For children and adolescents with developmental disorders, do psychosocial interventions delivered by non-specialist care providers in community settings produce greater benefits than standard care in developmental, daily skills, behavioral, and family outcomes?

Do specific variables (e.g., age, treatment type, setting, intensity) mediate outcome for psychosocial interventions delivered by non-specialist providers in community settings?

We will search MEDLINE, PsycINFO, EMBASE, CINAHL, and the Cochrane Central Register of Controlled Trials (CENTRAL) for studies published during or after 1992.

We will include randomized controlled trials and prospective group comparison quasi-experimental research designs; qualitative research, quasi-experimental pre/post comparisons, and single subject research designs will be excluded.

Developmental disorders, specifically intellectual disability with or without an autism spectrum disorder (that is, individuals with an IQ < 70 regardless of autism status and excluding individuals with autism spectrum disorders who have IQ > 70).

We will include children and adolescents (ages birth to 18) who have an intellectual disability and autism spectrum disorder; children who have developmental disabilities or autism spectrum disorders with IQ > 70 (on average) will be excluded.

We will include studies examining psychosocial interventions (e.g., cognitive adaptation, behavior analytic interventions, parent education programs) in which the direct treatment (therapy) services to the child were provided by non-specialist providers (e.g., general practitioner, nurse practitioner, teacher, social worker, parent, aide, lay community worker); studies examining other types of interventions (e.g., psychopharmacological agent) or studies in which a psychosocial intervention was provided by a specialist (e.g., psychiatrist, psychologist, psychiatric nurse) will be excluded.

We will include all studies comparing psychosocial intervention to a non-treatment control, a different treatment, or treatment as usual (e.g., standard care); studies with no comparison group or with a comparison in which two arrangements of one treatment method are compared will be excluded.

We will include studies in which the direct treatment services provided to the child or adolescent were delivered in community settings including homes, clinics, and schools. Studies in which treatment was provided at university clinics, and psychiatric hospitals will be excluded except for parent training programs in which the training was delivered in a clinical or university setting (these will be included as the parent is expected to provide the intervention to their child in the home setting).

Primary outcomes

Daily skills (e.g., adaptive behavior).

We will compare outcomes for the treatment and comparison groups after treatment has concluded or at a time point in which the comparison group has not yet received treatment.

Secondary outcomes

Development (e.g., IQ, developmental progression, language).

Severity of behavioral symptoms.

Educational outcomes.

Family life (e.g., parent stress, family functioning, parenting skills).

We will compare outcomes for the treatment and comparison groups after treatment has concluded or at a time point in which the comparison group has not yet received treatment.

All variables will be extracted independently by two coders and disagreements will be resolved through consensus resolution and use of a third opinion if a consensus cannot be reached. We will code16 variables including: study location, research design, child diagnostic characteristics, mean pretreatment chronological age, mean pretreatment level of functioning, outcome measures, study reported results, effect sizes (standardized mean difference using Cohen's d), treatment description, treatment setting, treatment format (instructional format), treatment density, total hours of treatment, treatment provider characteristics, training requirements, and treatment supervision requirements.

We have adapted the Cochrane Risk of Bias tool due to the inclusion of both randomized and non-randomized trials. We will assess eight items: sequence generation, allocation concealment, performance bias, detection bias, attrition, selective outcome reporting, protection against contamination, and baseline imbalance.

Due to the large variability we expect to find in type of psychosocial interventions and outcome measures we do not anticipate using meta-analytic techniques to synthesize the data; thus narrative synthesis is planned at this stage. We will create harvest plot matrixes (e.g., Crowther et al., 2011; Ogilvie et al., 2008) which will be created to help visual display treatment effects.

We will compare effects by type of intervention across key variables, including, but not limited to, methodological quality, age, severity of intellectual disability, type of disability, treatment provider, treatment setting, resource allocation. We will use the harvest plot matrices to help visually discriminate differences of effect across intervention type and these variables.

Brian Reichow

230 South Frontage Road

New Haven, CT 06520

United States

brian.reichow@yale.edu

Yale Child Study Center, World Health Organization, University of Verona

Dr Brian Reichow, Yale Child Study Center
Dr Chiara Servili, World Health Organization
Dr Taghi Yasamy, World Health Organization
Dr Shekhar Saxena, World Health Organization
Dr Corrado Barbui, University of Verona

Ms Jan Glover, Yale Medical Library

27 March 2012

30 June 2013

Autism Speaks, World Health Organization, Yale School of Medicine

None known

None

English

Italy, Switzerland, United States of America

Subject indexing assigned by CRD

Autistic Disorder; Child; Child Development Disorders, Pervasive; Humans; Intellectual Disability; Psychotherapy

None

26 July 2012

15 March 2013