Researchers from Norway and New York (Eldevik, Eikeseth, Jahr, & Smith, 2006) recently compared 2 groups of children who were either receiving behavioral treatment for 15 hours/week or eclectic (i.e. using multiple treatment approaches) for 12 hours/week. After 2 years of treatment, the behavioral group made larger gains than the eclectic group.
Many studies support the fact that behavioral treatment is effective and several researchers have stated that this treatment must be intensive, meaning 40 hours of 1:1 intervention. In the 1st study of intensive behavioral intervention, Lovaas (1987) showed clinically significant gains for 17 out of 19 children with autism while the control group of the other 19 children with only 10 hours or less of intervention showed decreases in IQ. Although this study has received some criticism due to focusing mostly on IQ changes for treatment outcome and because researchers have not been able to replicate the dramatic results, it is the most commonly cited study in ABA treatment for autism and helped lay the groundwork for many years of research. Other research studies have also shown positive gains with intensive behavioral intervention, but none as dramatic as the Lovaas study in 1987. In this paper, ABA is synonymouse with Discrete Trial Training (DTT). However, as we have discussed in previous postings, it is important to note that DTT does not equal ABA, rather, DTT is one ABA approach.
The authors in this paper stated that less intensive ABA (10-20 hours/week) is acceptable when cost, availability of professionals, or concerns from parents about the stress of 40 hours/week on a child & the family are concerns. The National Research Council recommends 25 hours/week, but they do not state that this intervention must be DTT, or even ABA, but the NRC does state that practices should be evidence-based. In other words, it is not clear whether or not more natural, peer group settings are as effective (or more effective). In most situations of quality education, eclectic approaches are provided which have some 1:1 intensive ABA and more natural, peer group instruction.
In this study, 13 boys participated in the behavioral group and 14 boys and 1 girl participated in the eclectic group. All children attended regular education classrooms for about 20 hours/week.
Treatment staff were required to complete intensive training in both the behavioral and eclectic treatment group.
The behavioral intervention used was based on the Lovaas manual (1981) and the Behavioral Interventions for Young Children with Autism book (Maurice, Luce, Green) (1996). An updated Lovaas manual is now available too, if you are interested in this teaching approach!
The eclectic treatment approach consisted of alternative communication (symbol communication or sign language), ABA, total communication, sensory integration therapy, TEACCH program, and other methods based on the experiences of the teachers.
Treatment outcome was determined by IQ testing (Bayley and Stanford-Binet), standardized language assessments (Reynell and PEP-R), the Vineland Adaptive Behavior Scale, nonverbal IQ (Merrill-Palmer), and a pathology scale (No Words, Not Affectionate, No Toy Play, No Peer Play, Stereotypical Behaviors, Severe Tantrums, and Not Toilet Trained).
No significant differences in number of hours of treatment were observed between the 2 groups, both groups received about 12 hours/week of 1:1 for 20 months. The major difference between the 2 groups were the approaches used, not the intensity. In the eclectic group, ABA was not used nearly as often as in the behavioral group.
The behavioral group overall did slightly better than the eclectic group, but the difference was not clinicially significant. In this study, the children partipating had lower intellectual functioning overall to start than children in similar studies. Also, in this study, lower intensity treatment was used (less hours) than in other high intensity treatments in other studies (more hours). The quality of treatment was also not monitored as closely in the eclectic group as in the behavioral group. So, it is possible that quality of intervention may account for the difference rather than the actual approach used. The types of intervention used in the eclectic group were also not closely monitored so it is not clear how much time was spent on each treatment approach. It is also possible that the difference between the 2 groups may be due to the more systematic, data-driven approach in ABA and that the ABA approach was manualized whereas the other approaches were not necessarily.
The fact that this study was a retrospective one (meaning they went back and looked at these children after 2 years) rather than a planned and controlled investigation is also a limitation that the authors noted.
1) I think this study will get a lot of attention and I appreciate that there are researchers out there trying to answer some of the important questions about number of hours and types of interventions. I also think that the researchers did an excellent job identifying the limitations of their study and ideas for future research. My only concern about this paper is that some people may leap to conclusions which the authors probably did not intend to make the readers believe. I hope to see more studies like this one from this group of researchers addressing some of the limitations they mentioned in their paper and maybe some of the ones I list in this posting.
2) I don’t think that this study proves that lower intensity treatment is not as effective as higher intensity treatment (nor do I think that was the intention of the authors in this paper) as this study did not directly compare number of hours and the children in this study had lower intellectual functioning to start than children in the other research studies. I think it is an important study to directly compare # of hours with the same treatment approach within the same study - there is so much controversy over this and the truth is, we still do not know much about the # of hours of intervention. Many studies have demonstrated effectiveness of ABA treatments with low-intensity (e.g. PRT studies) and many studies have demonstrated effectiveness of ABA with high-intensity (e.g. Lovaas). I think this study highlights the importance of doing this research, but it does not answer the question about low-intensity vs. high-intensity.
3) I don’t think this study demonstrates that Discrete Trial Training is superior to other types of intervention because the differences between the groups were not very big and it is not clear what exactly the children were receiving in the eclectic group. I would like to see a research study comparing evidence-based treatment approaches, it is not clear to me that the eclectic group used evidence-based treatment programs and if they did, which ones did they use and which ones were effective or not? Also, I would like to see research looking at more structured ABA approaches such as Discrete Trial Training compared to more play-based types of ABA such as Pivotal Response Training (PRT). I also think it is important to look at treatments that are backed by research, comparing an evidence-based approach (Discrete Trial Training) to a potentially non evidence-based approach may be the main reason a difference between the groups was observed.
4) Although I think the measures used were interesting, I would like to have seen actual behavioral data on the children - did the children look any different after 2 years of treatment? Did they use more spontaneous language, were changes in joint attention observed, did the play skills become more age-appropriate, and did the children look more “normal” in behavioral observations? The problem with only using standardized measurements is that this information gets lost, a simple behavioral observation of the child in a play situation with the experimenter and with the parent, would have been sufficient. This type of measurement is frequently used in more naturalistic training studies which focus on generalization of skills to the natural environment and I think that kind of measurement, especially in these important studies, would shed more light onto real treatment outcome (rather than just changes in standardized measures).
5) Another thing I would like to have seen in this study is information about generalization of skills - was there a difference in this between the two groups? How was generalization measured or wasn’t it? I think that any study looking at treatment outcome for ABA should report generalization - if the skills did not generalize to real-life situations, it is not clear how valuable those skills are.