1.- It’s going to be an oral presentation. I did it already. but what need to be corrected is the 1st slide.

A) 1st slide needs to be in bullet point and under the note section at the bottom,  please add notes about the bullet points and according to article for all 3 slides so that I could present well.

B) the other slides: results, and Individual variability are fine, but add notes because I am totally clueless of what to say when presenting.

C) I will attach the article and the power point

ISSN: 1751-8423 (print), 1751-8431 (electronic)

Dev Neurorehabil, 2016; 19(3): 162–174
! 2016 Taylor & Francis. DOI: 10.3109/17518423.2014.933983

O R I G I N A L A R T I C L E

Effect of community interventions on social-communicative abilities of
preschoolers with autism spectrum disorder

Sara Van der Paelt, Petra Warreyn, & Herbert Roeyers

Research Group Developmental Disorders, Department of Experimental Clinical and Health Psychology, Ghent University, Ghent, Belgium

Abstract

Purpose: To evaluate the effect that different intervention methods have on the social-
communicative abilities of children with autism spectrum disorder (ASD) in community
settings. Methods: Intervention based on applied behaviour analysis was compared with a more
specific intervention programme targeting imitation and joint attention (JA) and with treatment
as usual in a sample of 85 children with ASD. Objective measures for imitation, JA, pretend play,
language, autism severity and parent report measures were used to assess the effect of six
months of intervention. Results: Results revealed no differences between the intervention
methods. There was, however, great individual variability in outcome within each treatment
method. Conclusion: These results suggest that it is important to focus on ‘‘What works for
whom’’ instead of trying to find a one-size-fits-all-

treatment for children with ASD.

Keywords

Applied behaviour analysis, social
communication, treatment

History

Received 5 January 2014
Accepted 9 June 2014
Published online 16 July 2014

Introduction

A reliable diagnosis of an autism spectrum disorder (ASD) is

possible from the age of two onwards [1]. Since age at the

start of the treatment is an important predictor for the effect of

an intervention [2], it is essential to start the most beneficial

treatment soon after diagnosis. Of the available interventions,

applied behaviour analysis (ABA) is widely recognized as one

of the most well-established treatments for ASD. Many

controlled studies show gains in language, cognitive and

adaptive functioning, following intensive behavioural inter-

vention [e.g. 3–6]. A review of five recent meta-analyses on

ABA concluded that no other comprehensive treatment model

for children with ASD has the same amount of empirical

support [7]. Several other reviews also concluded ABA is the

treatment of choice for children with ASD [8, 9]. A few

critical comments should, however, be added to these positive

results. Almost all studies in favour of ABA evaluated very

intensive programmes, usually of 20–40 hours a week.

Implementing such programmes on a large scale in clinical

practice is difficult because of a lack of financial resources

and trained staff. Studies with less intensive forms of

ABA

(12–20 hours) do not seem to be as effective as the traditional

type of ABA [10]. Furthermore, most studies have looked

primarily at the effects of the intervention on cognitive and

adaptive abilities. Recently, the effect on the core symptoms

of ASD has also received some attention [11], but most of the

evidence for ABA comes from studies that did not look at the

effects on social and communicative abilities. It is also

important to note that there has been little attention to

variability in outcome. The first results on ABA from the

study by Lovaas [12] already suggested that ABA did not have

positive effects for all children with ASD. Some recent studies

tried to explain the variability in outcome by looking at

factors that predict success of the intervention. Treatment

intensity, age, severity of autism and cognitive functioning are

all related to outcome [13–16]. It is important to note that

ABA is a generic term for a group of intervention methods,

which have the use of principles of operant conditioning in

common. However, these principles can both be used in a very

structured and directive way (discrete trial teaching) as in a

naturalistic behavioural approach (also called incidental

teaching) where the interests of the child are followed.

Apart from comprehensive treatments like ABA, there is

recent attention for specific training programmes in social-

communicative abilities. Most of these interventions, which

differ in the specific abilities targeted, share the use of

behavioural techniques. However, compared to ABA, the time

frame is usually shorter, and the number of treatment goals is

more limited. Social-communicative abilities are proposed

by several researchers as important treatment goals because

they are considered to be pivotal areas of development [e.g.

17,18]. Young children with ASD already show clear deficits

in social-communicative abilities such as imitation and joint

attention (JA), and these deficits are the first to raise parental

concern [19–21]. Because these abilities are important for

further language and social-cognitive development, children

with ASD will miss chances to develop a range of abilities

from early on [22–24]. Targeting these pivotal developmental

behaviours in young children with ASD should help prevent

further developmental delays.

Correspondence: Sara Van der Paelt, Research Group Developmental
Disorders, Department of Experimental Clinical and Health Psychology,
Ghent University, Henri Dunantlaan 2, B-9000 Ghent, Belgium.
Tel: 0032(9)2649412. Fax: 0032(9)2646489. E-mail:
Sara.Vanderpaelt@ugent.be

Several researchers have shown that imitation and JA

abilities of children with ASD can improve with a specific

training [25–27]. Moreover, effects seem to generalize to

abilities that have not been targeted. For instance, it has been

shown that children with ASD exhibit gains in language after

a training in JA or symbolic play and that the effects of

reciprocal imitation training generalize to language, pretend

play and JA [17, 28]. Research also demonstrated that these

treatments may also be effective when provided at a low

intensity. A study by Goods, Ishijima, Chang, and Kasari [29]

revealed that children who did not respond to ABA after one

year of treatment, benefitted from an intervention targeting

JA, symbolic play and regulation. Thirty minutes twice a

week of their regular ABA intervention was replaced by the

social communication intervention. After 12 weeks, these

children demonstrated greater play diversity, initiated more

interactions and showed more engagement. Moreover, inter-

vention techniques targeting social-communicative abilities

can be taught to parents. Several recent studies have shown that

parents can use these techniques effectively, and that social-

communicative abilities of their children improve more than in

children receiving standard care [30, 31].

It should be noted, however, that most of these studies rely

primarily on observational methods, not on standardized

assessment of the social-communicative abilities. Effects on

standardized language and cognitive tests are somewhat

mixed, with some studies reporting greater improvement than

in standard care [28] and other studies not [30, 32].

Although ABA and social communication interventions

are broad categories of therapy methods, which can be used

in a number of ways, in general, these treatments are

promising. However, most studies took place in research

settings under controlled circumstances. It is also important

to look at the effect of interventions when they are being used

in clinical practice, because clinicians tend to adapt evidence-

based programmes to characteristics of the child or setting

[33]. This limits the external validity of randomized

controlled trials (RCT), which test the use of an intervention

under ideal circumstances, which cannot be achieved in the

real world. Results from studies that have compared com-

munity-based interventions were not always in favour of ABA

and have found similar improvements as in treatment as usual

(TAU) [16, 34].

In Belgium, where this study was conducted, the majority

of children with ASD do not have access to intensive early

intervention services. Supplementary to regular or special

education, children with ASD are entitled to a couple of hours

a week of publicly funded intervention. This is usually

provided by the multidisciplinary team of a rehabilitation

centre, serving children with developmental and learning

disabilities. A regular care system with a lack of access to

intensive intervention services is not unique to Belgium and is

therefore an important issue that has not received sufficient

attention in the international literature. The number of studies

that have looked at community interventions in children with

ASD is already low, but even less data are available on low-

intensive interventions in community settings. To our know-

ledge, no study has looked at the effect of this type of

intervention on a broad array of objective measures of social-

communicative abilities in children with ASD.

The main goal of this study was to evaluate the effect of

different methods of intervention, used in community settings,

on social-communicative and related abilities in young

children with ASD. Measures of adaptive behaviour, symptom

severity and social-emotional functioning were used in

addition to social communication measures, in order to

provide a broad picture of children’s progress. More specif-

ically, intervention based on ABA was compared with a more

specific intervention programme targeting imitation and JA

and with TAU, in a large sample of children with ASD

receiving community intervention in comparable settings. A

second aim was to map the individual variability in the

different intervention groups.

Methods

Participants

Ninety-two children were recruited from 16 publicly funded

specialized multidisciplinary treatment centres for children

with developmental disabilities. Children qualified for these

services based on a diagnosis of ASD and their need for

multidisciplinary intervention (at least by two therapists of

a different discipline). The children were diagnosed with

ASD by a multidisciplinary team based on criteria of the

Diagnostic and Statistical Manual of Mental Disorders, Fourth

Edition, Text Revision (DSM-IV-TR) [35]. Eight children

received a working diagnosis [36] because they were con-

sidered at risk for ASD due to their young age or because they

did not meet full criteria. Five treatment centres used ABA, six

used a specific training in imitation and JA (imitation/JA) and

five provided TAU. Since this was a community intervention

study, there was no random assignment. Seven children

dropped out before the end of the study, because they started

special education
1

(n¼5), because of practical issues (n¼1)
or because parents chose to end the treatment (n¼1). These
children were excluded from further analysis. The remaining

85 children, aged 22–75 months, were divided into three

groups, based on the intervention method. Their cognitive level

was assessed before the start of the study by the treatment

centre with the Dutch version of one of the following tests:

Bayley Scales of Infant Development, second edition (n¼42)
[37], Wechsler Preschool and Primary Scale of Intelligence,

third edition (WPPSI-III, n¼13) [38], WPPSI – Revised (used
because the WPPSI-III was not yet available in Dutch in all

treatment centres at the time of the assessment, n¼13) [39],
Snijders-Oomen Non-Verbal Intelligence Test – Revised

(n¼14) [40], Psychoeducational Profile – Revised (n¼2)
[41] and McCarthy Developmental Scales (n¼1) [42]. There
was no initial difference between the groups in the mean age,

F(2, 82)¼1.77, p¼0.17, severity of autism symptoms
(Autism Diagnostic Observation Schedule, ADOS) [43], F(2,

82)¼2.52, p¼0.09, amount of intervention before the start of
this study (information provided by the treatment centre), F(2,

82)¼0.30, p¼0.74 and the educational level of the parents,
F(2, 82)¼0.34, p¼0.72 for maternal years of education and
F(2, 82)¼0.44, p¼0.64 for paternal years of education. Mean
scores and standard deviations are presented in Table I.

Furthermore the distribution of gender, �2(2)¼0.49,
p¼0.78, diagnosis, �2(2)¼2.79, p¼0.25 and IQ
�

2
(8)¼13.95, p¼0.08, was also equivalent in the three

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 163

groups. The only initial difference between the groups was in

type of education, �2(6)¼24.39, p50.001. The decision about
what type of education a child with ASD will follow is usually

made based upon the child’s ability to function in a quite large

group of typically developing children. Children in special

education are believed to function better in a smaller group,

with more individual support. Most children in the

TAU

group followed regular education (with and without extra

support at school). The children in the imitation/JA group

were mainly in regular education with extra support or in

special education. The ABA group had a larger part of

children who did not go to school yet compared to the other

two groups (see Table II). In all groups, a part of the children

in regular education received 1:1 intervention at school to

help them integrate in a regular school setting. Usually, this

support is given for 100 minutes a week (in some children 50

or 150 minutes).

Interventions

ABA

Therapists were trained in ABA-techniques by a Board

Certified Assistant Behaviour Analyst before the start of the

study and also received monthly/bimonthly supervision by

this person during the period they participated in this study.

The training focussed on the verbal behaviour approach [44].

This specific type of ABA intervention is based on the theory

of Skinner [45] who specified different categories of verbal

behaviour such as mands (requesting for desired objects and

activities), tacts (naming objects), echoics (imitating lan-

guage) and intraverbals (answering questions). According to

the verbal behaviour approach, each of these categories

should be taught explicitly to children with ASD. For

example, a child who can tact certain objects will not

automatically be able to mand for those same objects and will

need explicit instruction to learn this. Therapists used the

Verbal Behaviour Milestones Assessment and Placement

Program (VB-Mapp) [46] to evaluate progress and determine

new targets. This program puts an emphasis on language,

cognitive and social communication goals (play, imitation and

social skills in groups). The verbal behaviour approach

combines both incidental teaching as well as discrete trial

teaching. Treatment fidelity was checked by the first author

through a questionnaire the therapists were asked to complete

every week, in which they described the amount of time spent

on ABA, the goals, methods and behaviour of the child during

the sessions. In addition, a video of an intervention session

was made to check if the therapists implemented the

techniques correctly
2
. Since this was a community interven-

tion study, it was expected that therapists would adapt the

intervention and combine it with other intervention methods

[33]. For this reason, only broad fidelity criteria were used to

check whether the main teaching techniques (such as

systematic use of prompts and reinforcement) of ABA were

used. All therapists achieved sufficient levels of treatment

fidelity. From the written descriptions and the analysis of the

video, we could conclude that therapists used mainly

incidental teaching techniques: they followed the child’s

lead and used activities/toys the child was motivated for, to

prompt for more complex behaviours. Therapists used several

strategies to provoke communication in the child (e.g. putting

objects in the visual field of the child, but out of their reach).

When the child initiated communication, therapists used

prompts and prompt fading to increase the complexity of the

child’s communicative attempt. In all children, manding was

targeted; and in the majority of the children, this remained the

main goal of the ABA intervention throughout the study.

Other goals were determined individually and included

imitation (in 65% of the children), play (50%), listener

responding (following instructions; in 25% of the children),

tacting (in 20% of the children), social skills with peers (in

10% of the children) and intraverbals (in 5% of the children).

These skills were taught trough differential reinforcement,

prompt and prompt fading strategies, reinforcer assessment,

following the child’s lead and imitating the child. In some

children (15%), discrete trial training was used in addition to

incidental strategies. In 25% of the children, parents attended

the ABA sessions on a regular basis and implemented some of

Table II. Distribution of gender, IQ, diagnosis and school placement in
the different intervention groups.

Characteristic

ABA Imitation/JA TAU

Gender
Male 15 (75%) 24 (80%) 29 (83%)
Female 5 (25%) 6 (20%) 6 (17%)

IQ
555 11 (55%) 12 (40%) 4 (11%)

55–70 3 (15%) 7 (23%) 10 (29%)
71–85 3 (15%) 5 (17%) 11 (31%)

86–115 3 (15%) 6 (20%) 9 (26%)
4115 0 (0%) 0 (0%) 1 (3%)

Diagnosis
Clinical diagnosis 20 (100%) 26 (87%) 31 (91%)
Working diagnosis 0 (0%) 4 (13%) 4 (9%)

Education type
Regular education 5 (25%) 4 (13%) 14 (40%)
Regular education with support 3 (15%) 13 (43%) 14 (40%)
Special education 4 (20%) 11 (37%) 1 (3%)
Not in school yet 8 (40%) 2 (7%) 6 (17%)

Percentages are within each treatment group.
ABA¼applied behaviour analysis; JA¼ joint attention; and

TAU¼ treatment as usual.

Table I. Child’s characteristics and parental education data in the
different intervention groups.

Characteristic
ABA

(n¼20)
Imitation/JA

(n¼30)
TAU

(n¼35)

Age
a

M (SD) 44.47 (16.28) 51.79 (11.29) 49.09 (13.49)
Range 21.64–69.40 30.36–71.92 24.41–70.73

Autism severity
M (SD) 6.55 (2.21) 5.97 (2.44) 5.23 (1.85)
Range 1–10 1–10 1–8

Time elapsed since start of the intervention
a

M (SD) 8.12 (11.04) 8.89 (7.06) 7.31 (7.36)
Range 0.20–41.25 1.15–23.26 0.30–36.38

Maternal years of education
M (SD) 13.30 (2.39) 13.00 (2.23) 13.43 (1.87)
Range 6–17 6-17 12–17

Paternal years of education
M (SD) 12.45 (2.01) 13.03 (1.81) 12.91 (2.26)
Range 6–15 12–17 6–17

ABA¼applied behaviour analysis; JA¼ joint attention; and
TAU¼ treatment as usual.

a
In months.

164 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

the techniques at home. In addition to the ABA intervention,

children in this group received TAU targeting mainly

cognitive, language, fine, gross motor and visual-spatial

abilities and daily living skills.

Imitation/JA intervention

The treatment centres in this group had previously taken part in

a study to test the effect of the imitation/JA-intervention [27]

and had incorporated the principles of this intervention in their

regular care. They all received a training manual with 24

sessions describing specific activities to stimulate imitation (of

actions with objects, gestures, body movements and symbolic

imitation), following and initiating imperative and declarative

JA. The training manual briefly described how to use different

levels of prompting to increase the imitation and JA ability of

the children (see [27] for a more detailed description of the

intervention). It is important to note that although this

intervention method had some similarities with the ABA

intervention, such as the use of prompts and rewards, it was

also clearly different in a number of ways. First, this

intervention used prescribed activities to stimulate imitation

and JA, while the ABA intervention used mainly activities

for which the child was motivated to teach new skills. Second,

the ABA intervention relied heavily on the prompt-behaviour-

reward sequence, which was repeated over and over again,

within a session as well as in several subsequent sessions, until

a child could perform a specific skill spontaneously. Prompting

had a far less important place in the imitation/JA programme,

which focussed more on spontaneous behaviour of the children

elicited by the specific situation. Treatment fidelity in the

imitation/JA group was also checked through a file the

therapists were asked to complete every week, in which they

described the amount of time spent on imitation and JA, the

specific goals, methods, whether they used sessions from the

manual and the behaviour of the child during the sessions. The

amount of time spent on these goals depended on the individual

needs of the children. In most children, around 30 minutes a

week was dedicated to direct training of these abilities, but

imitation and JA were also stimulated more indirectly during

other activities (e.g. teaching the child to ask for a drink during

snack time). Because the aim of this study was to look at the

effect of this training as it was used in practice, the therapists

could use the training as they would normally do. In the

majority of the children (63%), therapists used a combination

of sessions described in the manual and variations on those

exercises with other materials. In the other children, the same

teaching principles were used, but not with the exercises

described in the manual. For example, the manual describes

several exercises in which the therapist sabotages an activity

for the child (e.g. giving the child a broken crayon) and waits

for the child to ask for help. Furthermore, the therapists that did

not follow the manual used this principle (e.g. giving the child

an empty cup during snack time). Both describing using

specific sessions from the manual and describing similar

techniques as in the manual were considered sufficient to

achieve treatment fidelity (given the community focus of the

study). In addition to imitation and JA treatment targeted

cognitive, language, fine, gross motor and visual-spatial
abilities and daily living skills.
TAU

For each child, therapists completed a file, describing the

goals and methods of the intervention. The majority of the

therapy centres of this group used one or more methods

designed primarily for children with ASD. They mainly

included strategies from the TEACCH model (in 34% of the

children), social skills training with social scripts and role

play (17%) and Hanen (11%). In 34% of the children, no

ASD-specific method was used. In the last three months

before the post-test, the therapists of one child received a

training in ABA and started to use some ABA-techniques

with this child. A specific training for imitation and JA was

not used with any of the children. This implicated that the

intervention methods in TAU were clearly different from the

methods in the other two groups. Cognitive, language, fine,

gross motor and visual-spatial abilities and daily living skills

were targeted in most of the children. Social-communicative

abilities were also mentioned as a treatment goal in all

children. In some children, therapists targeted play (43%),

imitation (20%) and/or JA (11%).

Comparison of intervention techniques

Table III provides an overview of the intervention techniques

and programmes that were used by the therapists. It is

apparent from the table that in TAU the therapists mentioned

less-specific intervention techniques than in the other two

groups. Moreover, following the child’s lead and prompt

fading are techniques, which were used especially in the ABA

group, but rarely in the other two groups.

Treatment intensity

See Table IV for an overview of the average distribution of the

intervention time in the different groups. The total amount of

time spent on 1:1 intervention (in the treatment centre and for

Table III. Number of children in each group with whom certain
intervention techniques/programmes were used.

ABA
(n¼20)

Imitation/JA
(n¼30)

TAU
(n¼35)

TEACCH (visualisation, structured
tasks)

5 (25%) 2 (7%) 12 (34%)

Hanen 0 (0%) 2 (7%) 4 (11%)
Social skills training with modelling

and role-play
0 (0%) 2 (7%) 6 (17%)

Following the child’s lead 20 (100%) 3 (10%) 3 (9%)
Communicative temptations (e.g.

sabotage and placing objects in
sight and out of reach)

20 (100%) 26 (87%) 0 (0%)

Prompt fading 20 (100%) 3 (10%) 0 (0%)
Social scripts 0 (0%) 2 (7%) 0 (0%)
Imitating the child 4 (20%) 1 (3%) 0 (0%)
Discrete trial training 3 (15%) 0 (0%) 0 (0%)
PECS 1 (5%) 0 (0%) 0 (0%)
Shadowing (prompting social skills

with peers)
3 (15%) 0 (0%) 0 (0%)

Including parents in intervention
sessions

5 (25%) 0 (0%) 0 (0%)

Percentages are within each treatment group.
ABA¼applied behaviour analysis; JA¼ joint attention;

TAU¼ treatment as usual; PECS¼picture exchange communication
system.

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 165

some children also at school) was the same in all three groups,

F(2,82)¼2.17, p¼0.12. However, the total amount of time in
group intervention was not the same in the different groups,

F(2, 82)¼6.10, p¼0.003. Games-Howell post hoc procedure
(used because of inequality of variances and different sample

sizes) showed that children in the ABA group had signifi-

cantly less group intervention than children in the imitation/

JA group, 95% CI of difference [2.84–111.33], p¼0.04 or
TAU group, 95% CI of difference [52.17–141.90], p50.001.
There was no difference between the imitation/JA and TAU

group, p¼0.34.

Outcome measures

The severity index of the ADOS [43] was used to assess

severity of autism symptoms [48].

The Preschool Imitation and Praxis Scale (PIPS) [49] was

used to measure motor imitation. The PIPS consists of 30

items, of which 21 items measure bodily imitation (gestural

and facial imitation) and 9 procedural imitation. The bodily

imitation scale comprises meaningful (e.g. wave good-bye)

and non-meaningful (e.g. place one fist on top of the other)

actions. The procedural scale encompasses goal directed (e.g.

raise a toy bear by pulling a cord) and non-goal directed (e.g.

open a box, turn it upside down and put a block on the bottom

of the box) actions.

The structured version of the Test of Pretend play (ToPP)

[50] was used to asses three main types of pretend play: object

substitution, property attribution and reference to an absent

object. The test assesses the child’s ability to use him/herself

as the object of pretend play as well as the ability to use a doll

or teddy bear as agent. Moreover, the ability to combine play

acts into a script is tested. A nonverbal version, in which

actions were modelled, was used in children with a language

comprehension level of less than three years old. In children

with a better language comprehension, we used the verbal

version, in which besides modelled actions, also verbal

instructions were used. Every item consists of a part were

the child can produce original play and a part were the child is

asked to copy a modelled action (e.g. using an ambiguous

object as a hat for a doll) or to follow an instruction (‘‘show

me the bear is sad’’).

The abridged version of the Early Social Communication

Scales (ESCS) [51] was used to measure initiation of JA

(IJA), initiating behaviour request (IBR) and response to JA

(RJA). Four different mechanical toys (three wind-up toys

and a pop-up puppet) were activated in sight of the

children. The experimenter gave each toy to the child when

he or she requested it. The child could play with the toy for

30 seconds, after which the experimenter requested the toy

back and activated it again. This procedure was repeated

with each toy three times. Two of the toys were first placed

in a box that the child could not open by himself and were

given to the child in the box in order to elicit requesting to

open it. In order to assess RJA, four pictures (A4 size) of

Winnie the Pooh and friends were placed on the walls right

and left of the child, two in their visual field (at

approximately 60 degrees from the child’s midline) and

two behind the child (at approximately 150 degrees form

the child’s midline). After gaining the child’s attention, the

experimenter gazed at each of the four posters and said the

name of the child three times before looking back to the

child. If the child did not follow the gaze of the

experimenter to the first two posters, a pointing gesture

was added for the last two posters. Children received a

score from 0 to 4, depending on the number of posters they

followed the gaze and/or point to.

The coding of the ESCS was done with the Observer XT,

version 9.0, Wageningen, The Netherlands [52] by four

independent coders. Scores for IJA and IBR were based on

frequency counts of nonverbal and verbal communication

during the whole observation. Verbal communication was

included because we tested children up to six years old in our

sample. It can be expected that the older children become, the

more they will use language as a means for sharing attention.

Yoder et al. [53] also used the ESCS to count the frequency of

nonverbal and verbal JA, (called unweighted triadic commu-

nication) and showed that the frequency of the unweighted

triadic communication remained stable in siblings of children

with ASD between 15 and 34 months. Because this is a period

in which children become more verbal, these results suggest

that the amount of nonverbal JA decreased in that same

period. It seemed thus important for this study to use a

combined measure of nonverbal and verbal JA to be able to

draw conclusions on the effect of the interventions on JA.

The following nonverbal IJA behaviours were observed:

(1) making eye contact with the experimenter to share

interest, (2) alternating eye contact between an active/moving

toy and the experimenter, (3) proximal or distal pointing with

or without eye contact to share interest and (4) showing an

Table IV. Average distribution of the treatment time and total weekly
treatment duration in the different intervention groups.

Intervention type ABA Imitation/JA TAU

ABA-intervention

Individual 42% (27%) 0% 0%

Group 0% 0% 0%

Imitation/JA training

Individual 0% 14% (8%) 0%

Group 0% 1% (3%) 0%

Other ASD specific

Individual 1% (4%) 2% (6%) 7% (13%)

Group 0% 1% (4%) 15% (26%)

General speech-language therapy

Individual 12% (18%) 16% (12%) 20% (19%)

Group 0% 7% (10%) 5% (7%)

Occupational training

Individual 20% (17%) 19% (11%) 17% (12%)

Group 0% 4% (9%) 6% (9%)

Physiotherapy

Individual 16% (15%) 15% (12%) 12% (11%)

Group 2% (6%) 6% (9%) 8% (12%)

School intervention

Individual 6% (15%) 15% (19%) 12% (16%)

Group 0% 0% 0%

Total intervention time

Individual 178.50 (49.77) 196.67 (77.17) 162.00 (65.70)

Group 5.25 (12.82) 62.33 (119.49) 102.29 (107.21)

Total 183.75 (51.55) 259.00 (101.95) 264.29 (117.61)

Percentages are averages and standard deviations (between brackets)
within each treatment group. Total intervention time is presented in
minutes a week.

ABA¼applied behaviour analysis; JA¼ joint attention; and
TAU¼ treatment as usual.

166 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

object to the experimenter with eye contact. Verbal IJA was

defined as using one or more words to share interest with the

experimenter. The following nonverbal IBR behaviours were

coded: (1) making eye contact with the experimenter to

request something, (2) reaching for a toy, with and without

eye contact, (3) proximal or distal pointing with and without

eye contact to request and (4) giving an object to the

experimenter. Verbal IBR was defined as using one or more

words to request something. Nonverbal and verbal scores for

IJA and IBR were combined in a total IJA score and a total

IBR score. Interrater reliability was determined with single

measures intraclass correlations (ICCs) by double coding of

25% of the observations. The ICCs were 0.94 for nonverbal

IJA, 0.96 for verbal IJA, 0.87 for nonverbal IBR, 0.91 for

verbal IBR and 0.84 for RJA.

The Reynell Developmental Language Scales – Dutch

version (RTOS) [54] was used to assess expressive and

receptive language. Normative data, based on a sample of

Dutch speaking children, were available.

Parents were asked to fill out a screener version of the

Vineland Adaptive Behavior Scales [55]. This questionnaire

was used to measure adaptive behaviour. This questionnaire

has good psychometric properties: internal consistency

Cronbach’s alpha of 0.90, inter-observer agreement between

mothers and fathers ICC of 0.97, test–retest reliability ICC of

0.99 and adequate contents-, construct- and criterion

validity.

The Dutch version [56] of the Social Communication

Questionnaire [57] was used as a parent report measure of

symptom severity at home. The psychometric properties of

this instrument are good: internal consistency Cronbach’s

alpha of 0.90 and adequate construct validity.

The total problem score on the Dutch version [58] of the

Child Behaviour Checklist 1
1/2

–5 years, [59] was used as a

measure of social-emotional problems at home. This measure

has good psychometric properties: test–retest reliability r of

0.90, inter-observer agreement between mothers and fathers r

of 0.65 and adequate contents-, construct- and criterion

validity.

Procedure

Pre-and post-tests were administered in the treatment centres

of the children, both on two separate days, with approximately

one week in between. The first assessment started with the

ADOS, after which the PIPS was administered. The second

assessment consisted of the ESCS, ToPP and RTOS, in this

order. Both assessments took approximately 60–90 minutes.

Time between pre- and post assessment was six months. The

assessment was videotaped, and all the tests were scored

afterwards from the video.

Parents received the questionnaires from the therapy

centre. They were asked to fill them out at home and hand

them in afterwards in the therapy centre. Fifty-eight percent

of the parents filled out and returned the questionnaires both

at the pre- and post-test. The study design was prospectively

reviewed and approved by the Ethics Committee of

the Faculty of Psychology and Educational Sciences of

Ghent University, where the study was conducted. Parents

gave their written consent prior to the inclusion of their

children in the study.

Results

Comparison of treatment effect between the
intervention groups

We performed two repeated measures MANOVAs with time

(pre- vs. post-test) as a within group independent variable and

intervention group as a between group independent variable.

Imitation, pretend play, IJA, IBR, RJA, receptive language,

expressive language and ADOS symptom severity were

entered as dependent variables in the first analysis. The

results from parent questionnaires (adaptive behaviour, ASD

symptoms and social-emotional problems) were entered in a

separate analysis because these data were not available for all

children. Raw scores were used for imitation, pretend play,

language and adaptive behaviour because some children had

bottom scores on age equivalent scores (AE). Furthermore,

for social-emotional problems, we used the raw scores,

because some children were older than the upper age limit

of five years. Descriptive statistics for pre- and post meas-

urement in the different intervention groups of all dependent

variables are presented in Table V.

The repeated measures MANOVA with the test results

as dependent measures revealed no interaction between

intervention group and time, which means that the progress

children made, was not different in the three intervention

groups, F(16, 152)¼0.59, p¼0.89. There was also no
main effect of intervention group, which shows the groups

did not have significantly different scores on these meas-

ures, irrespective of the progress they had made during

the intervention phase, F(16, 152)¼1.41, p¼0.14.
However, there was a main effect of time, indicating that

irrespective of the intervention group, post scores differed

significantly from pre scores, F(8, 75)¼14.69, p50.001,
��2¼0.61. Univariate analyses showed a significant increase
in imitation, F(1, 82)¼20.59, p50.001, ��2¼0.20, pretend
play, F(1, 82)¼38.86, p50.001, ��2¼0.32, receptive lan-
guage, F(1, 82)¼99.70, p50.001, ��2¼0.55, expressive
language, F(1, 82)¼47.68, p50.001, ��2¼0.37 and IJA,
F(1, 82)¼4.01, p¼0.048, ��2¼0.05 and a significant
decrease in ADOS severity score, F(1, 82)¼5.32, p¼0.02,
��2¼0.06. There was no significant difference between pre-
and post-test in IBR, F(1, 82)¼0.22, p¼0.64, or RJA,
F(1, 82)¼0.50, p¼0.48.

The repeated measures MANOVA with the parent report

variables also showed no difference in the progress of the three

intervention groups, F(6, 90)¼0.47, p¼0.83 and no differ-
ence between the groups, irrespective of time, F(6, 90)¼1.00,
p¼0.43. Again, there was a significant main effect of time,
F(3, 44)¼6.18, p¼0.001, ��2¼0.30. Univariate tests
revealed a significant increase in adaptive behaviour, F(1,

46)¼16.48, p50.001, ��2¼0.26. There was no difference
between pre- and post-test in social-emotional problems,

F(1,46)¼0.04, p¼0.84 or ASD symptoms, F(1, 46)¼3.87,
p¼0.06.

Individual variability

For imitation, pretend play, language abilities and adaptive

behaviour (parent report), AE scores were available (the

manuals of these instruments contained information to

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 167

convert the raw scores in AE scores), making it possible to

compare the progress children had made with a normal

developmental path. Difference scores of the pre- and post AE

scores were computed for each of these abilities. For each

intervention method, we categorized each child in one of four

groups, based on the progress they had made. A distinction

was made between children who declined, made limited

progress (0–2 months progress in AE in a six-month-period),

moderate progress (3–5 months progress in AE in a six-

month-period) or followed a normal developmental path/

progressed faster than the average normal developmental path

(�6 months progress in AE in a six-month-period). Figure 1
shows the percentage of children in each group for imitation

(a), pretend play (b), receptive language (c), expressive

language (d) and adaptive behaviour (e). The graphs show that

for all abilities measured and in all the intervention groups,

there was clear variability in outcome. It is apparent that for

each of the abilities and in each intervention group, a

substantial number of children either declined (between 5%

and 35% depending on the specific ability and intervention

method) or made limited progress (between 7% and 70%

depending on the specific ability and intervention method),

but also a substantial number progressed at an age-equivalent

developmental rate or even faster (between 20% and 63%

depending on the specific ability and intervention method).

For pretend play, receptive and expressive language only a

minority progressed at a moderate rate (most percentages

between 6% and 14%, except for receptive language in the

TAU group 34%). For imitation and adaptive behaviour, this

group was larger, but still consisted of less than one-third of

the children (between 18% and 33% depending on the specific

ability and intervention method). The distribution of the

amount of progress was not different in the three intervention

groups for imitation, �2(6)¼6.99, p¼0.32, pretend play,
�2(6)¼6.55, p¼0.37, expressive language, �2(6)¼5.76,
p¼0.45 and adaptive behaviour, �2(6)¼2.96, p¼0.81.
There was however a difference between the groups in the

distribution of the progress in receptive language,

�
2
(6)¼15.82, p¼0.02. Children in the TAU group were

equally distributed in minimal, moderate and age-equivalent

progress, whereas the majority (65%) of the children in the

ABA group showed minimal progress and most children in

the imitation/JA group either showed minimal (33%) or age-

equivalent (50%) progress.

Besides a categorisation for each outcome measure

separately, it seemed relevant to make a global categorisation

to describe the progress children made in general. The

categorisation was done post-hoc and was based on the

Table V. Descriptive statistics of the dependent measures at pre- and posttest in the different intervention groups.

ABA Imitation/JA TAU

Measure Pre Post Pre Post Pre Post

Imitation
a

M (SD) 16.30 (21.10) 23.00 (22.57) 34.80 (24.10) 40.37 (21.91) 39.97 (22.31) 44.69 (19.07)
Range 0–68 0–66 0–67 0–69 0–78 0–73

Pretend play
a

M (SD) 5.35 (7.49) 8.80 (9.13) 10.33 (8.62) 14.00 (9.25) 13.37 (9.36) 17.17 (9.06)
Range 0–25 0–27 0–30 0–29 0–33 0–33

Receptive language
a

M (SD) 12.75 (18.42) 20.30 (20.79) 29.27 (21.02) 36.80 (20.37) 36.03 (22.30) 43.31 (19.64)
Range 0–58 0–64 0–69 0–71 0–69 2–71

Expressive language
a

M (SD) 12.05 (20.68) 17.10 (22.94) 27.03 (23.78) 35.97 (22.76) 34.91 (25.45) 43.37 (22.30)
Range 0–69 0–65 0–67 0–74 0–75 0–76

ADOS symptom severity
M (SD) 6.55 (2.21) 5.70 (1.92) 5.97 (2.44) 5.43 (2.81) 5.23 (1.85) 4.97 (2.18)
Range 1–10 1–9 1–10 1–10 1–8 1–9

IJA
b

M (SD) 1.81 (2.40) 1.95 (2.02) 2.69 (1.34) 3.26 (2.11) 2.98 (1.55) 3.39 (1.96)
Range 0.00–9.67 0.00–6.56 0.08–5.33 0.00–7.51 0.24–6.94 0.40–7.56

IBR
c

M (SD) 1.88 (1.40) 2.18 (1.70) 2.10 (1.33) 1.98 (1.37) 2.25 (1.22) 2.27 (1.11)
Range 0.00–5.66 0.24–6.72 0.45–5.51 0.28–5.71 0.12–5.73 0.22–5.14

RJA
d

M (SD) 0.46 (0.39) 0.43 (0.33) 0.52 (0.29) 0.62 (0.29) 0.60 (0.30) 0.61 (0.30)
Range 0–1 0–1 0–1 0–1 0–1 0–1

Adaptive behaviour (PR)
a,e

M (SD) 51.00 (37.31) 56.17 (34.58) 61.24 (28.09) 71.53 (34.76) 68.25 (30.91) 74.80 (27.28)
Range 8–112 6–118 12–103 6–120 13–118 17–122

ASD symptoms (PR)
a,e

M (SD) 18.33 (7.79) 18.25 (7.11) 15.94 (6.39) 14.06 (7.27) 16.65 (6.12) 15.05 (5.46)
Range 6–27 7–28 5–26 1–28 6–30 6–29

Social–emotional problems (PR)
a,e

M (SD) 58.58 (29.49) 58.17 (31.21) 52.00 (21.38) 49.82 (22.36) 62.40 (27.31) 63.25 (27.28)
Range 15–111 16–114 15–86 9–80 26–121 19–112

a
Raw score.

b
Rate per minute of initiating joint attention.

c
Rate per minute of initiating behavioural request.

d
Proportion of responding to joint attention.

e
Parent report.

168 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

difference scores of the AE scores for imitation, pretend play,

receptive and expressive language. The AE score for adaptive

behaviour was not used in this study because these data were

missing for part of the children. The resulting (mutually

exclusive) categories were as follows:

(1) Good progress: These children followed for at least three

of the four outcome measures a normal developmental

path or faster (� 6 months progress in AE in a six-month-
period).

(2) Moderate to good progress: These were children who

made moderate progress (3–5 months progress in AE) on

at least three outcome measures or made moderate

progress on two outcome measures and followed a

normal developmental path or faster on the other two

outcome

measures.

(3) Mixed: These children followed a normal developmental

path or faster on one or two outcome measures, but made

limited (not more than two months progress in AE) or no

progress or even deteriorated on at least one other

outcome measure.

(4) Limited to moderate progress: These children made

moderate progress (3–5 months progress in AE) on one

or two outcome measures and limited progress (not more

than two months progress in AE), no progress or

deteriorated on the other outcome measures.

(5) Limited progress: These children made either limited

progress (not more than two months progress in AE), no

progress or deteriorated on each of the four outcome

measures.

Figure 2 shows the percentage of children in each outcome

category for the three intervention groups.

The graph shows that the largest part of the children either

made good progress (26% of the total group), limited progress

(19% of the total group) or showed a mixed profile (34% of

the total group). Few children made predominately moderate

progress (with 7% of the total group in the moderate to good

and 14% in the limited to moderate categories). The

distribution of the amount of progress was not different in

the three intervention groups, �2(8)¼11.90, p¼0.16.

Predicting outcome

Because we found no difference in the average scores for the

different intervention groups, we additionally performed a

multinomial logistic regression analysis to explore possible

factors contributing to the individual variability in the global

100%90%80%70%60%50%40%30%20%10%0%

100%90%80%70%60%50%40%30%20%10%0%
100%90%80%70%60%50%40%30%20%10%0%
100%90%80%70%60%50%40%30%20%10%0%
100%90%80%70%60%50%40%30%20%10%0%

Decline

0-2 months progress

3-5 months progress

≥ 6 months progress

Decline
0-2 months progress
3-5 months progress
≥ 6 months progress
Decline
0-2 months progress
3-5 months progress
≥ 6 months progress
Decline
0-2 months progress
3-5 months progress
≥ 6 months progress
Decline
0-2 months progress
3-5 months progress
≥ 6 months progress

ABA (a) (b)

(c)

(e)

(d)

Imitation/JA TAU

ABA Imitation/JA TAU

ABA Imitation/JA TAU ABA Imitation/JA TAU

ABA Imitation/JA TAU

Figure 1. Percentage of children in each of the different outcome categories in the applied behaviour analysis (ABA), Imitation/Joint attention
(Imitation/JA) and Treatment as usual (TAU) intervention groups for imitation (a), pretend play (b), receptive language (c), expressive language (d) and
adaptive behaviour (e).

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Limited progress

Limited-moderate progress

Mixed

Moderate-good progress

ABA Imitation/JA TAU

Good progress

Figure 2. Percentage of children in each of the different global outcome
categories in the applied behaviour analysis (ABA), Imitation/Joint
attention (Imitation/JA) and Treatment as usual (TAU) intervention
groups.

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 169

outcome categorisation. We chose a model with the global

progress categories over several multiple regression analyses

with the different outcome measures as dependent variables to

avoid making type-I errors due to multiple testing and

because predicting the global categorisation seemed more

clinically relevant. Age, IQ-category, autism severity and

treatment intensity (total number of hours of intervention a

week, number of hours of 1:1 intervention and number of

months in intervention before the start of the study) were

chosen as predictors because previous research pointed to

these variables as predictors of the outcome of an interven-

tion. Moreover, since there was a significant difference in the

distribution of the education types in the three intervention

groups, we also assessed whether children in different types of

education (regular education, regular education with support,

special education and not in school yet) progressed in

different ways. Intervention group and the interactions

between each of the continuous predictors and intervention

group were added to test for the possible moderating effect of

the intervention group (different effect of predictors on

outcome for children who received a different intervention).

In order to comply with the assumptions of logistic regres-

sion, two adaptations were made. First, we merged the

moderate to good and good progress categories and also the

limited to moderate and limited progress categories. This was

done to assure that there would be enough observations in

each cell of the IQ-category�progress-category-matrix,
education-category�progress-category-matrix and interven-
tion group�progress-category-matrix. Second, age was
omitted as predictor because this predictor violated the

linearity-assumption (there was no linear relationship between

age and the logit of the outcome variable). IQ, intervention-

group, autism severity, total number of hours of intervention,

number of hours of 1:1 intervention and amount of interven-

tion before the study started were forced into the model.

Education type and the interaction between the intensity

variables and autism severity with the intervention group were

entered in a second step with the forward entry method (only

variables that significantly improve the predictive ability of

the model are added).

The multinomial regression analysis showed that the only

significant predictors in the final model were IQ,

�2(6)¼21.04, p¼0.002 and education type, �2(6)¼17.02,
p¼0.01. Parameter estimates of the final model showed that
children with an IQ below 55 were less likely to belong to the

mixed progress group, than the least progress group, in

comparison to children with a normal IQ, �2(1)¼4.57,
p¼0.03. The odds ratio was 0.09 (with a 95% CI of [0.01–
0.82]. Moreover, children with an IQ below 55 were also less

likely to belong to the best progress group, than the least

progress group, in comparison to children with a normal IQ,

�
2
(1)¼3.75, p¼0.05. The odds ratio was 0.10, with a 95%

CI of [0.01–1.03]. There were no other significant IQ effects.

Furthermore, parameter estimates for education type, showed

that children in special education were more likely to belong

to the mixed progress group than the least progress group, in

comparison with children that did not go to school,

�2(1)¼4.30, p¼0.04. The odds ratio was 13.97, with a
95% CI of [1.15–196.16]. There were no other significant

education type effects. There was no significant contribution

to the model of total number of hours of intervention,

�2(2)¼3.89, p¼0.14, number of hours of 1:1 interven-
tion, �2(2)¼0.68, p¼0.71, number of months in interven-
tion, �

2
(2)¼0.52, p¼0.77, autism severity, �2(2)¼0.92,

p¼0.63 or intervention group, �2(4)¼3.37, p¼0.50.

Discussion

The first goal of this study was to compare the effect of

different methods of intervention, used in community settings,

on social-communicative abilities in young children with

ASD. Our results reveal that on average, there is no difference

between the intervention methods in the amount of progress

children make with respect to their social-communicative

abilities in a six-month-period. This is in contradiction with

several previous studies that have found that ABA [3, 10, 11,

60] or interventions targeting imitation or JA [27, 30, 31]

were more effective in stimulating these abilities than TAU.

There are several possible explanations for the lack of

difference between the interventions. First, we studied the

interventions in a community setting, where it is more

difficult to clearly distinguish one method from another than

when using a RCT. Clinicians tend to adapt evidence-based

interventions and often use combinations of interventions,

with a varied level of scientific support [33]. Indeed, also in

this research, few therapists in the imitation/JA group exactly

followed the intervention manual they were given. Instead,

most therapists only used part of the sessions described in the

manual, supplemented with variations to the sessions with

other materials. Moreover, the goals targeted in the three

intervention groups were partly similar, with social-commu-

nicative abilities being targeted in almost all children. This

means that despite the clear differences between the inter-

vention methods, they also had similarities which could have

been responsible for the similar effect observed with all three

methods. Our results are in line with previous studies that also

found no difference between ABA and TAU in community

settings [16, 34].

A second factor, that could explain the lack of difference

found in this study, is the low treatment intensity. Previous

studies [13, 15] have shown that the treatment intensity is

associated with the amount of progress children make.

Moreover, Eldevik et al. [10] found that less intensive ABA

(12–20 hours a week) is not as effective as the traditional 40

hours a week of ABA-intervention. In this study, children in

the ABA group only received one hour of ABA-intervention a

week on average (maximum 4.5 hours a week). Since ABA is

a method that relies heavily on the repetition of the prompt-

behaviour-reward-sequence, it is plausible that one hour per

week does not give a child with ASD an adequate amount of

opportunities to practice the skills that are being taught.

Moreover, imitation and JA were on average targeted for only

30 minutes a week in the imitation/JA group. Although

previous studies have shown that also low-intensive interven-

tions can lead to a bigger improvement in social-communi-

cative abilities compared to TAU [27, 29], it is plausible that a

more intensive training of these abilities would be more likely

to reveal differences with TAU.

Third, if children in the TAU group make on average the

same amount of progress than children receiving ABA and

170 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

imitation/JA interventions, this could point to a good quality

of regular care for children with ASD in Flanders (the Dutch

speaking part of Belgium, where the study took place).

Support for this claim can be found in the significant progress

that children in the TAU group have made on most outcome

measures, which contrasts with several previous studies that

found no progress or even a deterioration in TAU [3, 10, 30,

31, 60]. However, we cannot be sure that the participating

centres of the TAU group are representative of regular care in

Flanders. There could be a selection bias, with centres

providing more than average quality of care to children with

ASD, being more willing to participate in the study. Either

way, since children in the TAU group made substantial

progress, it was more difficult to find a significant difference

with the imitation/JA and ABA interventions.

Even though the intervention was less intensive than

interventions in a lot of earlier reported studies, we found

significant progress on almost all dependent measures,

irrespective of the method. Especially, the decrease in the

ADOS severity score is remarkable, since several previous

studies with more intensive interventions, failed to show a

decrease in symptoms of ASD on the ADOS [16, 32]. The

majority of the studies on early intervention use cognitive and

adaptive behaviour and sometimes also language as outcome

measures. This study is one of the first to show significant

progress with a low-intensive intervention on a comprehen-

sive standardized assessment battery of social-communicative

abilities in a large sample of preschoolers with ASD.

A second goal of the study was to assess the individual

variability in outcome. We found clear individual variability

for all outcome measures, for which AE scores were available.

There was no difference between the intervention groups in

the distribution of the amount of progress for four out of five

outcome measures. Only the distribution of receptive lan-

guage differed. Half of the children in the imitation/JA group

progressed at a rate equivalent to normal development or

faster. This is in line with studies showing a collateral effect

of training imitation and JA on language [18, 28]. However,

also 40% of the children in this group made very limited

progress or declined on receptive language, which means that

the training did not have the same effect on all children. In the

ABA-group, 65% made no progress or limited progress on

receptive language, while only 25% progressed at a normal

rate. Although on average, there was no significant difference

on any of the outcome measures between the groups at the

start of the study, the ABA-group had the largest part of

children who had no spoken language at the pre-test and

showed no or limited receptive language skills. Before

receptive language could be stimulated in these children,

certain other skills (for example basic interaction skills and

JA) may have had to improve first. This could account for the

large part of children of the ABA-group making limited

progress on receptive language.

Furthermore, the general categorisation of children, taking

into account progress on imitation, pretend play and language,

shows clear individual variability. The large individual

variability could imply that there are good and poor

responders to each of the intervention methods. Children

who manage to make the same amount of progress or

even more than typically developing children on

social-communicative abilities are children who seem to

respond well to the intervention. In general, 26% of the

children made good progress on most outcome measures.

Thirty-four percent showed a mixed profile, with good

progress on some but very little or no progress on other

outcome measures. Possibly, the limited intensity of the

intervention implies that therapists can only target some of the

social-communicative abilities at once in the six-month-

period, which could explain the discrepancy in the progress in

this group of children. However, there is also a substantial

part of the children (19%) that made little or no progress or

even deteriorated on imitation, pretend play and language. It

seems thus important to gain more information on the

characteristics of these good and poor responders. Possibly,

these are different for several intervention methods, making it

crucial to focus future research on the question ‘‘What works

for whom?’’ instead of trying to find a one-size-fits-all-

treatment for children with ASD.

Few studies report on the individual variability of outcome

in community interventions, making it difficult to compare

our results. An exception is the study by Perry et al. [14], who

divided 332 children in several categories based on their

outcome after completion of an intensive ABA intervention in

community settings. Although this study used different

outcome measures (autism severity, cognitive and adaptive

functioning) and the intervention was much more intensive,

the results are quite similar to ours. They found that 25% of

their sample made clear improvement (had typical rates of

development or better and a decrease in autism severity), 41%

moderate improvement (rate of development between 0.50

and average or significant improvement in autism severity)

and 33% minimal improvement or deterioration (rate of

development of less than 0.50 and no improvement in autism

severity).

Remarkably, only a minority of the children in our study

belonged to the category that made predominantly moderate

progress. This has important implications, because a lot of the

research on early intervention in ASD bases its conclusions on

group-averages. If few children are average, conclusions

based on averages may be less informative. It seems crucial

for future research to take into account more the variability in

outcome.

Our results show that part of the variability in progress can

be explained by differences in IQ and education type. That IQ

is a significant predictor for the outcome of an intervention

replicates earlier studies [e.g. 13, 14]. In our sample,

especially children with an IQ below 55 had a smaller

chance of making moderate or good progress compared to

children with an average IQ. To our knowledge, no study

before compared children with ASD in regular and special

education to children who do not go to school. Our results

reveal that children with ASD who do not go to school yet,

have a higher chance at making minimal progress compared

to mixed progress (good progress on some, but limited

progress on other outcome measures) than children in special

education. A possible explanation for this finding could be

that children who do not go to school receive substantially

less stimulation of their social-communicative abilities, which

could lead to a lesser amount of progress. An alternative

explanation is that these children are less used to following

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 171

instructions, which leads to less cooperation during testing.

However, causal inferences cannot be made from this kind of

analysis, which implies that a third unknown factor related to

the education type, could also be responsible for the results.

There was only a significant difference between no education

and special education, not between no education and regular

education. This could have been caused by a more even

distribution between mixed and good progress in the children

who attended regular education, while children in special

education were more in the mixed category. The different

progress dependent on education type is an important finding,

given the differences in distribution of education types in the

intervention groups. The ABA group had more children who

did not go to school yet, which may have influenced the

results. The predictive value of the intensity of the interven-

tion that was found in previous studies was not replicated in

this study. A possible explanation of this result is that

relatively small differences in intensity (e.g. one hour more or

less) do not have an equally big impact on the outcome as the

larger differences in intensity that were reported in other

studies [e.g. 13]. Whether age could predict the progress

children made, could not be assessed, since the assumption

of linearity was not met. Possibly the relationship between

age and progress is non-linear with peaks in progress at

certain ages.

A limitation of this study is that since random assignment

was not possible, we cannot assure that the three groups

differed only on the intervention method. There was, however,

no bias in the allocation to the treatment groups, because

parents chose for a particular treatment centre based solely

on the place of residence. Moreover, we showed that the

treatment groups did not differ significantly on any of the

outcome measures at the start of the study. However, we

should note that the p-levels of the tests concerning the pre-

existing differences between the groups on the outcome

measures, age and IQ were all below 0.20, which is too low to

conclude that the groups were well-matched. Both IQ and age

are related to the outcome of intervention [13–15]. However,

since the children in the ABA group seemed to have a lower

IQ on average, they also seemed to be slightly younger, these

factors may have levelled each other out in their effect on the

progress of the children. Although there were no average age

differences, the youngest child in the imitation/JA group was

already 30 months, while the youngest children in the ABA

and TAU group were respectively 22 and 24 months old. It

seems, however, unlikely that this had a large effect on the

results, since the number of children in both the ABA and

TAU groups between 22 and 30 months was very limited.

Moreover, there were differences in education type, which is a

factor that was associated with overall progress. Although

quasi-experimental designs have obvious disadvantages, the

benefit of studying interventions in the real world is that it

maximizes external validity. RCT’s are important in showing

the effect of an intervention under ideal circumstances. It is,

however, equally important to show that a treatment is

effective in a naturalistic setting, where circumstances are

seldom ideal.

Second, since there was no comparison group, which

received no treatment, we cannot be sure that the progress we

observed was caused by the treatment and not merely by

maturation. However, the significant decrease in symptoms of

ASD after only six months of intervention does not seem to be

the result of maturation. Moreover, a substantial subgroup

made progress in social-communicative abilities at the same

rate as typically developing children, which is not what one

would expect without intervention, given that these abilities

are generally impaired in children with ASD [20, 21, 61].

A third limitation is the rather low response rate on the

parent questionnaires. We cannot rule out that parents who

were more motivated to fill out questionnaires differed in

some respects from parents who did not fill out the

questionnaires. For example, this could be parents who were

more involved with the intervention of their children or

parents who were higher educated, which are factors that

could have an influence on the progress children make. The

low response rate may be explained partly by the lack of

personal contact between researcher and parents. The children

were tested in the therapy centres, without the parents being

present. Because of this, the questionnaires were not handed

directly to the parents by the researcher but afterwards by the

therapists. Fourth, the assessments were conducted by the first

author or one of three research assistants. Only the research

assistants were blind for the treatment group. A last limitation

is the rather small sample size of the subgroups, especially of

the ABA group. A study with a larger sample size would have

had more power to detect differences between the intervention

groups. Given the large p values for the repeated measures

analyses of the interaction between intervention group and

time, it seems, however, unlikely that the lack of difference in

treatment effect can be attributed to limited power. On the

other hand, a larger sample size would have been useful to

assess the effect of the intervention method as a moderator in

the regression model. We found no significant moderating

effect of intervention group, but cannot exclude that this was

not due to limited power.

This study has important implications for clinical practice.

First of all, it does not seem beneficial to add low intensity

ABA or imitation/JA intervention to TAU for all children with

ASD. However, adding these interventions can possibly lead

to better results for a subgroup of children. Therefore more

research is needed to define the characteristics of those

children. Second, there is a substantial number of children

that makes limited or no progress. Possibly, these children

would benefit from a more intensive intervention or would

make more progress with a different intervention method.

Future research could address this question by studying poor

responders and looking at the effect of changes in treatment

intensity and method. Clinicians can address this issue by

regular evaluation of the treatment effect and adapting the

intervention method used, when progress fails to occur.

Concluding, on average, there is no different effect of low-

intensive ABA, imitation/JA training and TAU used in

community settings after six months intervention on social-

communicative abilities, symptoms of ASD, adaptive behav-

iour and social-emotional functioning. On average, children in

all three groups made significant progress on imitation,

pretend play, IJA, receptive and expressive language and

adaptive behaviour and showed a decrease in autism severity.

There was however great individual variability in the

outcome in all three intervention groups. While a substantial

172 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

proportion of children made good progress, there was a

subgroup that made hardly any progress. It seems important to

focus on ‘‘What works for whom’’ instead of trying to find a

one-size-fits-all-treatment for children with ASD.

Notes

1. Some treatment centres only provided intervention to

children in regular education.

2. The videos were collected for a separate study on the

interaction between therapist and child during an ABA-

session [47] and were thus only available for the children

in the ABA-group.

Acknowledgements

We thank Sig vzw for their help in the recruitment and we

thank the treatment centres and families for participating.

Declaration of interest

The authors declare that they have no conflict of interest.

Partial funding for this research was provided by support

from Steunpunt Expertisenetwerken, Vlaamse Vereniging

Autisme and Sig vzw.

References

1. Chawarska K, Klin A, Paul R, Macari S, Volkmar F. A prospective
study of toddlers with ASD: Short-term diagnostic and cognitive
outcomes. Journal of Child Psychology and Psychiatry 2009;50:
1235–1245.

2. Granpeesheh D, Dixon DR, Tarbox J, Kaplan AM, Wilke AE. The
effects of age and treatment intensity on behavioral intervention
outcomes for children with autism spectrum disorders. Research in
Autism Spectrum Disorders 2009;3:1014–1022.

3. Howard JS, Sparkman CR, Cohen HG, Green G, Stanislaw H.
A comparison of intensive behavior analytic and eclectic treatments
for young children with autism. Research in Developmental
Disabilities 2005;26:359–383.

4. Smith T, Groen AD, Wynn JW. Randomized trial of intensive early
intervention for children with pervasive developmental disorder
(vol 105, pg 269, 2000). American Journal on Mental Retardation
2000;105:508–508.

5. Sallows GO, Graupner TD. Intensive behavioral treatment for
children with autism: Four-year outcome and predictors. American
Journal on Mental Retardation 2005;110:417–438.

6. Remington B, Hastings RP, Kovshoff H, Espinosa FD, Jahr E,
Brown T, Alsford P, Lemaic M, Ward N. Early intensive behavioral
intervention: Outcomes for children with autism and their parents
after two years. American Journal on Mental Retardation 2007;112:
418–438.

7. Reichow B. Overview of meta-analyses on early intensive behav-
ioral intervention for young children with autism spectrum
disorders. Journal of Autism and Developmental Disorders 2012;
42:512–520.

8. Eikeseth S. Outcome of comprehensive psycho-educational inter-
ventions for young children with autism. Research in
Developmental Disabilities 2009;30:158–178.

9. Vismara LA, Rogers SJ. Behavioral treatments in autism spectrum
disorder: What do we know? Annual Review of Clinical
Psychology 2010;6:447–468.

10. Eldevik S, Eikeseth S, Jahr E, Smith T. Effects of low-intensity
behavioral treatment for children with autism and mental retard-
ation. Journal of Autism and Developmental Disorders 2006;36:
211–224.

11. Strauss K, Vicari S, Valeri G, D’Elia L, Arima S, Fava L. Parent
inclusion in Early Intensive Behavioral Intervention: The influence
of parental stress, parent treatment fidelity and parent-mediated

generalization of behavior targets on child outcomes. Research in
Developmental Disabilities 2012;33:688–703.

12. Lovaas OI. Behavioral treatment and normal educational and
intellectual-functioning in young autistic-children. Journal of
Consulting and Clinical Psychology 1987;55:3–9.

13. Mazurek MO, Kanne SM, Miles JH. Predicting improvement in
social-communication symptoms of autism spectrum disorders
using retrospective treatment data. Research in Autism Spectrum
Disorders 2012;6:535–545.

14. Perry A, Cummings A, Geier JD, Freeman NL, Hughes S,
Managhan T, Reitzel JA, Williams J. Predictors of outcome for
children receiving intensive behavioral intervention in a large,
community-based program. Research in Autism Spectrum
Disorders 2011;5:592–603.

15. Virues-Ortega J, Rodriguez V, Yu CT. Prediction of treatment
outcomes and longitudinal analysis in children with autism
undergoing intensive behavioral intervention. International
Journal of Clinical and Health Psychology 2013;13:91–100.

16. Zachor DA, Ben Itzchak E. Treatment approach, autism severity
and intervention outcomes in young children. Research in Autism
Spectrum Disorders 2010;4:425–432.

17. Ingersoll B, Schreibman L. Teaching reciprocal imitation skills
to young children with autism using a naturalistic behavioral
approach: Effects on language, pretend play, and joint
attention. Journal of Autism and Developmental Disorders 2006;
36:487–505.

18. Whalen C, Schreibman L, Ingersoll B. The collateral effects of joint
attention training on social initiations, positive affect, imitation, and
spontaneous speech for young children with autism. Journal of
Autism and Developmental Disorders 2006;36:655–664.

19. Kozlowski AM, Matson JL, Horovitz M, Worley JA, Neal D.
Parents’ first concerns of their child’s development in toddlers with
autism spectrum disorders. Developmental Neurorehabilitation
2011;14:72–78.

20. Paparella T, Goods KS, Freeman S, Kasari C. The emergence of
nonverbal joint attention and requesting skills in young children
with autism. Journal of Communication Disorders 2011;44:
569–583.

21. Vanvuchelen M, Roeyers H, De Weerdt W. Do imitation problems
reflect a core characteristic in autism? Evidence from a literature
review. Research in Autism Spectrum Disorders 2011;5:89–95.

22. Luyster RJ, Kadlec MB, Carter A, Tager-Flusberg H. Language
assessment and development in toddlers with autism spectrum
disorders. Journal of Autism and Developmental Disorders 2008;
38:1426–1438.

23. Poon KK, Watson LR, Baranek GT, Poe MD. To what extent do
joint attention, imitation, and object play behaviors in infancy
predict later communication and intellectual functioning in ASD?
Journal of Autism and Developmental Disorders 2012;42:
1064–1074.

24. Charman T. Why is joint attention a pivotal skill in autism?
Philosophical Transactions of the Royal Society of London Series
B-Biological Sciences 2003;358:315–324.

25. Kasari C, Freeman S, Paparella T. Joint attention and symbolic play
in young children with autism: A randomized controlled interven-
tion study. Journal of Child Psychology and Psychiatry 2006;47:
611–620.

26. Ingersoll B. Brief report: Pilot randomized controlled trial of
reciprocal imitation training for teaching elicited and spontaneous
imitation to children with autism. Journal of Autism and
Developmental Disorders 2010;40:1154–1160.

27. Warreyn P, Roeyers, H. See what I see, do as I do. Promoting joint
attention and imitation in preschoolers with autism spectrum
disorder. Autism 2013. [Epub ahead of print]. doi: 10.1177/
1362361313493834.

28. Kasari C, Paparella T, Freeman S, Jahromi LB. Language outcome
in autism: Randomized comparison of joint attention and play
interventions. Journal of Consulting and Clinical Psychology 2008;
76:125–137.

29. Goods KS, Ishijima E, Chang YC, Kasari C. Preschool based
JASPER intervention in minimally verbal children with autism:
Pilot RCT. Journal of Autism and Developmental Disorders 2013;
43:1050–1056.

30. Casenhiser DM, Shanker SG, Stieben J. Learning through
interaction in children with autism: Preliminary data from

DOI: 10.3109/17518423.2014.933983 Effect of community interventions on social-communicative abilities 173

asocial-communication-based intervention. Autism 2013;17:
220–241.

31. Schertz HH, Odom SL, Baggett KM, Sideris JH. Effects of joint
attention mediated learning for toddlers with autism spectrum
disorders: An initial randomized controlled study. Early Childhood
Research Quarterly 2013;28:249–258.

32. Green J, Charman T, McConachie H, Aldred C, Slonims V, Howlin
P, Le Couteur A, Leadbitter K, Hudry K, Byford S, et al. Parent-
mediated communication-focused treatment in children with autism
(PACT): A randomised controlled trial. Lancet 2010;375:
2152–2160.

33. Stahmer AC, Collings NM, Palinkas LA. Early intervention
practices for children with autism: Descriptions from community
providers. Focus on Autism and Other Developmental Disabilities
2005;20:66–79.

34. Magiati I, Charman T, Howlin P. A two-year prospective follow-up
study of community-based early intensive behavioural intervention
and specialist nursery provision for children with autism spectrum
disorders. Journal of Child Psychology and Psychiatry 2007;48:
803–812.

35. American Psychiatric Association. Diagnostic and statistical
manual of mental disorders, 4th ed. (text revision). Washington,
DC: Author; 2000.

36. Charman T, Baird G. Practitioner review: Diagnosis of autism
spectrum disorder in 2-and 3-year-old children. Journal of Child
Psychology and Psychiatry 2002;43:289–305.

37. van der Meulen BF, Ruiter SAJ, Spelberg HC, Smrkovsky M.
Bayley scales of infant development. Nederlandse versie. BSID-II-
NL. Lisse: Swets Test Publishers; 2002.

38. Hendriksen JG, Hurks PM. WPPSI-III-NL Nederlandstalige
bewerking: Technische handleiding. Amsterdam: Pearson
Assessment and Information BV; 2009.

39. Vander Steene G, Bos A. Wechsler preschool and primary scale of
intelligence. Vlaams-Nederlandse Aanpassing, 2nd ed.
Testinstructie. Lisse: Swets Test Publishers; 1997.

40. Tellegen PJ, Winkel M, Wijnberg-Williams BJ, Laros JA. Snijders-
Oomen, niet verbale intelligentietest: SON-R 2–7: Handleiding en
verantwoording. Lisse: Swets Test Publishers; 1998.

41. Pameijer N, van Beukering T. Handelingsgerichte diagnostiek. Een
praktijkmodel voor diagnostiek en advisering bij onderwijsleerpro-
blemen. Leuven/Voorburg: Acco; 2007.

42. van der Meulen BF, Smrkovsky M. MOS 2–8 Mc Carthy
Ontwikkelingsschalen: Handleiding. Lisse: Swets & Zeitlinger;
1985.

43. Lord C, Rutter M, DiLavore P, Risi S. Autism diagnostic
observation schedule: Manual. Los Angeles: Western
Psychological Services; 1999.

44. Sundberg ML, Michael J. The benefits of skinner’s analysis of
verbal behavior for children with autism. Behavior Modification
2001;25:698–724.

45. Skinner BF. Verbal behavior. Englewood Cliffs, NJ: Prentice Hall;
1957.

46. Sundberg ML. Verbal behavior milestones assessment and place-
ment program: A language and social skills assessment program for
children with autism or other developmental disabilities. Concord:
AVB Press; 2008.

47. Van der Paelt S, Warreyn P, Roeyers H. ABA under the looking
glass: Teaching social-communicative abilities to children with
autism spectrum disorder 2014. [Epub ahead of print].

48. Gotham K, Risi S, Pickles A, Lord C. The autism diagnostic
observation schedule: Revised algorithms for improved diagnostic
validity. Journal of Autism and Developmental Disorders 2007;37:
613–627.

49. Vanvuchelen M, Roeyers H, De Weerdt W. Development and initial
validation of the Preschool Imitation and Praxis Scale (PIPS).
Research in Autism Spectrum Disorders 2011;5:463–473.

50. Lewis V, Boucher J. Manual of the test of pretend play. London:
The Psychological Corporation; 1997.

51. Mundy P, Delgado C, Block J, Venezia M, Hogan A, Seibert J. A
manual for the abridged Early Social Communication Scales
(ESCS). Unpublished manuscript. Departement of Psychology,
University of Miami, Coral Gables, Florida; 2003.

52. Noldus Information Technology. The observer XT: The neXT
generation of observation software. Reference manual, version XT
9.0. Wageningen, The Netherlands: Author; 2009.

53. Yoder P, Stone WL, Walden T, Malesa E. Predicting social
impairment and ASD diagnosis in younger siblings of children with
autism spectrum disorder. Journal of Autism and Developmental
Disorders 2009;39:1381–1391.

54. Schaerlaekens A, Zink I, Van Ommeslaeghe K. Reynell
Taalontwikkelingsschalen. Handleiding – Tweede versie. Lisse:
Swetz & Zeitlinger; 2003.

55. Scholte EM, van Duijn G, Dijkxhoorn Y, Noens I, van Berckelaer-
Onnes IA. Vineland Screener 0-6 jaar. Leiden: PITS; 2008.

56. Warreyn P, Raymaekers R, Roeyers H. Handleiding Vragenlijst
Sociale Communicatie. Destelbergen: SIG vzw; 2004.

57. Rutter M, Bailey A, Lord C. Social Communication Questionnaire
(SCQ). Los Angeles: Western Psychological Services; 2003.

58. Verhulst FC, van der Ende J. Gedragsvragenlijst voor kinderen van
1

1/2
-5 jaar. Rotterdam: Erasmus MC – Sophia Kinderziekenhuis;

2000.
59. Achenbach TM, Rescorla LA. Manual for the ASEBA preschool

forms & profiles. Burlington, VT: University of Vermont, Research
Center for Children, Youth, & Families; 2000.

60. Eikeseth S, Smith T, Jahr E, Eldevik S. Intensive behavioral
treatment at school for 4-to 7-year-old children with autism – A
1-year comparison controlled study. Behavior Modification 2002;
26:49–68.

61. Jarrold C. A review of research into pretend play in autism. Autism
2003;7:379–390.

174 S. Van der Paelt et al. Dev Neurorehabil, 2016; 19(3): 162–174

Copyright of Developmental Neurorehabilitation is the property of Taylor & Francis Ltd and
its content may not be copied or emailed to multiple sites or posted to a listserv without the
copyright holder’s express written permission. However, users may print, download, or email
articles for individual use.

• Effect of community interventions on social-communicative abilities of preschoolers with autism spectrum disorder

Introduction
Methods
Results
Discussion
Notes
Acknowledgements
Declaration of interest
References

Procedure

ASD children

92 children recruited

16 Funded multidisciplinary centers

Using DSM-IV-TR

8 children with Dx. at HR for ASD

5 treatment centers used ABA

6 treatment centers used Imitation & TAU

7 children dropped out

85 children, ages 22-75 months

5 treatment centers TAU

Characteristics investigated

Age

Autism severity

Time elapsed

Maternal & paternal

Gender

IQ

Education

Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR).
Children divided into 3 intervention groups.
1

Results
Based on 3 interventions
ABA – Applied Behavior Analysis
JA- Joint attention
TAU- Treatment as usual.
Study was analyzed based on comparison of treatment between intervention groups.
Using 2 repeated MANOVA Analysis
Results from questionnaires were analyzed differently.
Descriptive statistics for pre & post measurement in the different intervention groups of all dependent variables.
Test results revealed no interaction between intervention group & time, which means that the progress children made, was not different in the 3 intervention groups.
There was an effect of time irrespective of intervention group.
Time
Intervention group

There was no stastical difference in intervention groups and the various intervention measures carried out.
2

Individual variability
Progress Age Outcomes % of Progress
1 Measurement of interventions 22- 75 months Imitation, Pretend Play, language abilities & adaptive behavior. All three intervention groups (ABA, TAU, & JA)
Good progress > 6 months in AE 3 or 4 outcomes. 26%
Moderate to good progress 3-5 months in AE 3 outcomes 7%
Mixed progress Normal development plan 1-2 outcomes 34%
Limited to moderate progress 3-5 months 1-2 outcomes 14%
Limited progress No progress or deteriorated Not ˃ 2 months, 19%

This is a simplified table of how the results were presented. For the various interventions they employed, there had to be measurements. They classified progress into categories of good to little progress (5 categories). Then, defined the ages that would signify if a child has undergone improvement or not, depending on how good they perform in different outcomes like Imitation, pretend play etc. Then analyze the total progress percentages of all three interventions.
3

Predicting Outcomes
1. No difference in average scores for different intervention groups.
Used global progress categories over multiple regression analysis.
This enables better analysis of outcomes.
Predictors: Age, IQ-category, autism severity & treatment intensity & number of months in intervention.
2. Significant difference in education types in intervention groups.
3. Found that children in different types of education progressed differently
regular education,
regular education with support,
special education and not in school yet.
4. Merging of progress, intervention groups & analyzing of continuous predictors to ensure enough observations.
5. Age was omitted as a factor.
6. Education type & IQ were the only significant predictors using multinomial regression analysis.

total number of hours of intervention a week, number of hours of 1:1 intervention.
Predictors were used as outcomes- as stated.
4