Early (Intensive) Behavioural Intervention in Autism : Variables and Outcomes

While this summary relates to only one recently published research paper, it is important to highlight the issues raised and the conclusions reached given the scarcity of evaluative evidence concerning early behavioural treatment based on the Lovaas model.

The current authors review existing studies about the the age of the children at which intervention is initiated, the number of hours per week and the total duration of programmes, and characteristics pertaining to the children which may be significant in predicting outcomes.

Their own conclusions support the view that positive outcomes can he achieved with fewer than the 30 or 40 hours per week considered in some quarters to be a necessary minimum, but that greater positive change appears to be associated with the timing of the intervention (children below 3 years of age rather than above 3 years), and the extent of improvement appears predictable from the overall duration (months) of the implementation of the behavioural programme.

Limitations of the study are acknowledged, including difficulty in having specific and direct knowledge of the content of each session of the programmes, the lack of direct comparability between younger and older subgroups of children in respect of initial skill levels, and the relative narrowness of the skill domains measured at pre- and post-intervention assessments. Nevertheless, this study provides useful and objective data in a field more noted for heat than light ; but it does also highlight the complexity of the interacting variables and the difficulty in forming clear judgements about potential treatment usefulness in individual cases.

M.J. Connor         March 2001



Much of the controversy that has surrounded the whole issue of early and intensive behavioural approaches (such as that advocated by Lovaas and his colleagues, e.g. 1987) concerns the lack of clarity concerning precisely how one might identify those children for whom such an approach is likely to more effective than any other, the use of control groups and the way participants are allocated to target and control samples, baseline and re-test measures, and the level and duration of the treatment offered.

Luiselli et al (2000) have attempted to explore some of these issues by examining the outcomes of home-based behavioural interventions in relation to the age of the children and the intensity of the programme.

They begin with a review of research which demonstrates that children diagnosed as autistic who receive intensive treatment before the age of 5 years, using a form of applied behavioural analysis, can show marked improvements in intellectual functioning, language, social performance, and self care.

Reference is made to the summary of evidence prepared by Green (1996) who concluded that the best outcomes are associated with very early initiation, of treatment with the children aged between 2 and 3 years, and when the programmes are maintained for no fewer than 30 hours per week and continue for at least 2 years.

On the other hand, the frequent use of the term "intensive" may lead to some confusion given the lack of clarity about what precisely "intensive" means. Not only may it be impractical to follow the exact course of the programme carried out at home, and one is aware of the concept of "therapy drift", but also there are questions whether "intensity" refers to the number of hours per week, or the number of learning opportunities that are available, or the amount of time the child works within specifically behavioural procedures as opposed to incidental teaching interactions.

There is also the major question of how a child's particular characteristics will relate to the outcome of a behavioural intervention. Some predictive validity has been applied to the variables of IQ, general developmental level, and language skill, leading to questions such as whether the outcome of some particular intervention will vary according to the age of the child when the treatment is started.

In respect of the number of hours, the current authors indicate that only Lovaas (opp.cit) has produced comparative data and he concluded that significant improvement was shown among the children who received 40 hours of intervention per week while control children who received only 10 hours per week were not shown to make meaningful progress. Some support for the importance of high intensity in terms of hours comes from the work of Weiss (1999) who reported positive outcomes for autistic children whose mean age was 41 months and who received 40 hours per week of home-based behavioural intervention over 2 years.

However, this retrospective study did not include a control group so the findings may be less than conclusive.

Meanwhile, other studies, such as that of Birnbrauer and Leach (1993) have shown that one can achieve positive results with a behavioural intervention using fewer than 30 to 40 hours per week.

This begs the question whether there is an interaction between intensity of the programme (i.e. precisely what form the daily sessions follow) and the number of hours per week, and between the programme content and the age of the child when the treatment began.

The complexity of these issues is further indicated by reference to the study of Sheinkopf and Siegel (1998) which differed from the Lovaas format in that the intervention was wholly home-based, continued for less than 2 years (mean 16 months), and was maintained for less than 30+ hours per week (mean = 19 hours). Nevertheless, the intensive home treatment group scored significantly higher on a post-treatment IQ measure than did a control group, and also showed a reduction in symptom severity. Further, the IQ score differences for the children in the treatment group did not show a positive relationship with the number of hours of treatment each week, suggesting no clear association between intensity (hours) and therapeutic outcome.

The purpose of the current retrospective study by Luiselli et al (opp.cit) was to explore outcomes among young children with autism who received a home-based behavioural intervention for less than 20 hours a week with a view to determining whether benefits varied according to child age, and length of treatment.

The participants were 16 children who had been formally diagnosed with autism or pervasive developmental disorder and who had been selected randomly from the population of children who had received home-based treatment. 8 children had begun treatment below the age of 3 years ( mean = 2.6 ), and 8 had begun treatment above the age of 3 years ( mean = 3.9). Baseline assessment had included standardised measures including the Early Learning Accomplishment Profile, together with direct observations, parent interviews, and video tapes of the children's performance.

Each child had been treated according to an individually-planned intervention programme which was implemented by well trained and experienced therapists who received weekly supervision and consultation from the director of the centre for early childhood education. The treatment was wholly home-based involving multiple weekly visits by the therapists using intervention styles based upon methods developed by Lovaas and his colleagues so that the children received discreet trial instruction and additional teaching opportunities based upon criterion-referenced and behaviour-specific learning objectives.

Strategies included modelling, prompting, positive reinforcement, etc.

The programmes had continued either until the child reached the age where it was necessary to move to the next educational stage, or when the child had achieved learning of a sufficient level to warrant some alternative provision, or when funding had precluded the continuation of the programme.

For each child there were recorded the number of hours per week of treatment, the cumulative number of months over which the programme had been provided, and the total hours of treatment for the whole treatment period. The assessment measures that were given before treatment and at the end of the treatment covered the areas of communication, cognition, fine and gross motor skills, social and emotional development, and self care.

The results showed that children who began the prograrnme above 3 years of age averaged significantly more hours of treatment per week than the children who started the programme below 3 years of age. Meanwhile, the younger group had a longer overall period on the programmes and were, on average, exposed to more hours in total. However, these latter measures did not differ significantly from those applied to the older group.

Both groups showed significant changes from the baseline assessment to the follow up assessment in all areas of functioning as listed above. However, the levels of improvement did not show significant differences between the two groups of children for any of the areas assessed.

A further analysis looking at each of the areas of functioning was carried out but only duration of treatment was found to be a valid predictor of change and significant for the areas of communication, cognition, and social-emotional growth.

The authors summarised the three major findings as follows:

· All 16 children in the sample demonstrated significant and positive changes in all the areas of functioning assessed in the formal measures used.

· The degree of change observed did not differ significantly between the two groups of children.

· Overall improvements for all the children on the communication, cognitive, and social-emotional domains was predicted by the length of time ( number of months) spent in the intervention programme.

Additional information showed that the younger group scored consistently lower than the older group in all areas of functioning tested at the time of baseline assessment, but that those children showed greater changes in their scores between the first and follow up assessment when compared to the older children.

The authors recognise that what is now needed is a comparative study in which initial skill level is matched between two groups of children, and where the number of hours per week and the duration of treatment are both controlled.

The authors also acknowledge that the nature of the home-based treatment was similarly described for all the children but that it was not possible to control precisely what happened within sessions in terms, for example, of the number of discrete trials, or the amount of time that the therapist interacted directly with the child. One could only record the amount of time that each child was exposed to instruction each week. In other words, further studies should involve a more behaviour-specific definition of the treatment thus to be clear what is meant by treatment intensity.

Yet a further weakness was that of limiting the study to "simple" pre- and post- intervention measures, although anecdotal evidence concerning the children's learning readiness or language production or level of challenging behaviour was consistently produced, with the report that all children showed positive changes in skill acquisition and a reduction in the frequency of challenging behaviours.

The study supported some of the earlier quoted findings that positive gains can be achieved despite using fewer hours than the 30 or 40 originally recommended, with the speculation that there has been some over-estimation of the minimum number of treatment hours per week that are required for positive benefit. Sheinkopf and Siegel are quoted again as speculating whether the positive outcomes are linked to the direct involvement of the parents rather than professionals in the implementation of the programmes.

In their conclusion, the authors recognise the limitations of their study, but hold that the evidence does suggest that beginning a home-based behavioural intervention before a child is 3 years of age would yield a greater level of developmental change than would be the case when introducing such interventions among children who are older than 3 years. Furthermore, the greater duration of the intervention in terms of months, independent of the number of hours per week, may be associated with better outcomes.

In any event, the number of hours involved in this current study was less than the time allocated in previous projects but it was still possible to demonstrate meaningful improvements in a number of areas assessed on a formal developmental rating scales.

The overall conclusion was that there still needs to be ongoing research and that this must involve rigorous experimental control of the various variables associated with the intervention and associated with the children if there is to be true enhancement of, and knowledge about, intensive behavioural programmes, and the production of guidelines for decision-making.


Luiselli J., Cannon B., and Sisson R. 2000 Home-based behavioural intervention for young children with autism/pervasive developmental disorder. Autism 4(4) 426438

Birnbrauer J. and Leach D. 1993 The Murdoch Early Intervention Programme afier 2 years. Behaviour Change 10 63-74

Green G. 1996 Early behavioural intervention for autism what does research tell us? In

Maurice C., Green G., and Luce S. (Eds) Behavioural Intervention for Young Children with

Autism. Austin, Texas : Pro-Ed

Lovaas 0. 1987 Behavioural treatment and normal educational and intelectual functioning in young autistic children. Journal of Consulting and Clinical Psychology 55 3-9

Sheinkopf S. and Siegel B. 1998 Home-based behavioural treatment of young children with autism. Journal of Autism and Developmental Disorders 28 15-23

Weiss M. 1999 Differential rates of skill acquisition and outcomes of early intensive behavioural intervention for autism. Behavioural Interventions 14 3-22

M.J. Connor         March 2001

This article is reproduced by kind permission of the author.

© Mike Connor 2001.

Back to NAS Surrey Branch Welcome Page