J Autism Dev Disord (2015) 45:3168–3182

DOI 10.1007/s10803-015-2474-x

ORIGINAL PAPER

Managing Repetitive Behaviours in Young Children with Autism

Spectrum Disorder (ASD): Pilot Randomised Controlled Trial
of a New Parent Group Intervention
Victoria Grahame1 • Denise Brett2 • Linda Dixon1 • Helen McConachie2 •

Jessica Lowry2 • Jacqui Rodgers3 • Nick Steen2 • Ann Le Couteur1,2

Published online: 3 June 2015

Ó Springer Science+Business Media New York 2015

Abstract Early intervention for autism spectrum disorder Introduction

(ASD) tends to focus on enhancing social-communication
skills. We report the acceptability, feasibility and impact on Autism spectrum disorder (ASD) is a common neurode-
child functioning of a new 8 weeks parent-group inter- velopmental disorder that affects as many as 1 % of chil-
vention to manage restricted and repetitive behaviours dren and has a considerable impact on a child’s
(RRB) in young children with ASD aged 3–7 years. Forty- development (Baird et al. 2006). One key diagnostic do-
five families took part in the pilot RCT. A range of primary main of ASD is restricted, repetitive and stereotyped pat-
and secondary outcome measures were collected on four terns of interests, behaviours and activities (ICD-10, World
occasions (baseline, 10, 18 and 24 weeks) to capture both Health Organization 1992; DSM-5, American Psychiatric
independent ratings and parent-reported changes in RRB. Association 2013). Narrowness of focus, inflexibility,
This pilot established that parents were willing to be re- perseveration of interest in activities and insistence on
cruited and randomised, and the format and content of the sameness reflect the restrictedness aspect of this domain,
intervention was feasible. Fidelity of delivery was high, while repetition is demonstrated in repetitive speech, rou-
and attendance was 90 %. A fully powered trial is now tines, rituals and rhythmic stereotypies (Leekam et al.
planned. 2011). The recent publication of DSM-5 criteria for a di-
agnosis of ASD includes sensory sensitivities within the
Keywords Restricted and repetitive behaviour
Parent restricted repetitive behaviour (RRB) domain for the first
group intervention
Clinical global impression of time. This new addition acknowledges the range of sensory
improvement
Randomised controlled trial symptoms experienced by individuals with ASD (Leekam
et al. 2007a, b).
The frequency and types of RRB displayed by children
with ASD can vary according to age and cognitive ability.
In a sample of 121 children with ASD, Militerni et al.
(2002) found that children with ASD aged 2–4 years dis-
played more motor and sensory behaviours e.g. spinning or
& Victoria Grahame pulling coat zipper up and down while children aged
victoria.grahame@ntw.nhs.uk 7–11 years displayed more complex behaviours such as
1
insistence on sameness. They also found that children with
Complex Neurodevelopmental Disorders Service,
Northumberland, Tyne and Wear NHS Foundation Trust,
lower IQ showed a more frequent interest in motor and
Walkergate Park, Benfield Road, sensory behaviours, with those with higher IQ demon-
Newcastle upon Tyne NE6 4QD, UK strating more complex RRB. However not all RRB is re-
2
Institute of Health and Society, Newcastle University, garded as problematic or warranting a targeted
Newcastle upon Tyne, UK intervention. Some RRB may indicate an area of relative
3
Institute of Neuroscience, Newcastle University, strength or a special skill. For some people, childhood
Newcastle upon Tyne, UK circumscribed interests may lead to employment in adult

123
J Autism Dev Disord (2015) 45:3168–3182 3169

life (Leekam et al. 2011). However, for the majority of children displaying intermittent RRB were excluded as the
individuals, RRB can have a disruptive impact on both the behaviour could not be reliably elicited in the weekly
child and their family. RRB can interfere with the child’s session.
ability to learn new skills, engage in daily living activities In acknowledging the unmet need for a parent-mediated
and have been shown to take up large amounts of time intervention focussing on RRB, we have developed in
(Cunningham and Schreibman 2008; Dunlap et al. 1983; collaboration with parents a new group based intervention
Rapp and Vollmer 2005). RRB can also be stigmatizing to support parents to recognise, understand and manage
(Cunningham and Schreibman 2008), and lead to agitation RRB in young children with ASD. The Managing Repeti-
or aggression if interrupted (Gordon 2000). Higher levels tive Behaviours Programme (MRBÓ) is designed to be run
of RRB are also associated with heightened anxiety, for by group leaders with experience working with young
example, insistence on sameness, may be a maladaptive children with ASD. MRB utilises parent group learning
coping response to anxiety (Rodgers et al. 2012; Lidstone alongside opportunities for mutual support and sharing of
et al. 2014). RRB can interfere with family functioning and strategies. The aim is that through helping parents to gain
have been reported as among the most stressful behaviours knowledge and skills in managing their child’s RRB, this
for parents to manage (Bishop et al. 2007) in addition to would lead to improvement in the child’s overall clinical
provoking negative parenting styles (Greenberg et al. function and reduction in RRB.
2006).
An evidence base for the efficacy of parent-mediated Outcome Measurement
interventions for young children with ASD is emerging
(Oono et al. 2013). Such interventions may enable parents A wide range of measurement tools have been used in the
to capitalise on teachable moments as they occur in the study of RRB, mainly questionnaires and diagnostic in-
environment and provide learning opportunities during terviews (Honey et al. 2012). A limited range of structured
naturally occurring situations. Building on these real world observation methods focusing specifically on RRB have
experiences may also facilitate generalization of the child’s been developed. Watt et al. (2008) and Barber et al. (2012)
learning across contexts (Carter et al. 2011). However, investigated RRB in toddlers using a videotaped behaviour
most ASD-specific early intervention programmes focus on sample of the Communication and Symbolic Behaviour
social communication (Green et al. 2010; Magiati et al. Scales Developmental Profile (CSBS-DP; Wetherby and
2014; Oono et al. 2013). Parents rarely receive specific Prizant 2002). Boyd et al. (2010) developed the Direct
advice on their child’s RRB and much of the existing Observation of Repetitive Behaviours in Autism (DOBRA)
evidence comes from single-case behavioural studies coding system which measures RRB, appropriate be-
(Mulligan et al. 2014). Techniques used include differential haviour, problem behaviours, and interference. Harrop
reinforcement of variability in behavioural responding to et al. (2014) developed a coding system using items from
reduce routines and insistence on sameness (Miller and validated RRB questionnaires to measure RRB in toddlers
Neuringer 2000), response interruption and redirection for and preschool children. However, the main limitation of
vocal and motor stereotypy (e.g. Ahrens et al. 2011), direct observation in RRB is the failure to capture the range
functional communication training for compulsive be- and variety of RRB.
haviour (Kuhn et al. 2009) and techniques such as visual An additional issue in measurement of change in RRB is
schedules or video based modelling to help tolerate chan- the need for individualised measurement of outcomes.
ges to routine or expand repetitive play (Hine and Wolery Arnold et al. (2003) proposed an individualised target
2006). While these studies offer evidence of success, symptom assessment procedure (target behaviour) involv-
limitations include implementation at specialised centres, ing quantified behaviour ratings of parent reports. Arnold
highly individualised delivery with expert clinicians, and and colleagues found this assessment highly convergent
small sample sizes. One recent development of a parent- with their other outcome measures, evidence that indi-
mediated intervention targeting RRB is the Family-Imple- vidualised outcomes can reflect a true change in behaviour
mented Treatment for Behavioural Inflexibility (FITBI; in heterogeneous groups.
Boyd et al. 2011). Direct instruction and naturalistic be- This paper reports the findings of a feasibility and ac-
havioural teaching methods were used to reduce targeted ceptability pilot randomised controlled trail of a new parent
RRB, through a therapist working with parent and child for group intervention (Managing Repetitive Behaviours). The
12 weekly sessions for 1–2 h. There was a reduction in the objectives of the study were to (a) investigate the feasibility
occurrence of RRB for the five child participants, and an and acceptability of the intervention, (b) review the re-
increase in their engagement with more appropriate ac- search protocol (including the randomisation and retention
tivities. However, as the intervention involved discrete processes; and (c) explore the utility of the outcome mea-
teaching trials with the child during the weekly sessions, sures to assess the effects of the intervention on RRB,

123
3170 J Autism Dev Disord (2015) 45:3168–3182

overall child functioning and parents’ reported self- Vineland Adaptive Behaviour Scales II (VABS II; Sparrow
efficacy. et al. 2005) The VABS II measures aspects of an indi-
vidual’s level of adaptive functioning. The parent/caregiver
rating form was used which focuses on four domains;
Methods communication, daily living skills, socialization and motor
skills. The assessment was undertaken with parents as an
Participants interview by trained research psychologists. The domain
composite scores provide an adaptive behaviour composite.
Inclusion criteria for the study were parents of young
children aged 3–7 years 11 months, with a confirmed Demographics
clinical diagnosis of ASD who identified a problematic
RRB to work on during the group. Parents were required to In the baseline interview, parents were asked about their
have sufficient spoken English to take part in assessments child’s age, gender, diagnosis and ethnicity; previous in-
and the group-based intervention, be willing to be ran- terventions, current medication and additional diagnoses.
domised, agree to maintain current medication regime, and Information was also obtained on parents level of educa-
agree not to try any other new intensive interventions tion, employment status, family structure, and if they had
during the course of the study. There were no child ex- attended a previous course or intervention. Socioeconomic
clusion criteria as the intervention is designed for parents status was calculated using Townsend’s Index of Depriva-
of children with ASD across a range of abilities; from tion based on the parent’s postcode (Townsend et al. 1988).
profound intellectual disability to average or above average
intelligence. A total of 64 families expressed an interest in Primary Outcome Measures
taking part; of these, two were excluded who did not meet
the inclusion criteria (one child was too young, one parent Taking into account the strengths and weaknesses of
did not have sufficient English) and 17 eventually declined standardised questionnaire-based measures, observational
to participate. Forty-five completed baseline assessments techniques and individualised outcomes, the decision was
and progressed to randomisation (see Fig. 1), a sufficient made to include a combination of approaches in this pilot
number to estimate key parameters for a future trial study to evaluate both RRB and impact on child and family
(Julious 2005; Lancaster et al. 2004). functioning.

Measures Clinical Global Impressions: Improvement scale (CGI-I;

Guy 1976) The CGI-I requires the clinician to assess how
Baseline Characterisation much symptoms have improved or worsened relative to the
child’s baseline state using a seven-point scale (1—very
Autism Diagnostic Observation Schedule-2 (ADOS-2; Lord much improved; 2—much improved; 3—minimally im-
et al. 2012) The ADOS-2 is a semi-structured observa- proved; 4—no change; 5—minimally worse; 6—much
tional assessment that involves social interaction between worse; or 7—very much worse). An independent panel of
the examiner and the child; children were assessed with expert ASD clinicians, blind to group allocation, rated
Modules 1, 2 or 3 according to language level and global improvement in overall interaction between parent
chronological age. The examiner scores elements of the and child and how much the child’s RRB had changed over
child’s behaviour in two domains, social affect and RRB. the 24 weeks since baseline. The clinicians rated inde-
The scores for the domains are combined into a total score. pendently using information from all time points (e.g.
Severity scores were also calculated according to Lord et al. questionnaires, ADOS 2 videos, target behaviour vignettes
(2012) ranging from 1 to 10, with scores of 1–2, 3–4, 5–7, and videos of parent child interactions) before reaching
8–10 indicating minimal to no evidence, low, moderate and consensus. Ratings of 1 and 2 were regarded as ‘im-
high degree of autistic impairment, respectively. The provement’ at week 24.
ADOS 2 was conducted by trained research psychologists.
Measurement of RRB: Target Behaviour Vignette As part
Social Responsiveness Scale (SRS; Constantino and Gru- of the baseline characterisation, two repetitive behaviours
ber 2005) The SRS is a 65-item quantitative measure of of most concern for each parent/caregiver were identified
the severity and type of social impairments that are char- and questions asked of the parent about their duration,
acteristic of ASD, completed by the caregiver. Higher total impact and possible triggers and functions. The protocol
scores on the SRS indicate greater severity of social for measuring change was developed by The Research
impairment. Units on Paediatric Psychopharmacology and Psychosocial

123
J Autism Dev Disord (2015) 45:3168–3182 3171

Expressed an Interest
(n=64)

Declined to participate (n=17)

Not included (n=2)

Randomized (n= 45)

Immediate intervention (n= 25) Delayed intervention (n=20)

Dropped out during intervention
Allocation (n= 2)

Assessed at time point 1 (n = 23) Assessed at time point 1 (n = 20)

1st Outcome
Lost to follow-up (n=2)
Assessment
Moved away = 1
Time pressure = 1

Assessed at time point 2 (n = 23) Assessed at time point 2 (n = 18)

2nd Outcome Lost to follow-up assessment (n=2)
Assessment Work commitments = 1
Practical problem = 1

3rd Outcome Assessed at time point 3 (n = 23) Assessed at time point 3 (n=18)
Assessment

Analysed (n = 23) Analysed (n= 20)

Analysis 1 family did not attend
intervention but completed all
assessments.

Fig. 1 CONSORT diagram

Interventions (RUPP Autism Network; Arnold et al. 2003). with the procedure developed by Arnold et al. (2003), after
At each outcome assessment point, the parent was asked all data were collected, a panel of blinded autism re-
‘‘At the beginning of the study you said you were con- searchers independently rated change in each target be-
cerned about (parent defined target behaviour at baseline). haviour. Strong agreement was achieved (intraclass
How has it been in the last couple of weeks?’’. The parent correlation coefficient = 0.81). The average rating for each
responses at each time point contributed to a written vi- behaviour was used in analysis. The target behaviours were
gnette by the researcher (who was blind to group inter- first categorised as: 1—repetitive motor movements; 2—
vention status). Identifying features such as gender and age rigidity, adherence to routine and insistence on sameness;
were not included in each vignette. The baseline vignette 3—preoccupation with restricted pattern of interest, limited
detailing the target behaviour was paired with a vignette play; or 4—unusual sensory interests. Each pair of vi-
from each follow-up, to provide a comparison. In keeping gnettes were rated on a nine point scale of improvement/

123
3172 J Autism Dev Disord (2015) 45:3168–3182

deterioration (1—normal; 2—markedly improved; 3— and duration of each behaviour by the actual time of the
definitely improved; 4—equivocally improved; 5—no video, and then multiplying by ten).
change; 6—equivocally worse; 7-definitely worse; 8—
markedly worse; 9—disastrously worse). Arnold et al. Secondary Outcome Measures
(2003) defined a positive response as a rating of three or
less and reported that the target behaviour measure was Parent Self-Efficacy (Sofronoff and Farbotko 2002) This
highly convergent with the CGI-I. 15-item questionnaire completed by parents/carers mea-
sures behaviours typically exhibited by children with ASD
Measurement of RRB: Repetitive Behaviour Questionnaire- including RRB. Parents indicated ‘yes’ or ‘no’ to whether
2 (RBQ-2; Leekam et al. 2007a, b) The RBQ-2 is a the child displayed each of the behaviours in the previous
20-item questionnaire completed by parents/carers that month and then rated their confidence in managing the
measures the frequency/severity of RRB known to occur in behaviours on a six-point scale ranging from 0 (no confi-
both ASD and typical development. The RBQ-2 was de- dence) to 5 (complete confidence). A mean self-efficacy
veloped using items from the RBQ (Turner 1995) and the score was then calculated by dividing the total confidence
Diagnostic Interview for Social and Communication score by the number of behaviours reported as displayed.
Disorders (DISCO; Wing et al. 2002). Leekam et al. (2007a,
b) found evidence for a four factor structure for the ques- Teacher-Completed RBQ-2 The teacher of each child was
tionnaire aligned with the RRB described in the interna- also asked to complete the same 20 item Repetitive Be-
tional classification systems for ASD; factor 1—repetitive haviour Questionnaire-2 (RBQ-2; Leekam et al. 2007a, b)
motor movements; factor 2—rigidity, adherence to routine questionnaire as parents/carers to assess the frequency and
and insistence on sameness; factor 3—preoccupation with severity of RRB in a school setting at each time point. This
restricted pattern of interest; and factor 4—unusual sensory was to assess whether there was any difference in presen-
interests. In a sample of typically developing 2-year-olds, tation of RRB between home and school and potential
Leekam et al. (2007a, b) found the RBQ-2 to have good generalisation of results across settings.
internal consistency, inter-item validity, and across samples
validity for the two geographical sub samples. Lidstone Focus Groups
et al. (2014) also found the RBQ-2 to be a suitable measure
of RRB in a sample of children with ASD aged 2–17 years, Focus groups were held at the end of the trial and after data
showing good internal consistency. collection was complete. They were designed to discuss with
parents three broad topics: the research process, acceptability
Measurement of RRB: Parent–Child Interactions Parents of the MRB intervention and impact of the intervention on the
and children were videoed interacting with a standardised participants, their children and the family. These groups were
set of toys for 10-min at baseline and at each outcome led by two independent facilitators with knowledge of ASD.
assessment. The frequency and duration of RRB observed, Questions asked included ‘How did you feel about being
alongside parent strategies to manage these behaviours, randomised?’, ‘How was your experience of being in the
were coded using a scheme previously developed by Shafi group?’, and ‘What impact has the MRB group had on your
(2009; available on request from the corresponding author). family’s quality of life?’. Three focus groups with parents
The RRB coded were those likely to be observed in a short were completed, with between three and six parents attending
play interaction, and consequently are narrower and more each one. Framework analysis (Ritchie and Spencer 1994)
detailed than the four categories described above. The was used to extrapolate main themes from the transcripts.
types of RRB coded were: narrow repetitive interests; Framework analysis begins deductively from the objectives
stereotyped behaviour/non-functional interests; specific set for the focus groups, but also uses an inductive approach
sensory interests; unusual or repetitive motor movement; from the accounts of the participants, i.e. new themes can
repetitive words/sounds. Parents’ response to these RRB emerge from the discussion with participants.
were categorised into four types: non-intervening; pre-
venting; engaging; and distracting/developing. Ratings Procedure
were undertaken by two blinded trained researchers. To
assess inter-rater reliability, 25 % of videos were double A positive ethical opinion was received from Newcastle
coded, half immediately after training and half during the and North Tyneside Research Ethics Committee (11/NE/
coding process. Intraclass correlations for each of the 0379). Referrals were made through clinicians in Child and
categories of RRB and parent strategies ranged from 0.70 Adolescent Mental Health and Early Years Education
to 0.91. Twelve videos did not have a total duration of teams in North-East England who discussed the research
10-min; raters pro-rated these videos (dividing frequency with parents of children with ASD who met the inclusion

123
J Autism Dev Disord (2015) 45:3168–3182 3173

criteria. They then returned expression of interest forms to (e.g. visual cues) and delivery techniques such as prompt,
the research team, who telephoned the families to arrange distract and redirect. It also includes original interactive
an initial home visit to take informed consent; baseline activities developed in collaboration with parents such as
assessment was conducted in a clinical setting. Parents quizzes. MRB places emphasis on discussion, developing
completed the RBQ-2, VABS II, SRS and parent self-ef- and sharing ideas, experiences and strategies, importantly
ficacy. The research psychologist conducted the target building parent’s knowledge and confidence to manage their
behaviour interview, ADOS 2 and video recorded the child’s RRB.
parent child interactions. Eligible participants were ran-
domised in a 1:1 ratio to ‘immediate intervention’ or ‘de- Fidelity
layed intervention’ within four blocks of referrals (www.
randomization.com). Group interventions ran consecutive- Three independent raters were randomly allocated 50 % of
ly from May 2012–May 2013 in four different locations. the recorded group intervention sessions to rate for fidelity to
Assessments took place immediately at the end of the the treatment manual, using a checklist developed for the
group intervention (10 weeks) and at two further time study. Raters rated fidelity to the manualised session content
points, 18 and 24 weeks after start of intervention. All using a three point scale (0—not at all; 1—briefly covered
outcome assessments included target behaviour vignette, but insufficiently; 2—covered adequately), and therapeutic
parent RBQ-2, parent self-efficacy and parent–child inter- best practice including techniques used, generic acceptable
action. Families remained under the clinical responsibility therapeutic components, and undesirable components rated:
of local teams and continued to receive their existing 0—not at all; 1—minimal evidence; 2—several examples.
routine care during the course of the study. At the end of
the study the parents in the delayed group were offered the Data Analysis
MRB intervention.
Patterns of recruitment, retention and participation in the
Intervention intervention were examined. Independent t tests, Chi
square tests and Fisher’s exact test were used to compare
The intervention involves eight weekly 2 hour sessions. The baseline scores on the characterisation and outcome mea-
group is designed to be delivered by early years profes- sures. The target behaviour vignettes, parent RBQ-2, par-
sionals with knowledge and experience of working with ent–child interactions, and self-efficacy were evaluated for
young children with ASD and their families. The interven- completion rates and distribution of scores to assess ap-
tion focuses on helping parents understand lower and higher propriateness of parametric analysis. Transformations were
order RRB, identify potential developmental and environ- used for the parent–child interactions; all other outcome
mental factors that may trigger RRB for their child, and measures had normal distribution. Repeated target be-
teaches parents to use a functional analytic approach to plan haviour vignettes, parent RBQ-2 and parental self-efficacy
appropriate behavioural strategies which are effective for were analysed using a multi-level model (occasions nested
their child and family. Functional analysis helps parents to within subjects), that is, first, whether a linear trend over
understand their child’s RRB, where and how to intervene in time was an appropriate model and then whether the trend
order to manage this specific behaviour. Each parent is also differed by allocation. Due to the number of separate codes
given individual support [weeks 2 and 6] to further specify (10), estimates of variability in parent–child interactions
and review one of their chosen target RRB. Parents video the were based on 2 (group) 9 4 (time-point) mixed analyses
target behaviour at home. This target behaviour is the focus of variance (ANOVA). Groups were compared on the CGI-
for parents to practise the new skills they are learning, thus I using Fisher’s exact test. For all analyses, a p value of
ensuring that strategies are individually tailored for each 0.05 was accepted as significant. Partial eta-squared (g2p)
child, e.g. reinforcing other desirable behaviour. statistic was used as a measure of effect size for ANOVA,
The MRB parent group intervention has adapted and where values [0.01, 0.06, and 0.14 indicated a small,
synthesised components of existing good practice for tar- medium, and large effect size respectively (Cohen 1988).
geting and modifying RRB, challenging behaviour and so-
cial communication difficulties. It incorporates
psychoeducation on ASD and RRB, helps parents formulate Results
their child’s behaviour using a basic functional analytic
framework, facilitates group discussion and tailoring of Participation and Attrition
strategies to manage RRB. The group also provides the op-
portunity to conceptualise RRB as a communicative function Of the 25 families randomised to immediate intervention,
and therefore utilises adapted materials for communication one family did not attend the intervention, but completed

123
3174 J Autism Dev Disord (2015) 45:3168–3182

follow-up assessments. Two families completed half of the Acceptability of the Intervention
sessions before dropping out of the course and research.
For the delayed intervention group, two families were lost Four principal themes emerged: (1) most participants re-
to follow-up at time point 2 (Fig. 1). There was a com- ported that they had little knowledge of RRB before at-
pletion rate of 89 % for all families (from baseline to tending the intervention; (2) comments about the effort
6-month follow-up). All results from previous outcome participants had made to attend and ‘‘get the most out’’ of
measures were carried forward for those who dropped out, the intervention; (3) positive comments about the inter-
in line with an intention to treat analysis. Of the 22 families vention and the opportunity to ‘‘share experiences’’ with
who participated in the intervention, there was an average other parents who were ‘‘going through the same things’’
attendance rate of 90 % at the sessions. Participant char- and (4) difficulties completing homework diaries or
acteristics are shown in Table 1. videoing their child engaging in a RRB although ac-
knowledging it was an ‘‘opportunity for the professionals to
Baseline Equivalence of Groups see behaviour first hand and comment on it’’.

There was no difference between allocation groups on Impact of the Intervention on the Participants, their
demographic variables or baseline characterisation of Children and the Family
measures (see Tables 1, 2).
Three themes relating to the impact of the MRB inter-
Fidelity of Intervention ventions were elicited: (1) positive changes that the course
had had on them personally as participants such as feeling
A panel of expert observers rated adherence to the content more confident, having an increased awareness and more
of the ‘Managing Repetitive Behaviours’ (MRB) manual as knowledge about RRB and feeling more equipped to deal
98 % and delivery of therapeutic best practice as 93 %. with their child’s RRB. ‘‘It has made me more aware of my
actions and made me think about targeting other repetitive
Acceptability of Intervention behaviours’’ (2) changes in their child’s RRB in terms of
frequency, duration or intensity, or that their child’s be-
Focus Groups haviour was becoming more manageable: ‘‘The group has
affected my child’s RRB in an extraordinary way. He now
Each group followed a semi-structured topic guide with the knows he has boundaries within his RRB which was a huge
aim of considering three key topics (a) the research pro- issue for us’’ and (3) participants talked about their desire
cess; (b) acceptability of the MRB intervention; (c) impact to disseminate strategies to other people involved their
of the intervention on the participants, their children and child’s care, such as other family members and teachers.
the family. They frequently described wanting to get ‘‘everyone on the
same page’’ in their family to provide a consistent approach
Research Process to managing their child’s RRB:

Parents’ views about being part of a randomised controlled Outcome Measures

trial (RCT) were divided into four themes. Parents com-
mented about (1) Assessments and follow-ups, (2) Contact Clinical Global Impressions: Improvement Scale
with research team, (3) Randomisation and (5) Time in
project. Discussions revealed some procedural issues with There was a significant difference between immediate in-
randomisation; some terms about randomisation might need tervention group (IG) and delayed intervention group (DG)
to be clarified and contact details of who to contact if there on the blind rating of CGI-I (Fisher’s exact test p = 0.05,
is an issue to do with randomisation need to be made more two-tailed). Seven (30 %) participants in the IG met CGI-I
explicit in the future; ‘‘I found it hard to not let it slip about criteria for very much improved/much improved (respon-
taking part in the group.’’ Some parents did not fully grasp ders) compared to 1 (5 %) in the DG at time point (TP)3.
what was meant by the term ‘‘the researchers are blind to There were 6 (26 %) in the IG and 4 (20 %) in the DG who
who has attended the MRB programme’’ misinterpreting met criteria for minimally improved, while 10 (44 %) in
this to mean that they should conceal any new learning from the IT and 15 (75 %) in the DG met the criteria for no
researchers during the follow-up appointments. change.

123
J Autism Dev Disord (2015) 45:3168–3182 3175

Table 1 Participant demographics

Immediate (N = 25) Delayed (N = 20) Difference
(%) (%) p

Child
Age (months) 60.44 M (13.44 SD) 62.75 M (16.89 SD) 0.612*
Gender 24 male, 1 female 15 male, 5 female 0.074
Psychotropic medication 0 male, 1 female 0 male, 0 female 0.444
Melatonin 3 male, 1 female 3 male, 1 female 1.0
Child’s ethnicity 0.682
White background 22 (88) 16 (80)
Other background 3 (12) 4 (20)
Other diagnoses 0.938**
Yes 6 (24) 5 (25)
No 19 (76) 15 (75)
Previous interventions 0.071**
Speech therapy 14 (56) 9 (45)
Other 9 (36) 4 (20)
None 2 (8) 7 (35)
Mean ADOS (severity) 6.96 (SD 2.03) 7.05 (2.21) 0.766*
Mean SRS (total) 115.44 (SD 26.12) 126.0 (28.73) 0.204*
Mean VABS-II (adaptive behaviour composite) 68.40 (SD 14.04) 69.20 (14.76) 0.854*
Parent
Previous course attended 0.641**
Yes 12 11
No 13 9
Family make up 0.716
Married/living with partner 20 (80) 17 (85)
Single/divorced 5 (20) 3 (15)
Parent education 0.718**
School leavers 12 (48) 8 (40)
A—levels 7 (28) 5 (25)
University/post-graduate 6 (24) 7 (35)
Parent occupation 0.373**
Full-time employed 11 (44) 12 (60)
Unemployed/other 14 (56) 8 (40)
Socioeconomic status? 2.82 (3.14 SD) 2.39 M (2.88 SD) 0.637*
?
* t test; Fishers test; ** Pearson’s Chi square; Townsend index of deprivation

Measurement of RRB: Target Behaviour Vignette To analyse the four factors of the target repetitive be-
haviours individually, participants in the immediate group
From the boxplot, it can be seen that the immediate in- were matched with participants from the delayed group
tervention group had greater improvement at each outcome based on category of RRB of their specific target be-
assessment time point (i.e. lower rating; Fig. 2). The trend haviours (see Table 2). There was a significant change in
across visits was fitted separately for immediate and de- ratings of the types of RRB included in Factor 3 (Preoc-
layed intervention groups. The immediate impact of the cupations with restricted patterns of interest and limited
intervention characterised by the difference between the play; N = 9 in immediate, N = 8 in delayed). Independent
groups at the first outcome assessment visit was -0.45 samples t tests revealed a significant difference between
(95 % CI -1.23, 0.33). Then subsequently, the difference groups at TP 2, t(15) = 2.83, p = 0.01, with the immediate
in the average change between visits was -0.13 (95 % CI group having greater improvement i.e. lower scores (me-
-0.55, 0.29). dian 2.25, interquartile range 1.63) than delayed (median

123
 Table 2 Means and SD of outcome variables at baseline, time point 1, time point 2, time point 3
3176

Baseline Baseline equivalence Time point 1 Time point 2 Time point 3

Mean (SD) Mean (SD) Mean (SD) Mean (SD)

123
Immediate Delayed p Immediate Delayed Immediate Delayed Immediate Delayed

Parent total RBQ-2 40.76 (7.62) 42.35 (7.29) 0.482 38.55 (7.86) 42.00 (7.54) 38.88 (8.25) 41.65 (7.69) 38.68 (7.73) 40.67 (7.43)
Parent factor RBQ-2
Factor 1 9.56 (2.89) 10.65 (2.08) 0.164 9.32 (2.58) 10.35 (2.13) 9.44 (2.57) 9.30 (1.89) 9.44 (2.65) 9.45 (2.14)
Factor 2 14.44 (4.11) 14.65 (3.83) 0.862 13.60 (3.58) 14.10 (3.70) 13.36 (4.28) 14.70 (3.77) 13.40 (4.07) 14.22 (3.60)
Factor 3 15.24 (2.89) 15.10 (3.16) 0.878 14.15 (3.37) 15.40 (2.89) 14.24 (3.09) 15.85 (3.28) 14.24 (2.89) 15.35 (2.91)
Factor 4 7.28 (2.11) 7.75 (2.47) 0.495 6.40 (1.98) 7.75 (2.51) 7.24 (2.11) 8.10 (2.07) 6.84 (1.75) 7.46 (2.34)
Parent–child interactions
Child
Narrowed interest 4.60 (4.40) 4.70 (5.57) 0.918 4.56 (4.25) 4.14 (5.79) 4.35 (3.88) 5.04 (6.04) 4.54 (4.74) 7.45 (8.64)
Stereotyped behaviourb 4.96 (3.60) 4.85 (4.64) 0.724 2.36 (2.74)c 4.83 (4.17)c 2.47 (2.89) 3.51 (3.03) 2.58 (3.22) 3.49 (4.38)
Sensory 4.60 (4.65) 5.05 (5.16) 0.647 5.28 (5.98) 4.19 (4.73) 3.44 (3.14) 3.76 (7.63) 3.57 (4.50) 3.31 (5.53)
Motor 2.24 (3.05) 3.20 (4.70) 0.449 3.00 (3.45) 1.75 (1.92) 2.88 (3.98) 1.35 (1.87) 3.43 (5.88) 1.85 (2.50)
Higher level words 2.20 (2.20) 3.80 (5.24) 0.548 2.60 (4.08) 3.50 (4.35) 3.52 (5.78) 2.63 (2.81) 1.60 (2.22) 3.54 (4.50)
Lower level sounds 4.20 (6.01) 3.10 (3.20) 0.839 3.44 (5.35) 3.15 (4.62) 2.64 (2.91) 1.95 (1.99) 2.41 (3.33) 2.02 (1.89)
Parent
Non-intervening 4.24 (3.26) 4.30 (2.81) 0.997 4.00 (3.74) 5.52 (4.21) 3.61 (3.58) 3.72 (2.61) 4.42 (3.20) 5.24 (5.65)
Preventing 0.76 (1.45) 0.98 (1.54) 0.633 1.32 (3.05) 0.25 (0.64) 1.00 (1.68) 1.25 (4.29) 0.52 (1.20) 0.90 (2.73)
Engaging 4.76 (3.13) 6.33 (6.71) 0.481 4.44 (3.32) 5.51 (4.21) 4.36 (4.06) 5.08 (2.87) 5.11 (3.68) 5.13 (4.12)
Distracting/developinga 7.12 (5.31) 7.83 (4.57) 0.518 6.64 (5.87) 6.98 (4.70) 7.19 (3.88)c 4.57 (2.86)c 5.54 (4.51) 6.65 (5.49)
Specific target RRB* – – – 3.83 (1.08) 4.38 (1.23) 3.46 (1.38) 3.98 (1.58) 3.43 (1.59) 4.28 (1.34)
Both target RRB* – – – 3.91 (1.18) 4.12 (1.52) 3.55 (1.26) 3.69 (1.56) 3.52 (1.52) 4.01 (1.62)
Factor 1 – – – 4.25 (0.88) 3.85 (1.32) 4.45 (1.36) 3.85 (1.27) 4.70 (1.11) 4.45 (1.08)
Factor 2 – – – 4.00 (1.21) 3.54 (1.98) 3.63 (1.34) 3.54 (1.39) 3.75 (1.96) 4.38 (1.76)
Factor 3 – – – 3.69 (1.06) 4.44 (1.07) 2.69 (1.29)c 4.59 (1.49)c 3.03 (1.32)c 5.03 (0.81)c
Factor 4 – – – 3.58 (1.53) 5.33 (0.38) 3.92 (1.04) 4.33 (0.76) 2.58 (2.10) 4.00 (1.73)
Teacher total RBQ-2 33.79 (6.36) 35.25 (7.88) 0.499 33.90 (8.79) 35.13 (8.01) 32.11 (6.32) 33.56 (8.26) 31.25 (7.27) 33.63 (9.06)
Parent self-efficacya 2.94 (1.06) 2.54 (0.79) 0.173 3.38 (0.97)c 2.64 (0.77)c 3.41 (1.07)c 2.64 (0.74)c 3.61 (0.91)c 2.40 (0.70)c
Specific behaviour = Level of improvement in the specific behaviour worked on in the immediate intervention group matched with a behaviour chosen from delayed groups that was the same
category. Both behaviours = Level of improvement in both behaviours reported in both groups
* Factor 1—repetitive motor movement; Factor 2—rigidity, adherence to routine and insistence on sameness; Factor 3—preoccupation with restricted pattern of interest, limited play; Factor
4—unusual sensory interests
a
Significant interaction effects
b
Significant effects of time within both immediate and delayed groups
c
Significant group difference
J Autism Dev Disord (2015) 45:3168–3182
J Autism Dev Disord (2015) 45:3168–3182 3177

Fig. 2 Change in target behaviour vignette Fig. 3 Parent rated total RBQ-2 by visit by group

4.87, interquartile range 2.19). Similarly at TP 3, the im- at first time point was -2.83 (95 % CI -6.23, 0.57) not a
mediate group had greater improvement (median 3.00, in- clinically meaningful change. As can be seen in Table 3,
terquartile range 1.88) than the delayed (median 5.12, Factor 2 ‘Rigidity, adherence to routine and insistence on
interquartile range 0.88; t(15) = 3.71, p = 0.002). Change sameness’ appears to show a trend in the hypothesised di-
in other factor behaviours (Repetitive motor movements; rection; however, it was not statistically significant.
rigidity, adherence to routine and insistence on sameness;
unusual sensory interests) was not significant. Measurement of RRB: Parent–Child Interactions

Measurement of RRB: Parent RBQ-2 A square root transformation was used to normalise the
category ‘stereotyped behaviour and non-functional inter-
The dependent variable, RBQ-2 total score at each time ests’, while a log transformation was used on the other
point was analysed using a mixed model with variation categories of RRB, to allow parametric analysis. For one
between parents and variation between time points in- type of RRB ‘stereotyped behaviour and non-functional
cluded as random effects; baseline RBQ-2 was included as interests’ there was a significant main effect for time, F(3,
a covariate; and the difference between groups, a linear 129) = 3.35, p = 0.021, g2p = 0.072 (a medium effect
trend across time points together and their interaction were size) (see Table 2). There was also a reduction in the fre-
included as fixed effects. Maximum likelihood estimation quency of this RRB from baseline (m = 4.91, SD = 4.04)
procedures were used to generate interval estimates of the to TP 2 [m = 2.93, SD = 2.97, t(44) = 2.78, p = 0.008],
impact of the intervention. From the box plot, there was and from baseline to TP 3 [m = 2.98, SD = 3.76,
little evidence that the difference between groups changed t(44) = 2.80, p = 0.008], and a significant difference be-
over time (Fig. 3). tween groups at TP1, the immediate group had lower levels
The total RBQ-2 score across all participants and occa- of this RRB at TP1 than the delayed group, t(43) = 2.47,
sions was 39.9 (95 % CI 37.8, 42.0). The rate of change in p = 0.017. Analysis of the individual parent strategies
immediate group was 0.07 units per time point (95 % CI indicated a significant change in ‘distracting/developing’
-0.87, 1.00). The rate of change in delayed group was -0.66 strategy, as hypothesised. The data for this strategy was
units per time point (95 % CI -1.71, 0.38). The difference transformed using a square root transformation. The in-
between these figures was 0.73 (95 % CI -0.67, 2.14), teraction was significant, F(3, 129) = 2.82, p = 0.042,
indicating little evidence of a change over time in total RRB g2p = 0.062 (a medium effect size). At TP2, there was an
in either group or that the rate of change of total score dif- increase in the number of these strategies displayed in the
fered between groups. Estimated difference between groups immediate group, t(43) = 2.39, p = 0.021.

123
3178 J Autism Dev Disord (2015) 45:3168–3182

Table 3 Final model of RBQ-2: estimated parameters with 95 % confidence intervals

Variable Difference at 1st outcome Change in score per visit Difference in rate of change
assessment per visit
Control group Intervention group
m 95 % CI m 95 % CI m 95 % CI m 95 % CI

Factor 1: motor movements -0.74 -2.22 0.74 -0.45 -0.88 -0.02 0.06 -0.33 0.45 0.51 -0.07 1.09
Factor 2: rigidity -0.40 -2.02 1.22 0.06 -0.39 0.51 -0.10 -0.50 0.30 -0.16 -0.77 0.45
Factor 3: preoccupation -1.57 -3.20 0.07 -0.03 -0.48 0.43 0.05 -0.36 0.45 0.07 -0.54 0.68
Factor 4: sensory interest -1.31 -2.26 -0.37 -0.14 -0.43 0.14 0.22 -0.04 0.48 0.36 -0.02 0.75
Total RBQ2 score -0.14 -0.31 0.03 -0.03 -0.09 0.02 0.00 -0.04 0.05 0.31 -0.03 0.11

Parent Self-Efficacy

Self-efficacy was analysed by fitting a sequence of mixed

models with time point scores nested within parents. Var-
iation between parents and variation between time point
scores within parents were included in the model as ran-
dom effects with normal distribution. Baseline self-efficacy
was included as a covariate. The difference between groups
was included as a fixed effect. The boxplot of the raw data
indicates the intervention had a positive effect on parent
self-efficacy (Fig. 4). There was a significant variation
between parents ru = 0.85 (95 % CI 0.68, 1.07) and be-
tween outcome assessment visits within parents, re = 0.49
(95 % CI 0.42, 0.56). Baseline self-efficacy was a sig-
nificant predictor of reported self-efficacy at follow-up
outcome assessment visits. By including baseline self-ef-
ficacy, we explained some of the variation between parents,
ru decreased to 0.71 (95 % CI 0.56, 0.90). Fitting an im-
pact of the intervention, our results suggest that the inter-
vention increased self-efficacy by 0.74 (95 % CI 0.34,
1.14). Fig. 4 Parent self-efficacy by visit by group

Teacher RBQ-2
established that families were willing to be recruited and
There were no significant changes over time in teacher- randomised, parents found the format and content of the
rated total RBQ-2 scores or factors of the teacher RBQ-2. intervention acceptable and attrition was small. Group
However not all RRB behaviours identified by parents are leaders were able to deliver the programme in a consistent
likely to occur at school e.g. rigid restricted bath time fashion as evidenced by a high degree of fidelity to the
routine. treatment manual. The properties of the outcome measures
have also been assessed.
The study was not designed as a fully powered trial.
Discussion However, there is preliminary evidence that this new
MRBÓ parent-group intervention led to greater gains in
To our knowledge, this is the first randomized controlled parent self-efficacy and improvement in overall interaction
trial of a parent group intervention designed to help parents between parent and child (CGI-I). Although 30 % of
recognise, understand and manage RRB in young children children in the intervention group being classified as
with ASD. Results from this pilot trial support the feasi- definitely ‘improved’ at 6 months may appear somewhat
bility and acceptability of both this new intervention, and low, nevertheless another 26 % made some improvement,
the recruitment and research procedures. This pilot trial and for families any change in difficult-to-manage

123
J Autism Dev Disord (2015) 45:3168–3182 3179

behaviours may be of clinical significance; 75 % of the also asked to complete the RBQ2, to provide a measure of
delayed group were observed to make no change at all. RRB in a different setting. In the previous development
There is also a suggestion that for parents in the inter- work, parents had identified the importance of recording
vention group who reported their chosen target RRB to be the levels of and any changes in RRB in other social
of the category ‘preoccupations with restricted patterns of contexts (such as nursery/education setting). Teachers were
interest and limited play’ (Factor 3), there was evidence of able to complete the teacher-version of the RBQ-2 but
a significant improvement in independently rated target indicated that some items were not relevant to the educa-
behaviour vignettes compared to the delayed group. tion setting (e.g. ‘insist on wearing the same clothes or
Equipping parents with environmental strategies in which refuses to wear new clothes’). Education staff were also
they can engage, distract or develop their child’s RRB may aware that they may not see the same types of RRB that
have more utility when the target behaviours of concern are parents reported in the home. In the development of
rigidity, routines or preoccupations. Perhaps behaviours therapeutic interventions for children with ASD, the
that may have a cognitive component or be more likely to recognition of the difficulties many children experience
be related to executive control, may be the most appro- transferring newly acquired skills from one setting to an-
priate for parents to attempt to address, compared to some other (generalisation) is widely recognised and reported
of the repetitive sensorimotor behaviours that may repre- (e.g. Parsons and Mitchell 2002; Swettenham 1996; Green
sent separate neurological pathways regulated by internal et al. 2010). A recently completed report on outcome
sensory mechanisms. measures has also highlighted the need for a questionnaire
MRB is a parent-implemented intervention. Central to appropriate for young children with ASD which can be
its ethos is the assumption that parents can increase their used to measure RRB across settings (McConachie et al.
understanding and management of RRB which in turn will 2015).
build parents’ confidence in their parenting skills. This The target behaviour vignette was practical, reliably
study reports improvement in parent self-efficacy that was rated, and sensitive to change. One important advantage of
maintained at 6-month outcome measurement. Given that this outcome measure is that it captures the idiographic
repetitive behaviours are reported to be some of the most account of the particular behaviour that is most relevance
stressful of ASD behaviours to manage (Bishop et al. to the family, whilst at the same time the vignette rating
2007), it is a strength of this parent-group intervention that procedure provides a method for obtaining quantitative
parents reported large gains in confidence in managing data (change score) independent of parent bias and judged
their child’s RRB. by ‘blind’ expert raters. The results are suggestive of an
It is possible that teaching parents how to consider RRB impact of the intervention in particular for rigidity of
using a functional analytic framework approach enabled routines and preoccupations leading to limited opportuni-
parents of young children with ASD to both consider the ties for play (Factor 3) (Hine and Wolery 2006; Honey
communicative function of their child’s RRB and imple- et al. 2007).
ment tailored strategies to manage them, which in turn may Turning to the parent child interactions, although this
have led to the improved outcomes reported. measure provided a way of directly observing RRB that
could be blind-rated, it is unlikely that for each child
Utility of Outcome Measures the behaviours of concern would be elicited in a 10-min
play interaction. This measure appears to be most
A strength of the study is that RRB and the possible impact useful with directly observable RRB such as stereo-
of the intervention were measured in several different typical motor movements and sensory behaviours and
ways. The CGI-I (a ‘blind’ independent measure regularly less so with preoccupations, routines and sameness. The
used in the evaluation of medication and other interven- parent self-efficacy scale was found to be an appropri-
tions in child mental health research) showed promise as an ate measure suitable for this young, ASD population
appropriate overall outcome measure for detecting treat- and their parents. In this study the measure identified a
ment response. clear treatment effect for those in the immediate inter-
Three different approaches to the measurement of RRB vention group.
were included in this study. In line with previous research
(Leekam et al. 2007a, b; Lidstone et al. 2014), the RBQ-2 Limitations
was found to be an acceptable measure for use with parents
of children with ASD. However the use of a total score of The major limitation of the current study is sample size;
RRB was not apparently sensitive to change. Despite this although it was adequate for a phase-II pilot, definitive
limitation, the parent RBQ2 may have utility in future conclusions about efficacy of the intervention cannot be
studies exploring generalisation over time. Teachers were drawn from the data.

123
3180 J Autism Dev Disord (2015) 45:3168–3182

Implications for Further Research part in this pilot RCT. Without their generosity this project would not
have been possible. We would also like to thank the course trainers
Jan Raine, Linda Barker, Joyce Frater, Adele Fearon, and Carol
This new parent group intervention MRB, has been devel- Richardson. We thank Kathryn Burn, Deborah Garland, Fiona
oped in collaboration with parents of young children with Gaultier, Anna Hodgson, Emma Honey, Kelly McGurk, Rachel
ASD utilising an ‘active’ research methodology. It is a Proud, Laura Rosby, Dr. Ellie Smith and Laura Surley for their
psychoeducational intervention that incorporates evidence contributions to the research. Finally we acknowledge the expertise of
Sue Leach (previously employed by the UK Mental Health Research
based knowledge of ASD, the principles of a functional Network), members of the MRB Steering group (Simon Douglas, Dr.
analysis approach to understanding behaviours combined Mark Baggot) and the clinicians who helped recruit and support
with the mutual support of group peer learning. Parent in- families with ASD during this research. This paper describes inde-
volvement has informed every aspect of the design of the pendent research commissioned by the National Institute for Health
Research (NIHR) under the Research for Patient Benefit programme
intervention, from training materials to the acceptability of (PB-PG-1010-23305). The views expressed are those of the authors
the devised programme, piloting the feasibility of outcome and not necessarily those of the NHS, the NIHR or the Department of
measures and the use of video feedback as a strategy for Health.
working on an agreed target RRB, through to contributing
to the design of the research protocol. This collaborative
‘active’ research model of working jointly with parents has References
in the opinion of all those involved in this study been of
Ahrens, E. N., Lerman, D. C., Kodak, T., Worsdell, A. S., & Keegan,
great mutual benefit. Parent involvement has brought new C. (2011). Further evaluation of response interruption and
perspectives and helped ensure that both the development of redirection as treatment for stereotypy. Journal of Applied
the intervention and the research is relevant to the concerns Behavior Analysis, 44, 95–108. doi:10.1901/jaba.2011.44-95.
and needs of parents and families of young children with American Psychiatric Association. (2013). Diagnostic and statistical
manual of mental disorders (5th ed.). DC: Washington.
ASD. Planning research into complex interventions such as Arnold, L. E., et al. (2003). Parent-defined target symptoms respond
parent-mediated interventions should address service user to risperidone in RUPP autism study: Customer approach to
research priorities; this can be facilitated through adoption clinical trials. Journal of the American Academy of Child and
of an ‘active’ research model. Future research also needs to Adolescent Psychiatry, 42, 1443–1450. doi:10.1097/00004583-
200312000-00011.
identify a reliable measure of observable RRB, capturing all Baird, G., Simonoff, E., Pickles, A., Chandler, S., Loucas, T.,
topographies of RRB, across a range of ages and cognitive Meldrum, P., & Charman, T. (2006). Prevalence of disorders of
abilities. Promising results have been found using the the autism spectrum in a population cohort of children in South
DOBRA coding system (Boyd et al. 2010); however, im- Thames: The Special Needs and Autism Project (SNAP). Lancet,
368, 210–215. doi:10.1016/S0140-6736(06)69041-7.
plementation of this coding system needs to occur in a Barber, A. B., Wetherby, A. M., & Chambers, N. W. (2012). Brief
larger, fully powered study to assess its sensitivity. It is report: Repetitive behaviors in young children with autism
desirable that any intervention to target RRB should con- spectrum disorder and developmentally similar peers: A follow
sider how best to work across different social contexts up to Watt et al. (2008). Journal of Autism and Developmental
Disorders, 42, 2006–2012.
(home, educational and other setting); this will also require Bishop, S. L., Richler, J., Cain, A. C., & Lord, C. (2007). Predictors
a tool to measure change in RRB across settings. of perceived negative impact in mothers of children with autism
spectrum disorder. American Journal of Mental Retardation,
112, 450–461. doi:10.1352/0895-8017(2007).
Boyd, B. A., McDonough, S. G., Rupp, B., Khan, F., & Bodfish, J. W.
Conclusion (2011). Effects of a family-implemented treatment on the
repetitive behaviors of children with autism. Journal of Autism
This study has shown that the Managing Repetitive Be- and Developmental Disorders, 41, 1330–1341. doi:10.1007/
haviours: MRBÓ programme was acceptable to parents and s10803-010-1156-y.
Boyd, B. A., Rupp, B., & Bodfish, J. W. (2010). Direct observation of
feasible to implement. The estimates of variability in out- repetitive behaviors in autism. Unpublished manual: University
come measures indicated large gains in parental self-effi- of North Carolina at Chapel Hill.
cacy and suggestive improvements in overall functioning Carter, A. S., Messinger, D. S., Stone, W. L., Celimli, S., Nahmias, A.
for the child including an improvement in the target RRB S., & Yoder, P. (2011). A randomized controlled trial of Hanen’s
‘More Than Words’ in toddlers with early autism symptoms.
in the immediate intervention group. A multi-site trial is Journal of Child Psychology and Psychiatry and Allied Disci-
now required to establish the efficacy of this intervention, plines, 52, 741–752. doi:10.1111/j.1469-7610.2011.02395.x.
which should be powered to examine potential moderators Cohen, J. (1988). Statistical power analysis for the behavioral
and mediators of treatment effects. sciences. Hillsdale, NJ: Lawrence Erlbaum Associates.
Constantino, J. N., & Gruber, C. P. (2005). Social Responsiveness
Scale (SRS). Los Angeles, CA: Western Psychological Services.
Acknowledgments We are very grateful to the many parents of Cunningham, A. B., & Schreibman, L. (2008). Stereotypy in autism:
children with ASD in the North East of England who have supported The importance of function. Research in Autism Spectrum
the development of the MRBÓ group-based intervention and taken Disorders, 2, 469–479. doi:10.1016/j.rasd.2007.09.006.

123
J Autism Dev Disord (2015) 45:3168–3182 3181

Dunlap, G., Dyer, K., & Koegel, R. L. (1983). Autistic self- in children with autism spectrum disorder. Research in Autism
stimulation and intertrial interval duration. American Journal of Spectrum Disorders, 8(2), 82–92.
Mental Deficiency, 88, 194–202. Lord, C., Rutter, M., DiLavore, P. C., et al. (2012). Autism diagnostic
Gordon, C. T. (2000). Commentary: Considerations on the pharma- observation schedule, second edition: ADOS-2. Torrance:
cological treatment of compulsions and stereotypies with Western Psychological Services.
serotonin reuptake inhibitors in pervasive developmental disor- Magiati, I., Tay, X. W., & Howlin, P. (2014). Cognitive, language,
ders. Journal of Autism and Developmental Disorders, 30, social and behavioural outcomes in adults with autism spectrum
437–438. disorders: A systematic review of longitudinal follow-up studies
Green, J., Charman, T., McConachie, H., Aldred, C., Slonims, V., in adulthood. Clinical Psychology Review, 34, 73–86. doi:10.
Howlin, H., et al. (2010). Parent-mediated communication- 1016/j.cpr.2013.11.002.
focused treatment in children with autism (PACT): A ran- McConachie, H., Parr, J., Glod, M., Hanratty, J., Livingstone, N.,
domised controlled trial. The Lancet, 375(9732), 2152–2160. Oono, I. P., et al. (2015). Systematic review of tools to measure
doi:10.1016/S0140-6736(10)60587-9. (eScholarID:81710). outcomes for young children with autism spectrum disorder.
Greenberg, J. S., Seltzer, M. M., Hong, J., & Orsmond, G. I. (2006). HTA Report (forthcoming).
Bidirectional effects of expressed emotion and behavior prob- Militerni, R., Bravaccio, C., Falco, C., Fico, C., & Palermo, M. T.
lems and symptoms in adolescents and adults with autism. (2002). Repetitive behaviors in autistic disorder. European Child
American Journal of Mental Retardation, 111, 229–249. doi:10. and Adolescent Psychiatry, 11, 210–218. doi:10.1007/s00787-
1352/0895-8017(2006). 002-0279-x.
Guy, W. (1976). Clinical global impression. In: ECDEU assessment Miller, N., & Neuringer, A. (2000). Reinforcing variability in
manual for psychopharmacology, revised (pp. 218–222). Rock- adolescents with autism. Journal of Applied Behavior Analysis,
ville, MD: National Institute of Mental Health. 33, 151–165. doi:10.1901/jaba.2000.33-151.
Harrop, C., McConachie, H., Emsley, R., Leadbitter, K., & Green, J. Mulligan, S., Healy, O., Lydon, S., Moran, L., & Foody, C. (2014).
(2014). Restricted and repetitive behaviors in autism spectrum An analysis of treatment efficacy for stereotyped and repetitive
disorders and typical development: Cross-sectional and longitu- behaviors in autism. Journal Autism Developmental Disorders,
dinal comparisons. Journal of Autism and Developmental 1, 143–164.
Disorders, 44(5), 1207–1219. Oono, I. P., Honey, E. J., & McConachie, H. (2013). Parent-mediated
Hine, J., & Wolery, M. (2006). Using point-of-view video modeling early intervention for young children with autism spectrum
to teach play to preschoolers with autism. Topics in Early disorders (ASD). The Cochrane Database of Systematic Reviews,
Childhood Special Education, 26(2), 83–93. 4, CD009774. doi:10.1002/14651858.CD009774.pub2.
Honey, E., Leekam, S., McConachie, H., & Turner, M. (2007). Parsons, S., & Mitchell, P. (2002). The potential of virtual reality in
Repetitive behaviour and play in typically developing children social skills training for people with autistic spectrum disorders.
and children with autism spectrum disorders. Journal of Autism Journal of Intellectual Disability Research, 46, 430–443.
and Developmental Disorders, 37(6), 1107–1115. Rapp, J. T., & Vollmer, T. R. (2005). Stereotypy I: A review of
Honey, E., Rodgers, J., & McConachie, H. (2012). Measurement of behavioral assessment and treatment. Research in Developmen-
restricted and repetitive behaviour in children with autism tal Disabilities, 26, 527–547. doi:10.1016/j.ridd.2004.11.005.
spectrum disorder: Selecting a questionnaire or interview. Ritchie, J., & Spencer, L. (1994). Qualitative data analysis for applied
Research in Autism Spectrum Disorders, 6, 757–776. doi:10. policy research. In A. Bryman & R. G. Burgess (Eds.), Analysing
1016/j.rasd.2011.10.011. qualitative data (pp. 173–194). London: Routledge.
Julious, S. A. (2005). Sample size of 12 per group rule of thumb for a Rodgers, J., Glod, M., Connolly, B., & McConachie, H. (2012). The
pilot study. Pharmaceutical Statistics, 4, 287–291. relationship between anxiety and repetitive behaviours in autism
Kuhn, D. E., Hardesty, S. L., & Sweeney, N. M. (2009). Assessment spectrum disorder. Journal of Autism and Developmental
and treatment of excessive straightening and destructive be- Disorders, 42, 2404–2409. doi:10.1007/s10803-012-1531-y.
haviour in an adolescent diagnosed with autism. Journal of Shafi, D. (2009). How parents of young children with autism manage
Applied Behavior Analysis, 42, 355–360. doi:10.1901/jaba.2009. repetitive behaviour: An observational study. Unpublished Clin.
42-355. Psy. D. thesis, Newcastle University.
Lancaster, G. A., Dodd, S., & Williamson, P. R. (2004). Design and Sofronoff, K., & Farbotko, M. (2002). The effectiveness of parent
analysis of pilot studies: Recommendations for good practice. management training to increase self-efficacy in parents of
Journal of Evaluation in Clinical Practice, 10, 307–312. doi:10. children with Asperger syndrome. Autism: The International
1111/j.2002.384.doc.x. Journal of Research and Practice, 6, 271–286.
Leekam, S. R., Nieto, C., Libby, S. J., Wing, L., & Gould, J. (2007a). Sparrow, S. S., Cicchetti, V. D., & Balla, A. D. (2005). Vineland
Describing the sensory abnormalities of children and adults with adaptive behavior scales (2nd ed.). Circle Pines, MN: American
autism. Journal of Autism and Developmental Disorders, 37, Guidance Service.
894–910. doi:10.1007/s10803-006-0218-7. Swettenham, J. (1996). Can children with autism be taught to
Leekam, S. R., Prior, M. R., & Uljarevic, M. (2011). Restricted and understand false belief using computers? Journal of Child
repetitive behaviors in autism spectrum disorders: A review of Psychology and Psychiatry and Allied Disciplines, 37, 157–165.
research in the last decade. Psychological Bulletin, 137, Townsend, P., Phillimore, P., & Beattie, A. (1988). Health and
562–593. doi:10.1037/a0023341. deprivation: Inequality and the North. London: Croom Helm.
Leekam, S., Tandos, J., McConachie, H., Meins, E., Parkinson, K., Turner, M. (1995). Repetitive behaviour and cognitive functioning in
Wright, C., et al. (2007b). Repetitive behaviours in typically autism. Unpublished Ph.D. thesis, University of Cambridge,
developing 2-year-olds. Journal of Child Psychology and Cambridge.
Psychiatry and Allied Disciplines, 48, 1131–1138. doi:10.1111/ Watt, N., Wetherby, A. M., Barber, A., & Morgan, L. (2008).
j.1469-7610.2007.01778.x. Repetitive and stereotyped behaviors in children with autism
Lidstone, J., Uljarevic, M. (joint first authors), Sullivan, S., Rodgers, spectrum disorders in the second year of life. Journal of Autism
J., McConachie, H., Freeston, M., et al. (2014). Relations among and Developmental Disorders, 38, 1518–1533. doi:10.1007/
restricted and repetitive behaviours, anxiety and sensory features s10803-007-0532-8.

123
3182 J Autism Dev Disord (2015) 45:3168–3182

Wetherby, A., & Prizant, B. (2002). Communication and symbolic disorders: Background, inter-rater reliability and clinical use.
behavior scales developmental profile—First normed edition. Journal of Child Psychology and Psychiatry and Allied Disci-
Baltimore, MD: Paul H. Brookes. plines, 43, 307–325.
Wing, L., Leekam, S. R., Libby, S. J., Gould, J., & Larcombe, M. World Health Organisation. (1992). International Classification of
(2002). The diagnostic interview for social and communication Diseases (ICD 10. Geneva: World Health Organisation.

123
Copyright of Journal of Autism & Developmental Disorders is the property of Springer
Science & Business Media B.V. 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.