Ludwig S and Harstall C in Mitchell K and Carr A,

2001,

Reviewed by Jane Barlow,
Institute of Health Sciences,
University of Oxford

Methodology

A computer-based search was conducted of a range of databases for critical reviews related to Autism using only the search word ‘autism’. The databases searched included the following: MEDLINE; Pre MEDLINE, EMBASE; Best Evidence; HTA; EED; DARE; Cochrane Database of systematic Reviews; ISTAHC database; HealthSTAR; Psycinfo; CINAHL; ERIC; Dissertation Abstractions; etc.

Only critical reviews were included in which the scientific validity of the primary research studies had been appraised using a specific set of criteria.

Critical Appraisal

Overall, this was a rigorously conducted systematic review. However, the primary studies included in the three systematic reviews are of variable quality and many have methodological weaknesses. See below for further details.

Key Findings

Three systematic reviews were included evaluating a number of comprehensive treatment programmes for young children including Lovaas Therapy, Rutgers Autism Programme, TEACCH programme, Denver Model, LEAP Programme, and the Autism Pre-School Programme (ECRI, 2000; BCOHTA, 2000; Smith, 1999).

ECRI (2000) included 14 studies of the effectiveness of a range of comprehensive treatment programmes.

BCOHTA (2000) included 4 studies evaluating the effectiveness of early intensive behavioural therapy (Lovaas therapy) and the TEACCH Programme (different primary studies were included by BCOHTA and ECRI).

Smith (1999) included 9 studies of behaviour programmes including Lovaas therapy, the TEACCH programme (not the same study as the one included in the above reviews) and the Denver model.

Lovaas Therapy

Lovaas therapy utilises time-intensive (40-hours per week) behavioural intervention techniques aimed at treating children with autism between the ages of 2-3 years. A central aspect of the therapy is teaching imitative skills, teaching of expressive and early abstract language and interactive play, and later early academic tasks, socialisation skills, cause and effect relationships.

ECRI (2000) included 5 studies of the effectiveness of Lovaas therapy, one of which was a follow-up study. A range of methodologies were used including pseudo randomised controlled study; single group pre and post design; non-randomised controlled design; case-control retrospective study. These studies were subject to some of the following methodological weaknesses:

• Small sample sizes
• Non random design
• No control group
• High drop-out and no intention-to-treat analysis
• Children on medications not excluded
• Poor treatment integrity
• Unrepresentative sample
• Possibility of maturation bias
• Treatment groups not treated equally
• Few standardised outcome instruments used
• Few independent observations of outcome
• Many skills were not assessed by the studies.

The results of these studies are not therefore very reliable.

One study showed normal functioning in 47% of the intervention group compared with none of the control group. The second study showed an average increase of 22 IQ points over intake on standardized tests at 2-year follow-up; the third study show no normal functioning in the treatment group, but 44% achieved IQ>49, and the fourth study showed no normal functioning in the treatment group but a mean IQ increase of 25 points.

Overall, all studies reported positive treatment outcomes. However, while children who participated in the Lovaas programme appeared to improve, methodological weakness in the study designs mean that it is not clear to what extent this was due to the Lovaas therapy.

The BCOHTA (2000) and Smith (1999) reviewed the same studies as above. Smith (1999) concludes that while methodological flaws hinder the ability to draw conclusions, two studies of Lovaas therapy show that children make major long-lasting improvements as a result of this treatment, and that a number of subsequent studies have also produced favourable results.

Rutgers Autism Programme

This is an early intensive behavioural programme that is provided on a full-time basis over the course of 2 years. It is similar to Lovaas therapy but differs in that it does not provide the staff to deliver the treatment.

ECRI (2000) included one study evaluating the effectiveness of this programme. This study was a single group pre-post design with a sample of 20 children and the ECRI conclude that the results of the study could not be used to answer any questions about effectiveness.

TEACCH Programme

This is a developmental programme that aims to maximise adaptation through structured teaching of new adaptive skills; develop environmental modifications to accommodate the child’s deficits; maintain close collaboration between teacher and parent; provide a continuity of structured teaching throughout life; prevent the development of further behavioural problems.

ECRI (2000) included 3 studies of the effectiveness of the TEACCH programme. These studies included a pseudorandomised controlled trial; single group design in which all children act as their own control; single group pre-post design. Overall, these studies were subject to some of the following methodological weaknesses:

• Small sample size
• No random allocation
• No control group
• Maturation bias possible
• Use of non-validated measures
• Effect of drop-outs not analysed
• Unrepresentative samples
• Did not control for extraneous events i.e. children on medication
• No information on diagnostic criteria

The results of these studies are not therefore reliable.

One study showed that the treatment group improved significantly more than the control group on scores on the Psychoeducational Profile. The second study showed that there were significant treatment effects with regard to parental guidance, appropriate child behaviour; inappropriate child behaviour and stress. The outcome for the third study (single group pre-post design) is not summarised.

Overall, although significant treatment effects were found with regard to imitation, perception, fine motor skills, gross motor skills and cognitive performance, the functioning of the control groups was not that different from the experimental group, and the meaning of the results in practical terms is not therefore, clear.

Denver Model

This is a developmentally based programme – a joint therapeutic and educational programme for autistic children aged 2 to 6 years. It aims to develop the capacity of the child and the treatment plan consists of goals and objectives, instructional plans and activities, and data collection. It focuses on the development of communication and play skills, sensory activities, personal independence, and reducing unwanted behaviours (using functional behavioural analysis, communication training, and positive teaching of alternatives and redirection). This is achieved through the use of teaching, elicitation and shaping.

ECRI (2000) included 3 studies of the effectiveness of the Denver model. All three studies were single group, pre-post tests and suffered some of the following methodological weaknesses:

• Small samples
• No random allocation
• No control group
• Did not exclude children on treatment
• Non-validated measures
• Missing data not accounted for

The results of these studies cannot therefore be regarded as reliable.

One study showed significant effects in five out of six areas including cognition, perceptual/fine motor, cognition, language, social/emotional, and gross motor. The second study showed significant treatment effects in cognition, perceptual/fine motor, social/emotional, and language skills. The results for the third study were not provided.

LEAP Programme

The LEAP programme is a comprehensive pre-school service system designed to meet the needs of both normal and autistic children. It is a developmental programme, which includes some behavioural techniques. The programme includes training for parents involved with the programme, which is administered 5 days per week, for 3 hours per day in a classroom setting. The intervention is provided to ten normal children and 3-4 autistic children. In addition to teachers, speech, occupational and physical therapists work with the children. The individualised curriculums aim to assist the learning of language and functional skills, independent play and work skills, social interaction skills, and adaptive behaviour. Positive behavioural interventions are used to address behaviour problems. The programme provides group-based instruction that involves individualised learning in an integrated environment.

The ECRI (2000) included only one study of the effectiveness of the LEAP programme. This was a single group, pre-post design of 13 children with autism. The ECRI concluded that the small sample size, the confounding effects of maturation and the poor generalisability of the sample precludes the possibility of using the study to answer questions about effectiveness.

Autism Pre-School Programme

This programme uses a variety of standard behavioural and language development methods and is similar to the Rutgers Autism Programme in that it operates primarily in a consultative role. Caregivers are taught by a multidisciplinary team to empathise with the child, perform a functional behavioural analysis and plan, and evaluate strategies for changing behaviour. The development of language and social skills is given priority over individual behavioural problems.

The ECRI (2000) review included one study evaluating the effectiveness of this programme. This was an RCT with a sample of 36 children with autism. It demonstrated strong construct validity and the results are generalisable to pre-school children with autism.

This study showed a statistically significant improvement in language development. No other significant results were obtained. The treatment period for this study was only 12 weeks and it is concluded that this was not a sufficiently long time to assess the effects of the programme on overall symptoms of autism.

Princeton Child Development Programme

This programme offers 27.5 hours per week of behavioural intervention in multiple settings. Only one review included a study evaluating this intervention (Smith, 1999) but no results are provided.

Implications for Social Work Practice

There is limited evidence concerning the comparative efficacy and effectiveness of the different treatments.

Intensive behavioural intervention programmes (up to 40 hours per week) are effective in improving some aspects of children’s functioning including imitation, perception, fine motor skills, gross motor skills and cognitive performance.

There is insufficient evidence to establish a relationship between amount (intensity and duration) of any intensive intervention treatment programmes and outcomes such as intelligence, language development, and behaviour.

Policy Implication

The evidence provided to date is preliminary and further research is needed before it will be possible to make any policy recommendations concerning the use of intensive treatments for young children with autism.