i) conduct a systematic review of all the pre-school intervention literature in autism, including a detailed quality rating of each study, the type of intervention that is being tested and classification of outcome measures used;
ii) to undertake a meta-analysis of methodologically adequate studies, which will allow for the first time comparison of different approaches to intervention on comparative outcome measures.
We will use the following search terms to search all trials registers and databases: “autism” , “autism spectrum disorder”, “ASD”, “high function autism”, “high function ASD”, “Aspergar syndrome”, “pervasive developmental disorder”, “PDDNOS”, “intervention”, “communication”, "interpersonal", "speech", "interaction", "synchrony", “relationship”, “language”, “social” and “development”, "behavior therapy", "intensive behavioral intervention". Their search will be limited by age group from 0 to 6 years old and “randomized controlled trial”. Reference lists from identified trials and review articles will be manually scanned to identify any other relevant studies. The clinicaltrials.gov and the Cochrane Library website will also be searched for randomized trials that were registered as completed but not yet published.
Relevant studies will be identified by searching the following data sources: PsycINFO (from 1956 to January, 2011), Medline via Ovid (from 1950 to January, 2011), ERIC (from 1950 to January, 2011) and the Cochrane database . Reference lists from identified trials and review articles will be manually scanned to identify any other relevant studies. The clinicalTrials.gov and the Cochrane Library website will also be searched for randomized trials that were registered as completed but not yet published.
Types of study to be included
1. Participants comprise preschool children with a diagnosis of ASD or pervasive developmental disorder (PDD).
2. Randomized controlled trials.
3. Interventions delivered to the parents/guardians and/or directly to the child, by special educators, teachers, speech pathologists, psychologists, or other allied health professional students will be included.
4. Studies carried out while the children were at a preschool age between 0 and 6 years.
5. The control group will be those who did not received early intervention for autism.
6. Studies where the judgements of risk of bias fall into categories A or B in the Cochrane Collaboration tool for assessing risk of bias.
1. The study was not primary research on preschool autism.
2. The study was not an RCT.
3. The study did not assess a cognitive/behavioural intervention for preschool autism.
4. The study did not report adequately on any measurable data for health related outcomes relevant to the review.
5. The study design was not a randomized controlled trial.
6. The intervention used alternative medicine.
7. The intervention was a pharmacological one.
8. The intervention was not classified into behavioural, multimodal developmental or communication-focused model.
9. The control group received some early intervention for children with autism.
10. Studies where the judgements of risk of bias fall into categories C in the Cochrane Collaboration tool for assessing risk of bias.
Condition or domain being studied
Preschool children aged 0-6 years old with autism spectrum disorder.
Inclusion: Preschool children aged 0 to 6 with autism spectrum disorder.
Exclusion: Children aged over 6 and elderly people with autism spectrum disorder.
We classify interventions for preschool autism in three groups; i) behavioural interventions – based essentially on learning theory and on applied behaviour analysis; ii) communication-focused interventions, targeting social communication impairment, as the core symptom of autism; iii) multimodal developmental interventions targeting a range of aspects of children’s development.
The control will be those who did not receive early intervention for autism.
Autism behavioural symptoms: Qualitative impairment in social interaction; qualitative impairments in communication; restricted, repetitive, and stereotyped patterns of behaviour, interest, and activities (DSM-4-TR). These are the definitional symptoms of the disorder and key indicators of psychopathology. The measures used for the primary outcomes will be ones which were developed to measure outcomes specific to autism, e.g. Autism Diagnostic Observation Schedule or Autism Diagnostic Interview-Revised.
Non-specific developmental outcomes. These are not directly related by definition to autism diagnosis but are used in some studies as substitute outcomes – examples are IQ and cognitive development.
Intermediate outcomes relevant to the known development of autism – which might be candidates for surrogate endpoints. These outcomes are often defined as the proximal targets of intervention approaches from a developmental perspective. Examples are: measures of joint attention, parent-child interaction, imitation ability, symbolic play, social communication in an interactive setting, receptive language, expressive language.
Data extraction, (selection and coding)
Two of the authors, Y.T. and Y.H. independently will review abstracts of potentially the relevant studies. This will be followed by a consensus discussion with J.G. The quality of the RCTs will be coded independently by Y.T. and Y.H. and disagreement will be resolved by consensus discussions.
Risk of bias (quality) assessment
Risk of bias will be assessed by two independent review authors (Y.T. and Y.H.) and disagreements will be resolved by negotiation with a third review author (J.G.). We will use the Cochrane Collaboration tool for assessing risk of bias in these areas. The assessed risk of bias in studies will include in the following domains: sequence generation; allocation concealment; blinding; incomplete outcome data; selective outcome reporting; other sources of bias. The process will involve recording the appropriate information for each study (for example describing the method used to conceal allocation in detail) and evaluating whether there is risk of bias in that area (for example, was allocation adequately concealed). We will allocate studies to the three categories according to our judgement of each area or potential risk of bias: A. Low risk of bias; B. Moderate (or unclear) risk of bias; C. High risk of bias. Only studies where the judgements of risk of bias falls into categories A or B will be included in subsequent analyses.
Strategy for data synthesis
The categories of outcome measure mentioned above differ conceptually in important ways, and have been used in a systematic different way across trials of the different intervention types identified above. Our review aims to make comparison across these different types of intervention study, thus we will standardize and synthesize the various categories of outcome measure using an inverse variance method. The measures used for outcome are varied between studies and the standardized data will be heterogeneous. We will use a random effects model for the analyses, comparing type of intervention model effectiveness for each outcome using a standardized mean difference.
Analysis of subgroups or subsets
Relevant subgroup analyses will include:
• Severity of autism at baseline. This is a crucial element in evaluating autism studies.
• SES and other demographic variables. Sampling bias and external validity of studies is an important consideration.
• Age of child
• Type of intervention (our 3 groups as above)
• Parent-mediated vs child-mediated intervention delivery. A key distinguishing point between different studies in the area.
• Cognitive ability at baseline.
Contact details for further information
Room 4.321, Psychiatry Research Group, 4th Floor (East), Jean McFarlane Building, University Place, University of Manchester and Manchester Academic Health Sciences Centre, Oxford Road, Manchester, M13 9PL, UK
Organisational affiliation of the review
Department of Child and Adolescent Psychiatry, University of Manchester and Manchester Academic Health Sciences Centre
Dr Yoshiyuki Tachibana, Department of Child and Adolescent Psychiatry, University of Manchester and Manchester Academic Health Sciences Centre Professor Jonathan Green, Department of Child and Adolescent Psychiatry, University of Manchester and Manchester Academic Health Sciences Centre Dr Yeonhee Hwang, Special Support Education Research Center, Tohoku Fukushi University Dr Richard Emsley, Health Methodology Research Group, Department of Biostatistics, University of Manchester
Dr Claire Hodkinson, The John Rylands University Library, The University of Manchester
Anticipated or actual start date
08 June 2011
Anticipated completion date
14 December 2012
This study was supported by ‘Tohoku University Young Scientist Dispatch Program’.
Conflicts of interest
Subject index terms status
Subject indexing assigned by CRD
Subject index terms
Behavior Therapy; Child Development Disorders, Pervasive; Child, Preschool; Humans;
Date of registration in PROSPERO
08 September 2011
Date of publication of this revision
08 September 2011
Stage of review at time of this submission
Piloting of the study selection process
Formal screening of search results against eligibility criteria
Risk of bias (quality) assessment
PROSPERO This information has been provided by the named contact for this review. CRD has accepted this information in good faith and registered the review in PROSPERO. CRD bears no responsibility or liability for the content of this registration record, any associated files or external websites.