses on the Pragmatics Profile of Everyday Communication (Dewart & Summers, 1995) were scored on a
developmental rating scale and scores converted to percentages of the maximum score for each of 8 functions and 8
responses (Jolleff, Emmerson, Ryan, and McConachie, 2006). The DBC is a 96 item parent completed questionnaire designed
to assess level of mental health and behaviour dysfunction in children with developmental disability.
A standardised parent measure, the Parenting Stress Index (PSI, Abidin, 1995), was chosen to provide a measure of parent
stress before and after the intervention. The Beach Family Quality of Life Questionnaire (Summers et al., 2005) was chosen to
evaluate parent perception of their quality of life. The Parent Perception Questionnaire, a non-standardised instrument, was
also administered. This questionnaire assesses parents’ understanding of ASD and feeling of competence in terms of coping
and management of their child with autism. Parents are asked to rate their confidence, coping skills, knowledge,
understanding, family issues and planning on a numerical scale (1–5). The questionnaire was developed for this project as no
standardised assessment of these constructs specific to autism was available.
[()TD$FIG]
Number of children randomised
2006=34
2007=33
Number of children consenting to
be on waiting list
2006=13
2007=15
Children who consented to be involved in the study N=95
Some outcome
data available
28
Home based group
2006=19
2007=15
Centre based group
commencing study
2006=13
2007=16
Some outcome
data available
27
Some outcome
data available
29
withdrawal
2006=5
2007=1
no loss
to follow-up
No loss to
follow-up
one loss
to follow-up
withdrawal
2006=2
2007=2
Home based group
commencing study
2006=14
2007=14
Centre based group
2006=15
2007=18
Randomised to
Fig. 1. Recruitment, randomisation and follow-up of all comparison groups.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1556
7. Intervention
Both the home-based and centre-based programs were staffed by transdisciplinary teams (teachers, speech pathologists,
occupational therapists and psychologists).
7.1. Centre-based program
Children in the centre-based program attended 40 weekly 2 h sessions. The centre-based program comprised six
playgroups of 4–6 children, with six concurrent parent support and training groups. The child playgroup component
involved children participating in a manualised condensed preschool program with the goal of preparing children for
integration into regular preschool settings by focusing on the development of social play skills, functional communication
skills and participation in small group activities. During these activities, staff worked with children to address individual
needs through direct intervention (e.g., speech difficulties, sensory routines, social skills). Children were also given time to
engage in routines and activities that were less directed.
The manualised parent training and support group component operated concurrently with the child program and
provided parents with an opportunity to meet weekly with other parents and professionals to discuss a range of set topics
which were prioritised according to individual interests and needs and included positive behaviour support,
communication, self-help issues, school options, specialist services and sensory issues. Parents were also encouraged to
share information, concerns and achievements and form a support network.
7.2. Home-based program
Children in the home-based program received a visit for 2 h once a fortnight over a 40 week period (20 sessions
maximum) from a member of the trans-disciplinary team. Staff from the program worked in the home with the family to
address identified needs (e.g., speech, sensory, social skills). An individual program was designed for each child in
conjunction with parents and other professionals involved in the child’s program. Major areas of focus included: assisting
parents to manage their child’s behaviour in a positive manner, teaching functional communication skills, enhancing social
development, extending attending and play skills, addressing sensory processing issues, developing self-help skills such as
independent toileting, implementing visual communication supports, improving fine and gross motor skills and developing
pre-academic skills. During this time the team member worked with both the child and with their parent/s to develop skills
in working with their child. Training techniques included direct modelling of skills, constructive feedback to the family on
how they manage the child’s education program, and discussion of issues that were immediate to the needs of the family.
There was an emphasis on training and supporting parents to work effectively with their children through play and natural
routines. Visits were made to the pre-school or day care setting in order to observe the child, provide strategies to staff and
support generalisation of skills.
Treatment fidelity checks of both the home-based and centre-based programs were made by a member of the research
team throughout the two years of data collection. A review of the files of all children at the end of their HB and CB programs
was conducted and a checklist completed to determine compliance with HB or CB program requirements. Relevant sections
of the children’s Building Blocks files were copied by a research team member, checked by the research team and stored with
the research files. The checklist used by the research team included program based pre and post assessment checklists
covering all relevant developmental domains completed for each child, documentation of individual programs; that is
written individual plans, record of reviews and final evaluations for each child using program proformas. Progress notes were
also checked including notes about meetings with parents. For the HB program file checks were conducted to establish that
core domains outlined in the program such as communication development were addressed. In addition for the CB programs
files were checked and 10% of the CB sessions in all groups were observed to evaluate compliance with the manualised
Table 1
Validated tools used to measure dependent variables for children and adults.
Outcome domain Tool Method of completion Improvement
indicated by score
Communication and Social skills The Vineland Adaptive Behaviour
Scales II (VABS) (Sparrow, Balla, & Cicchetti, 2005)
Interview with parent Increase
Communication Reynell Developmental Language
Scales III (RDLS-III) (Edwards et al., 1997)
Child assessment Increase
Communication The Pragmatics Profile
(Dewart & Summers, 1995)
Interview with parent Increase
Adaptive functioning and
psychopathology
The Developmental Behaviour
Checklist (DBC) (Einfeld & Tonge, 2002)
Interview with parent Decrease
Stress Parenting Stress Index (PSI)
(Short form) (Abidin, 1995)
Interview with parent Decrease
Quality of life The Beach Family Quality of
Life Questionnaire (Summers et al., 2005)
Interview with parent Increase
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1557
program for the child and parent groups. Measures taken indicated a high level of treatment fidelity. Each child in the HB and
CB group had a completed individual program with all required sections and all CB groups for children and parents observed
by a team member were conducted as outlined in the Building Blocks Manual.
8. Other interventions
Parents/carers were asked to report any interventions utilised during the 12 month study period, in addition to the home-
based and centre-based treatment provided as part of this trial. For analysis these interventions were classified as either
autism specific interventions (i.e., ABA, Hanen, Floor time, Giant steps, RDI) or conventional therapies (including speech and
language, occupational therapies) (Carter et al., 2010).
9. Statistical power
Sample size calculations prior to recruitment indicated that with approximately 30 children enrolled in each group using
one-way ANOVA or ANCOVA with one covariate, that an effect size of 0.7 would be detected with over 80% power (
a
0.05).
9.1. Statistical analyses
Data were analysed using SPSS version 16.0. In all analyses, P-values of <0.05 were considered statistically significant. To
avoid type II errors no adjustment was made for multiple comparisons (Perneger, 1998). Descriptive statistics were obtained
for baseline and follow-up data for all variables to verify that the distributions were approximately normally distributed in
order that parametric statistics could be used. To assess whether changes over time were statistically significant, paired t-
tests were used within each group to obtain mean changes and their 95% confidence intervals (95% CI). To assess whether
there were statistical differences between outcome variables at follow-up, analysis of covariance was used with baseline
score as the covariate and group as the explanatory variable. Post hoc tests were used to compare between group differences
and obtain 95% CIs with the difference between the HB group and the centre-based group being of most interest.
Table 2 shows family characteristics. Family background information was obtained through pre-test administration of
the Beach Centre Family Quality of Life Scale (Summers et al., 2005), which was returned by post. Most families reported an
income of over $75,000 year and the mean number of family members supported by this income was 4.0.
10. Baseline measures
10.1. Children
At the commencement of the trial, the mean age of children was 3.5 years (range 2.2–5.0, SD = 0.61) (Table 3) and 90.5%
were male. The mean Griffiths developmental quotient was 62.4 (SD = 15.5, n= 80). Of the 82 completed ADOS assessments,
59 children were classified with autistic disorder, 13 autism spectrum disorder and 10 tested as not being on the autism
spectrum, although a referring clinician had made a diagnosis of autistic disorder, Asperger’s disorder or PDD-NOS according
to DSM IV.
Table 3 shows clinical characteristics and additional interventions received. Despite randomisation, the HB group had a
higher proportion of children with a diagnosis of autistic disorder and a lower mean GQ compared to the other groups. There
was no significant difference in the total number of educational interventions used across the groups but the WL group used
Table 2
Demographic characteristics.
Characteristic NNumber (%)
Language spoken at home 78 Language other than English spoken exclusively 2 (2.6%)
A language in addition to English 12 (15.4%)
Household income 78 More than $75,000 per year 45 (57.7%)
Between $60,000 and $75,000 10 (12.8%)
Between $50,000 and $60,000 11 (14.1%)
Between $40,00 and $50,000 4 (5.1%)
Less than $40,000 8 (10.3%)
Mothers’ education 73 High school 10 (13.7%)
College or post-high school training 28 (38.4%)
Bachelors degree 23 (31.5%)
Postgraduate education 12 (16.4%)
Mean (SD)
Family members supported by income 78 Number 4.0 (SD = 1.2)
SES
a
80 Ranking within NSW 73.0 (SD = 23.0)
Mothers’ age 75 Years 36.6 (SD = 4.3)
a
SES was determined according to the 2006 Postal Area Index of Relative Socio-economic Advantage and Disadvantage (Australian Bureau of Statistics,
2008). In 2006, the mean total household income in Australia was approximately $81,000 (Australian Bureau of Statistics, 2007).
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1558
significantly more ASD specific interventions (i.e., ABA, Hanen, Floor time, Giant Steps, RDI) than the CB group (P= 0.02) with
no other differences reaching significance (Carter et al., 2010).
Also of note, in the HB group in particular, some participants failed to score on the formal language comprehension
measure at both pre and post assessment; that is, these children were nonverbal receptively and expressively at the start and
finish of the intervention. Their language age was below the 12 month baseline for the Reynell Developmental Language
Scale (Edwards et al., 1997).
10.2. Within group changes for child measures
Table 4 shows that statistically significant improvements were made by all groups on the communication scale of the
Vineland Adaptive Behaviour Scales (Sparrow, Balla, & Cicchetti, 1984) with the CB group having the largest improvement.
Improvements were also made for the social scale of the Vineland in the CB and WL groups, but these were not statistically
significant. All groups improved for raw scores of the Reynell, with a statistically significant improvement for the CB group
for the expression raw score. The only improvement for Reynell standard scores was made in the CB group for the
comprehension standard score, but this was not significant. Assessment of functional communication using the Pragmatics
Profile showed that all three groups made large, statistically significant improvements. The CB and WL groups also improved
on the DBC and the HB group did not, although the changes were not statistically significant.
10.3. Within group changes for parent measures
Table 5 shows that within group changes in quality of life, as measured by the Beach, were small and varied slightly
between the groups. The HB group showed no improvement or a slight worsening in all domains but all changes were non-
significant. The CB group showed improvements in all domains with a significant improvement in the parenting domain and
total score and a marginally significant improvement in emotional wellbeing. The WL group improved in some domains with
a significant improvement in the parenting domain.
There were no significant changes in parent perception in the HB group. The CB group, however, had consistent
improvement across all subscales of the parental perception questionnaire, with significant improvements in confidence,
knowledge, understanding, planning and total score and a marginally significant increase in coping. The WL group had
significant improvements in knowledge, understanding, planning and total score.
Table 5 shows that parental stress as measured by the PSI total stress score increased in the HB group, decreased for the CB
group (neither significant) and decreased in the WL group (significant). The WL group also had significant improvements in
the defensive and parental distress sub-domains as measured on the PSI.
11. Between group comparisons
11.1. All children
Table 6 presents the between group comparison of the HB and CB groups for child outcomes. When follow-up scores are
adjusted for baseline scores, the CB group had close to statistically significant improvement for the communication scale and
a statistically significant improvement for the social scale of the VABS. The WL group improved significantly more than the
HB group for the social scale of the VABS but there were no other significant differences in the VABS between the WL and HB
or CB groups.
There was a statistically significant greater improvement in Reynell comprehension standard score for the CB group
compared to the HB group (7.3; 95% CI (13.9, 0.7), P= 0.02). There was a greater improvement for expression standard
score of the Reynell for the CB compared to HB group but this was not significant (3.0; 95% CI (9.0, 2.9), P= 0.31). For the
Table 3
Baseline characteristics with valid percent of group or mean (SD).
HB group CB group WL group
Total number 28 29 28
Age 41.5 (26.5–59.4) 43.1 (26.3–60.0) 43.7 (27.6–60.3)
ADOS
Autistic disorder 87.5% 69.0% 60.7%
ASD 14.3% 13.8% 17.9%
Non-ASD 0.0% 17.2% 21.4%
Griffiths Developmental Quotient (SD) 57.0 (11.7) 66.5 (17.7) 63.3 (15.5)
Mean number of ASD specific interventions
a
used
during the intervention period (SD)
0.22 (0.42) 0.14 (0.35) 0.54 (0.79)
Mean number of educational interventions
b
used
during the intervention period (SD)
2.37 (1.28) 2.41 (1.50) 3.11 (1.64)
a
Autism specific therapies include ABA, Hanen, floor time, Giant steps, RDI.
b
Includes all educational interventions identified by familie.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1559
Table 4
Within-subject changes in domains in child outcomes.
HB CB WL
Baseline
mean (SD)
Follow-up
mean (SD)
Change (95% CI) P-value
*
(N) Baseline
mean (SD)
Follow-up
mean (SD)
Change
(95% CI)
P-value
*
(N) Baseline
mean (SD)
Follow-up
mean (SD)
Change
(95% CI)
P-value
*
(N)
Vineland
a
Communication 64.4 (12.8) 68.4 (15.6) 4.0 (0.03, 7.9)
a
0.048 (26) 66.9 (12.5) 76.1 (17.1) 9.2 (3.6, 14.8) 0.002 (29) 68.5 (17.0) 74.2 (15.5) 5.7 (1.5, 9.8) 0.01 (28)
Social 68.7 (7.3) 66.4 (7.7) 2.3 (4.6, 0.02)
b
0.052 (27) 70.1 (7.3) 72.6 (11.2) 2.5 (1.1, 6.1) 0.17 (29) 70.8 (9.9) 73.1 (10.8) 2.4 (1.5, 6.2) 0.22 (28)
Reynell
a
Comprehension
Standard score
4.2 (9.2) 2.6 (8.4) 1.6 (5.5, 2.4)
b
0.82 (27) 5.5 (10.6) 10.5 (17.4) 5.0 (0.2,10.1) 0.12 (26) 7.2 (15.2) 5.7 (12.1) 1.5 (8.4, 5.5) 0.97 (26)
Comprehension
Raw score
6.9 (9.7) 17.5 (6.3) 10.7 (6.7, 14.6) <0.0001 (27) 11.3 (13.8) 23.7 (19.9) 12.4 (7.7, 17.1) <0.0001 (26) 12.2 (14.9) 22.0 (17.8) 9.7 (4.9, 14.5) 0.001 (26)
Expression
Standard score
3.4 (8.3) 2.8 (7.5) 0.6 (5.3, 4.1)
b
0.61 (26) 8.2 (16.6) 7.0 (15.1) 1.1 (7.9, 5.6) 0.69 (26) 6.0 (10.9) 4.4 (8.7) 1.6 (7.0, 3.8) 0.83 (26)
Expression
Raw score
3.2 (5.4) 8.8 (8.9) 5.6 (3.2, 7.9) <0.0001 (27) 6.9 (9.9) 11.4 (10.9) 4.5 (0.9, 8.1) 0.004 (26) 5.8 (7.9) 11.1 (9.9) 5.3 (2.7, 7.8) 0.001 (26)
Pragmatics profile
a
Total Q range 50.4 (17.5) 62.8 (19.4) 12.4 (9.0, 15.8) <0.0001 (27) 58.3 (16.8) 73.0 (19.0) 14.7 (10.1, 19.3) <0.0001 (29) 56.7 (16.2) 72.2 (18.8) 15.5 (10.2,20.8) <0.0001 (27)
DBC
b
Total 44.7 (19.0) 52.9 (29.3) 8.1 (3.4, 19.8) 0.16 (22) 58.5 (20.4) 55.7 (19.5) 2.8 (8.1, 2.4) 0.28 (22) 43.9 (21.9) 42.9 (24.3) 1.0 (7.9, 5.8) 0.76 (26)
P-value for Reynell from Wilcoxon paired test, P-values for all other instruments from paired t-tests.
a
Higher score indicates improvement.
b
Lower score indicates improvement.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1560
Table 5
Within-subject changes in domains for parent outcomes.
HB CB WL
Baseline
mean (SD)
Follow-up
mean (SD)
Change
(95% CI)
P-value (N) Baseline
mean (SD)
Follow-up
mean (SD)
Change
(95% CI)
P-value Baseline
mean (SD)
Follow up
mean (SD)
Change
(95% CI)
P-value (N)
Beach satisfaction
a
Family
interaction
4.0 (0.8) 3.9 (0.6) 0.1 (0.4, 0.3) 0.84 3.6 (0.8) 3.8 (0.8) 0.1 (0.2, 0.5) 0.41 3.8 (0.7) 3.9 (0.6) 0.1 (0.1, 0.3) 0.41
Parenting 3.8 (0.8) 3.7 (0.7) 0.1 (0.4, 0.2) 0.48 3.2 (0.7) 3.7 (0.8) 0.4 (0.2, 0.7) 0.001 3.6 (0.6) 3.9 (0.7) 0.3 (0.1, 0.6) 0.02
Emotional
wellbeing
3.5 (1.0) 3.5 (0.8) 0.0 (0.4, 0.4) 1.00 3.0 (0.8) 3.3 (1.0) 0.3 (0.0, 0.7) 0.06 3.3 (0.8) 3.4 (0.7) 0.2 (0.2, 0.5) 0.42
Physical
wellbeing
4.1 (0.6) 3.9 (1.0) 0.2 (0.6, 0.2) 0.32 3.9 (0.5) 3.9 (0.8) 0.0 (0.3, 0.3) 0.95 4.1 (0.5) 4.0 (0.8) 0.1 (0.4, 0.2) 0.67
Disability
support
4.0 (0.8) 3.8 (1.1) 0.2 (0.7, 0.2) 0.30 3.6 (0.7) 3.7 (0.8) 0.1 (0.2, 0.4) 0.49 3.9 (0.7) 3.9 (0.7) 0.0 (0.3, 0.2) 0.94
Total score 3.9 (0.7) 3.8 (0.6) 0.1 (0.4, 0.2) 0.53 (21) 3.5 (0.5) 3.7 (0.6) 0.2 (0.0, 0.4) 0.02 (23) 3.7 (0.5) 3.8 (0.5) 0.1 (0.1, 0.3) 0.37 (26)
Parental
perception
a
Confidence 3.8 (0.7) 3.7 (0.8) 0.1 (0.5, 0.2) 0.44 3.3 (0.7) 3.7 (0.7) 0.4 (0.2, 0.6) 0.002 3.9 (0.6) 4.0 (0.7) 0.0 (0.2, 0.3) 0.73
Coping 3.8 (0.8) 3.6 (0.9) 0.2 (0.5, 0.2) 0.27 3.0 (0.9) 3.3 (0.9) 0.3 (0.0, 0.5) 0.07 3.6 (0.8) 3.6 (0.7) 0.0 (0.3, 3.0) 0.79
Knowledge 3.8 (0.6) 3.9 (0.7) 0.1 (0.2, 0.4) 0.44 3.5 (0.6) 4.2 (0.4) 0.8 (0.5, 1.0) <0.0001 3.7 (0.5) 4.1 (0.5) 0.3 (0.1, 0.6) 0.007
Understanding 3.9 (0.6) 3.9 (0.8) 0.0 (0.5, 0.4) 0.85 3.4 (0.8) 4.1 (0.7) 0.7 (0.3, 1.1) 0.001 3.8 (0.6) 4.2 (0.6) 0.5 (0.2, 0.7) <0.0001
Family issues 3.8 (0.8) 3.8 (0.7) 0.0 (0.5, 0.5) 0.95 3.4 (0.7) 3.6 (1.0) 0.2 (0.2, 0.6) 0.25 3.7 (0.9) 3.7 (0.8) 0.0 (0.3, 0.3) 0.87
Planning 3.5 (0.9) 3.6 (1.0) 0.1 (0.6, 0.8) 0.79 3.0 (0.7) 3.7 (1.2) 0.7 (0.2, 1.2) 0.01 3.4 (0.8) 3.8 (0.8) 0.4 (0.1, 0.7 0.02
Total 3.8 (0.6) 3.7 (0.7) 0.0 (0.4, 0.4) 0.88 (23) 3.3 (0.6) 3.8 (0.6) 0.5 (0.3, 0.5) <0.0001 (23) 3.7 (0.5) 3.9 (0.5) 0.2 (0.0, 0.4) 0.047 (26)
PSI
b
Defensive 19.7 (6.7) 19.5 (6.5) 0.2 (2.3, 1.9) 0.86 22.1 (5.9) 20.8 (5.9) 1.3 (3.6, 0.9) 0.23 21.8 (5.5) 19.4 (5.5) 2.4 (4.2, 0.6) 0.01
PD 31.5 (10.2) 31.9 (9.3) 0.4 (3.1, 3.9) 0.82 34.6 (8.9) 33.9 (8.7) 0.8 (4.1, 2.6) 0.64 35.7 (9.7) 32.5 (9.5) 3.2 (6.0, 0.4) 0.03
P.CDI 26.3 (5.6) 27.5 (6.8) 1.2 (1.2, 3.6) 0.31 27.5 (4.5) 28.5 (6.3) 1.0 (1.2, 3.2) 0.35 27.3 (8.6) 25.5 (8.2) 1.8 (4.2, 0.6) 0.14
DC 34.9 (9.2) 35.4 (10.1) 0.5 (3.0, 4.0) 0.76 39.7 (8.9) 38.4 (6.0) 1.3 (4.8, 2.2) 0.45 34.7 (8.8) 33.1 (8.2) 1.6 (4.6, 1.3) 0.26
Total stress 90.4 (18.9) 92.7 (20.9) 2.4 (4.6, 9.3) 0.49 (20) 103.4 (19.6) 98.2 (20.1) 5.2 (16.1, 5.7) 0.33 (20) 97.1 (22.0) 90.4 (22.9) 6.7 (12.7, 0.8) 0.03 (22)
a
Higher score indicates improvement.
b
Lower score indicates improvement.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1561
Reynell standard comprehension and expression scores the WL group performed better than the HB group, but not
significantly. For the Reynell standard comprehension and expression scores the CB group performed better than the WL
group but not significantly.
There were no statistically significant differences between the three groups for the other child outcomes as measured by
the Pragmatics Profile or DBC.
12. Autistic Disorder children as classified on the ADOS
When the analyses for the Vineland and Reynell scores were limited to the children with autistic disorder only as
classified on the ADOS there was an increase in the magnitude of effect for better outcomes in the CB group compared to the
HB group for Vineland social (effect size 6.0, P= 0.02) and communication score (7.5, P= 0.05) and Reynell comprehension
(11.5, P= 0.001) and expression standard scores (5.9, P= 0.07), but the presence or absence of statistical significance of the
difference did not change.
13. Children not receiving other autism specific treatments
When the analysis was limited to children not receiving other autism specific treatments the positive effect size for the CB
group compared to the HB group decreased for the Vineland social (4.1, P= 0.10) and communication score (1.4, P= 0.69)
and Reynell comprehension (7.1, P= 0.07) and expression standard score (1.9, P= 0.59) and the differences were no longer
statistically significant for the Vineland social and communication scores and the Reynell comprehension score.
14. All parents
Table 7 presents the between group comparison of the HB and CB groups for parent outcomes. For parent related
outcomes measured by the BEACH, there were statistically significant differences favouring the CB group above the HB group
for parenting and favouring the HB group above the CB group for disability support. For parental perception, there was a
statistically significant difference favouring the CB above the HB group for knowledge. There was no significant difference
between groups for stress as measured by the PSI.
15. Staff time for programs
For the centre-based program children met as a group of six with two therapists weekly for 40 weeks, that is 40 2h
sessions with two staff; 160 staff hours of intervention per group. In addition parents met for 40 weeks with one staff, a
further 80 h of intervention. Therefore in total each group received 240 staff hours of intervention. Because there were six
children/parents in each group, each child/family received 40 staff hours. Children receiving home-based intervention
received 20 2 h sessions with a staff member (that is, 40 h of 1:1 intervention). Therefore, despite the fundamental
differences in the programs, including increased hours of intervention for the centre-based program, because of the group
setting and concurrent parent components, the actual staff hours required to deliver these different models was the same.
The increasing number of children with autism being identified at younger ages (Williams et al., 2005) focuses attention
on finding effective interventions for this population, increases the need to focus on parents as key deliverers of intervention
Table 6
Between group differences
*
for child outcomes.
HB vs CB HB vs WL CB vs WL Three group
comparison
Mean difference
(95% CI)
P-value Mean difference
(95% CI)
P-value Mean difference
(95% CI)
P-value P-value
Vineland Communication 5.8 (0.5, 12.1,)
a
0.07 2.6 (3.8, 9.0)
c
0.42 3.2 (3.0, 9.3)
a
0.31 0.19
Social 5.2 (0.7, 9.6)
a
0.02 5.2 (0.7, 9.7)
c
0.02 0.1 (4.3, 4.4)
c
0.98 0.03
Reynell Comprehension
standard score
7.3 (0.7, 13.9)
a
0.03 1.8 (4.9, 8.4)
c
0.60 5.5 (1.2, 12.2)
a
0.10 0.08
Comprehension
raw score
1.3 (4.9, 7.5)
a
0.67 1.5 (4.7, 7.7)
b
0.63 2.8 (3.4, 9.0)
a
0.37 0.66
Expression
standard score
3.0 (2.9, 9.0)
a
0.31 0.9 (4.9, 6.8)
c
0.75 2.1 (3.8, 8.0)
a
0.48 0.58
Expression
raw score
0.6 (3.4, 4.6)
b
0.76 0.1 (3.9, 4.0)
c
0.99 0.6 (3.3, 4.6)
c
0.75 0.94
Pragmatics Total Q range 3.1 (3.2, 9.3)
a
0.33 3.7 (2.7, 10.0)
c
0.25 0.6 (6.8, 5.6)
a
0.89 0.47
DBC Total 8.1 (3.6, 19.8)
a
0.17 9.4 (1.5, 20.2)
c
0.09 1.3 (10.0,12.6)
c
0.82 0.20
*
All comparisons adjusted for baseline scores.
a
Favours CB.
b
Favours HB.
c
Favours WL
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1562
and highlights the need to support parents and promote resilience. In a field characterised by a lack of methodologically
robust outcomes research (McConachie & Diggle, 2007), reliable information about the efficacy of different approaches is
scarce. The results of this study demonstrate that outcomes for children with autism and their parents in small centre-based
groups providing both structured programs for young children with an ASD and a simultaneous structured group program
for their parents, tend to be better overall for both children and parents than outcomes for children and parents receiving a
home-based program.
All three groups of children (CB, HB & WL) made gains over the intervention period in some domains of behaviour,
communication or social interaction as assessed by the outcome measures used. The gains made on the VABS and the
Pragmatics Profile suggest that children’s functional communication in particular improved. The change is significant for all
groups on the Pragmatics Profile and for the CB group on the VABS. While formal language raw scores (expression and
comprehension) improved significantly across the groups between pre and post assessments, the Reynell standard scores
show that these children were mainly loosing ground relative to language development of their typical peers. Reynell
standard scores decreased for both expression and comprehension in the HB, and WL groups and in the CB group for
expression. There was a positive but non-significant gain in the CB group for the mean comprehension standard score that is
encouraging.
Outcomes for parents were mixed. When CB and HB groups are compared for parental stress, as measured by the PSI,
baseline differences varied, indicating that although the HB group increased their stress scores on average during the trial
from a low baseline score, and the CB group decreased theirs from a high baseline score, the differences were not significant
when analysis adjusted for baseline scores.
Parents’ perception of competence showed consistent improvement across all subscales and a significant improvement in
total score at the end of the intervention for the CB group in contrast to variable outcomes for the HB parent group with no
significant change. This difference persisted when adjustments were made to account for baseline differences between the
HB and CB groups.
When CB and HB groups were compared accounting for baseline differences, the difference between the CB and HB
groups was supported for three subscales of quality of life. Comparison of the three groups indicates CB parents still showed
the most gains overall. For quality of life the CB group parents had the best outcomes followed by the WL group, which
showed some significant improvements on subscales measuring quality of life but no significant change in total scores. The
HB group showed most gains in disability support but the least improvement in quality of life over all.
In comparison with the CB and HB groups, the WL group also showed some significant improvements on parent
perception subscales but no significant change in total score. When CB and WL were compared adjusting for baseline
differences, only one subscale on the perception scale showed significant difference, parent knowledge improved more in the
CB program suggesting that the small group environment was more effective in improving parent understanding of autism
and its management, possibly because parents were able to share knowledge with each other as well as with professionals. In
addition there is a specific focus in the small group CB program on improving parent knowledge, whereas the parent
educational component in the home-based program is unstructured. The CB group also had better outcomes for parents in
Table 7
Between group differences
*
for parent outcomes.
Home vs CB Home vs WL Centre vs WL Three group
comparison
Mean difference
(95% CI)
P-value Mean difference
(95% CI)
P-value Mean difference
(95% CI)
P-value P-value
Beach satisfaction Family interaction 0.1 (0.3, 0.4)
a
0.92 0.1 (0.3, 0.4)
c
0.72 0.1 (0.3, 0.4)
a
0.80 0.93
Parenting 0.4 (0.0, 0.7)
a
0.05 0.3 (0.0, 0.6)
c
0.08 0.1 (0.3, 0.4)
c
0.71 0.11
Emotional wellbeing 0.0 (0.5, 0.5) 0.92 0.0 (0.5, 0.4) 0.89 0.0 (0.4, 0.4) 0.98 0.99
Physical wellbeing 0.2 (0.3, 0.6)
a
0.52 0.2 (0.3, 0.6)
c
0.51 0.0 (0.4, 0.4) 0.99 0.76
Disability support 0.4 (0.0, 0.8)
b
0.04 0.1 (0.2, 0.5)
b
0.45 0.2 (0.6, 0.1)
c
0.18 0.13
Total score 0.1 (0.2, 0.24
a
0.35 0.1 (0.1, 0.4)
c
0.36 0.0 (0.3, 0.3) 0.93 0.57
Parental perception Confidence 0.3 (0.1, 0.7)
a
0.09 0.2 (0.1, 0.6)
c
0.19 0.1 (0.3, 0.5)
c
0.46 0.20
Coping 0.2 (0.2, 0.6)
a
0.40 0.1 (0.3, 0.5)
c
0.55 0.1 (0.3, 0.5)
c
0.75 0.68
Knowledge 0.4 (0.1, 0.7)
a
0.01 0.2 (0.1, 0.5)
c
0.22 0.2 (0.1, 0.5)
c
0.13 0.04
Understanding 0.3 (0.2, 0.7)
a
0.25 0.3 (0.1, 0.7)
c
0.11 0.1 (0.3, 0.5)
a
0.74 0.26
Family issues 0.0 (0.5, 0.5) 0.99 0.0 (0.4, 0.5) 0.91 0.0 (0.5, 0.5) 0.92 0.99
Planning 0.2 (0.4, 0.8)
a
0.57 0.2 (0.4, 0.8)
c
0.49 0.0 (0.6, 0.6) 0.93 0.76
Total 0.3 (0.2, 0.6)
a
0.14 0.2 (0.1, 0.5)
c
0.26 0.1 (0.3, 0.4)
c
0.62 0.30
PSI Defensive 0.4 (2.3, 3.0)
a
0.77 1.5 (1.0, 4.1)
c
0.22 1.2 (1.4, 3.7)
a
0.36 0.44
PD 0.2 (4.0, 4.3)
a
0.94 2.3 (1.6, 6.1)
c
0.25 2.1 (1.8, 6.0)
a
0.28 0.42
P.CDI 0.1 (3.1, 3.2)
b
0.96 2.7 (0.3, 5.8)
c
0.08 2.8 (0.3, 5.9)
a
0.07 0.12
DC 0.2 (3.9, 4.4)
b
0.90 2.2 (1.6, 6.1)
c
0.25 2.5 (1.5, 6.4)
a
0.22 0.37
Total stress 3.4 (7.4, 14.2)
a
0.53 6.9 (3.4, 17.2)
c
0.18 3.5 (6.7, 13.8)
a
0.49 0.41
*
All comparisons adjusted for baseline scores.
a
Favours CB.
b
Favours HB.
c
Favours WL
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1563
terms of the value placed on the importance of parenting, suggesting greater empowerment and satisfaction with disability
support, possibly related to increased capacity to access support more effectively. Findings suggest that the CB model is more
effective in increasing parental resilience and capacity.
With the exception of the stress measure, these results suggest that parents in the CB group benefited most from the small
group structure and the parent program. A potential influence on outcomes for parents in the HB group generally is the more
severe nature of their child’s disability in terms of ASD diagnosis and cognitive ability reflected in baseline data (despite
randomisation between the CB & HB groups).
Analyses adjusted for diagnostic category, comparing children only with a diagnosis of autistic disorder, and other
treatments received, comparing children not receiving other autism specific treatments, reduced the sample size and hence
the power of the analyses. However the trend toward greater improvement for the CB group persisted.
It is important to note that the CB program was not an option for all children referred. Some children lacked sufficient
social maturity or to have high levels of ‘problem behaviours’, which would preclude them from benefitting from the small
group program at this stage in their development. These children (who were not eligible for the study) were assessed as
needing a home-based program to prepare them for the small group program. In addition for a few parents, participation in a
parent group, an essential part of the CB program, was not an option. Therefore it is evident that different types of programs
are required to best meet individual needs of children and families reflecting the heterogeneity of the autism population and
a key component of evidence-based practice. The evidence presented in this paper, however, suggests that, except where
contraindicated by specific family circumstances or child characteristics, a small group centre-based program should be
considered as the preferred option.
These findings are consistent with other research. Lack of statistically significant improvement on formal language
measures is characteristic of other methodologically robust outcome evaluations. For example, Keen, Roger, Doussin, and
Braithewaite (2007) found that a significant improvement in parent ratings of the child’s communication was not supported
by independent observation. Green et al. (2010) found that the addition of a parent training program in communication
facilitation was not associated with improvement on formal language measures, although there was a positive effect for
child-parent dyadic interaction and for parent perception of the child’s communication. The current study indicates that
functional communication can be improved and that a coordinated, structured, child program can have a positive effect on
formal comprehension measures at least. It should be noted that in this study in the HB group in particular, some participants
failed to score on the formal language comprehension measure at both pre and post assessment, that is, these children were
nonverbal receptively and expressively at the start and finish of the intervention. Their language age was below the 12
month baseline for the Reynell. As in this study, Tonge et al. (2006) also found that there was no significant change or
difference between ‘‘treatment as usual’’ and ‘‘treatment as usual plus parent training’’ groups on the DBC, a measure of child
psychopathology.
It is possible that the baseline difference in parent stress found in this study was not a random difference and that parents
who were visited in the home initially reported lower stress than those who had to travel to a new centre for intervention
and participate with other parents in a group. As the program progressed, however, parents in the CB group appear to have
benefited from group mentoring and support unavailable to the HB program parents. It is also possible that parents in the HB
group were feeling more stressed at the time the post intervention measurements were completed because a supportive
program had just finished, leaving the parents feeling abandoned. In contrast, parents involved in the CB group continued to
have a support network provided by each other, even though the formal intervention had finished. No effect on parent stress
immediately post treatment has also been shown by Keen et al. (2007) and Tonge et al. (2006), although decreased stress at 6
month follow-up was found in the latter study.
In this study the finding that the CB group parents did better overall is not surprising when the difference in intensity is
considered. CB group met weekly for 4 h (a total of total of 80 h over the program, 40 h child group & 40 h parent group) while
the HB group children and families had 1 2h visit per fortnight (40 h). In addition in the CB group the parents had 2 h per
week focused on their needs.
16. Strengths and weaknesses
Children participating in this study (HB, CB & WL groups) typically received several interventions in the course of the
study. There were differences between the groups in relation to the mean number of autism specific interventions they
received. When analysis was adjusted for this, although significance was lost, a trend toward the CB group benefiting more
than the HB group remained. One of the methodological issues for the conduct of RCTs of autism interventions is that it is not
possible to recruit participants receiving no treatment. In this study, however, the overall number of additional interventions
was similar across the three groups.
The heterogeneity of the autism population, which is reflected among the participants recruited for this study is an issue in
all autism trials. Despite random assignment to HB and CB groups there were significantly more children in the HB group
with a diagnosis of autistic disorder on the ADOS compared to autism spectrum or non-ASD than in the CB and WL groups,
suggesting that on average the children in the HB group had more severe autism. The mean DQ (a measure of developmental
ability) for the HB group was also lower than for the CB and WL groups. Analysis of the data has in part controlled for these
differences when comparing the groups and ‘levelled the playing field’ to enable comparison of the groups accounting for
differences at pre assessment. It is likely, however, that moderate developmental disability in the HB group, as compared to
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1564
mild developmental disability in the CB and WL groups, and greater proportion of participants with a diagnosis of AD in the
HB group are contributing factors to the minimal change in this group. In addition the home-based educational program
would have been more focused on preverbal, play and social initiation/reciprocity goals. Change in skills at this level would
be less likely to show up on the measures used in this study. This is another study design problem that applies to autism
trials; the lack of suitable standardised assessments for early developing social/communication skills. The more significant
impairment of the children in the HB group may also be a factor influencing outcome measures for parents, for example the
increase in parental stress over the duration of the study.
17. Loss to follow-up
In this study eligible and consenting participants were randomly assigned to either centre-based or home-based early
intervention. Some families, however, were lost to the study following randomisation because of a strong preference to
participate in the alternative program and in one case because staff made the decision to move a child from a centre-based to
an home-based program once he started the centre-based program and was showing signs of stress. Families were also lost
to follow up as a result of discontinuing the program for a variety of reasons including moving interstate and the death of a
parent. Some families completed the program but did not participate in all follow-up assessments for logistical reasons.
18. Measures
Finding valid and reliable standardised assessments is an ongoing challenge, in particular because current standardised
tools are often not suitable for measuring change in the more significantly impaired autistic population especially over the
relatively short time frame of many trials. In addition, not all outcomes measured for autism trials are important to children
with autism and their families or the professionals caring for them. The tools we chose reflect key areas of child development,
measure areas known to cause difficulties for children with autism and have been used in similar research. We also included
measures that would provide information about both child and family/parent outcomes. In addition, for trials where
blinding to treatment group is not possible for parents and children, outcome measures that do not require parent
completion are needed, but few are available.
19. Intensity of intervention and range of measures
Both the home-based (1 fortnightly session 2–3 h) and the centre-based group (1 weekly session 2–3 h) are low
intensity interventions. Children in the HB group received 4–5 visits during school terms, i.e., 18 sessions per year while the
CB group met for 9–10 sessions per term i.e., 36 sessions a year. The centre-based intervention for children and parents
accounted for a total of 144 shared hours of intervention, the home-based program 36 h in total. By comparison many
programs based on behavioural theory and interventions based on transactional developmental theory recommend
intensity of 20–40 h per week for a period of two or more years (1000 h). As such both interventions compared in this study
were of similar intensity, and known to be of lower intensity than other therapies. Children and families did access other
autism specific therapies during the intervention period and some of these are known to be of high intensity. Information
collected for this trial, however, did not include sufficient detail about contact hours of additional therapies to allow for
meaningful adjusted analysis.
The cost effectiveness analysis for this study was limited to staff costs and outcomes as monitored for the trial. When
costs are calculated on a staff hours per child basis the cost per child is $6383AU regardless of which treatment the child
received. This is a small cost compared to a range of other interventions currently available to children and families with
autism.
Given that most children in this study were concurrently receiving some form of therapy relevant to their problem profile
and autism specific interventions, it is problematic to assess the benefit of these programs in general, despite our attempts to
have a suitable comparison group (the WL group). In cases where a child and family are assessed as being equally eligible for
a home-based or centre-based program the results of this study suggest that the centre-based program is likely to provide
better outcomes.
In conclusion, the findings of this research support the provision of a coordinated intervention for both children with
autism and their parents that focuses on the development of social communication skills in particular for children and on the
provision of information, skill development and support for their parents. It is also clear, however, that small group centre-
based early intervention for autism will not meet the needs of all young children with autism and their families. Account
must be taken of the child’s developmental stage, the severity of autism characteristics and family capacity. Intervention
planning needs to be individualised to meet the needs of children with autism most effectively. There is a trend toward
increasingly earlier identification of larger numbers of children with an autism spectrum disorder. As a result there is an
increased focus on families, in relation to delivering intervention to very young children with autism and in relation to
supporting family resilience and capacity. The findings of this research provide some useful information for families and
service providers of early intervention.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1565
Acknowledgements
The authors would like to acknowledge the children and their families who participated in this research and research
assistance from Susan Dodd, Alison Parsons, Emma Pierce and Rebecca Sutherland.
References
Abidin, R. R. (1995). Parenting stress index (3rd ed.): Professional manual. Odessa, FL: Psychological Assessment Resources Inc.
Aman, M. G., Mcdougle, C. J., Scahill, L., Handen, B., Arnold, E., Johnson, C., et al. (2009). Medication and parent training in children with pervasive developmental
disorders and serious behaviour problems: Results from a randomized clinical trial. Journal American Academy Child Adolescent Psychiatry, 48, 1143–1154.
American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC.
Australian Bureau of Statistics. (2007). General Social Survey: Summary Results, Australia, 2006. From February 10, 2010 http://www.abs.gov.au/AUSSTATS/
abs@.nsf/Latestproducts/4159.0Appendix42006?opendocument&tabname=Notes&prodno=4159.0&issue=2006&num=&view=.
Australian Bureau of Statistics. (2008). Census of population and housing: Socio-economic indexes for areas (SEIFA), Australia – Data only, 2006. From January 28,
2010http://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/2033.0.55.001Main+Features12006?OpenDocument.
Autism Association of NSW (2004a). Starting blocks child component training manual. Building Blocks Early Intervention Service. Autism Spectrum Australia,
Sydney NSW 2087.
Autism Association of NSW (2004b). Starting blocks parent component training manual. Building Blocks Early Intervention Service. Autism Spectrum Australia,
Sydney, NSW 2087.
Carpenter, B., & Egerton, J. (2005). Early childhood intervention; possibilities and prospects for professionals, families and children. British Journal of Special
Education, 32, 176–183.
Carter, M., Roberts, J., Williams, K., Evans, D., Parmenter, T., Silove, N., et al. (2010). Interventions used with an Australian sample of preschool children with autism
spectrum disorders. Research in Autism Spectrum Disorders , doi:10.1016/j.rasd.2010.11.009.
Dawson, G., & Osterling, J. (1997). Early intervention in autism. In M. J. Guralnick (Ed.), The effectiveness of early intervention (pp. 307–326). Baltimore, US: Brookes
Publishing.
Dewart, H., & Summers, S. (1995). Pragmatics profile of everyday communication in children. Windsor, UK: NFER-NELSON.
Edwards, S., Fletcher, P., Garman, N., Hughes, A., Letts, C., & Sinka, I. (1997). Reynell developmental language scales. III. The university of reading edition. NFER-Nelson
Publishing Company.
Einfeld, S., & Tonge, B., (2002) Developmental Behaviour Checklist: Parent-Carer (DBC-P) Centre for Developmental Psychiatry and Psychology, Monash Medical
Centre, Clayton, Victoria 3168, Australia.
Green, J., Charman, T., McConachie, H., Aldred, C., Slonims, V., Howlin, P., et al. (2010). Parent-mediated communication-focused treatment in children with
autism (PACT): a randomised controlled trial. The Lancet, 375, 2152–2160.
Howlin, P., Magiati, I., & Charman, T. (2009). Systematic review of early intensive behavioural interventions for children with autism. American Journal of
Intellectual and Developmental Disabilities, 114, 23–41.
Jolleff, N., Emmerson, F., Ryan, M., & McConachie, H. (2006). Communication skills in Angelman Syndrome: Matching phenotype to genotype. International Journal
of Speech-Language Pathology, 8, 28–33.
Keen, D., Roger, S., Doussin, K., & Braithewaite, M. (2007). A pilot study of the effects of a social pragmatic intervention on the communication and symbolic play of
children with autism. Autism, 11, 63–71.
Krebs Seida, J., Ospina, M. B., Karkhaneh, M., Hartling, L., Smith, V., & Clark, B. (2009). Systematic reviews of psychosocial interventions for autism: An umbrella
review. Developmental Medicine and Child Neurology, 51, 95–104.
Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule—Generic: A standard measure
of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30, 205–223, doi:10.1023/
A:1005592401947.
Lord, C., Rutter, M., DiLavore, P. C., & Risi, S. (1999). Autism diagnostic observation schedule. Los Angeles, CA: Western Psychological Services.
Luiz, D., Faragher, B., Barnard, A., Knoesen, N., Kotras, N., Burns, L. E. et al. (2006). Griffiths Mental Development Scales – Extended Revised: 2 to 8 years. Oxford, UK.
McConachie, H., & Diggle, T. (2007). Parent implemented early intervention for young children with autism spectrum disorder; a systematic review. Journal of
Evaluation in Clinical Practice, 13, 120–129.
National Research Council. (2001). Educating Children with Autism. Committee on Educational Interventions for Children with Autism. In Lord Catherine &
Division of Behavioral and Social Sciences and Education. Washington DC: National Academy Press.
National Standards Report (2009). National Autism Centre Massachusetts. www.nationalautismcenter.org.
Ospina, M. B., Krebs Seida, J., Clark, B., Karkhaneh, M., Hartling, L., Tjosvold, L., et al. (2008). Behavioural and developmental interventions for autism spectrum
disorder: A clinical systematic review. PLoS ONE, 3, 1–32 www.plosone.org,doi:10.1371/journal.pone.0003755.
Perneger, T. V. (1998). What’s wrong with Bonferroni adjustments. British Medical Journal, 316, 1236–1238 Retrieved from http://www.bmj.com/content/316/
7139/1236.full.
Rickards, A. L., Walstab, I., Wright-Rossi, R. A., Simpson, J., & Reddihough, D. S. (2009). One-year follow-up of the outcome of a randomized controlled trial of a
home-based intervention programme for children with autism and developmental delay and their families. Child: Care Health and Development, 35, 593–602.
Roberts, J. M., & Prior, M. (2006). A review of the research to identify the most effective models of best practice in the management of children with autism spectrum
disorders. Australian Government Department of Health and Ageing. Retrieved from http://www.health.gov.au/internet/main/publishing.nsf/Content/
846804F6D67F34F3CA257280007853DE/$File/autrev.pdf.
Sparrow, S. S., Balla, D., & Cicchetti, D. (1984). Vineland adaptive behaviour scales. Circle Pines, MN: AGS.
Sparrow, S. S., Balla, D., & Cicchetti, D. (2005). Vineland Adaptive Behaviour Scales (second ed. (Vinelamd-2)). Circle Pines, MN: AGS.
Summers, J. A., Poston, D. J., Turnbull, A. P., Marquis, J., Hoffman, L., Mannan, H., et al. (2005). Conceptualizing and measuring family quality of life. Journal of
Journal of Intellectual Disability Research, 49, 777–783.
Tonge, B., Brereton, A., Kiomall, M., Mackinon, A., King, N., & Rinehart, N. (2006). Effects on parental mental health of an education and skills training programme
for parents of young children with autism: A randomised control trial. Journal of the American Academy of Child and Adolescent Psychiatry, 45(5), 561–569.
Vismara, L. A., Colombi, C., & Rogers, S. J. (2009). Can one hour per week of therapy lead to lasting changes in young children with autism? Autism, 13, 93–115.
Williams, K., Glasson, E., Wray, J., Tuck, M., Helmer, M., Bower, C., et al. (2005). Incidence of ASD in children in two Australian states. The Medical Journal of Australia,
182, 108–111.
J. Roberts et al. / Research in Autism Spectrum Disorders 5 (2011) 1553–1566
1566
