Saturday, March 10, 2007

Evidence Based Autism Interventions

The expression "evidence based" is often encountered in discussion of autism treatments or interventions. Parents, government officials, professionals and researchers will use the expression when discussing the effectiveness of various interventions in treating or curing autism. A well known article in the Journal of the Australian Medical Association, linked on the sidebar of this blog site, is entitled "Autistic Children Deserve Evidence Based Intervention". But what does the expression "evidence based" mean and which autism interventions, if any, meet the evidence based standard?

It is important for parents and political decision makers trying to determine what interventions to provide autistic persons to understand the concept of "evidence based" interventions. There have been many quack autism interventions offered both by self promoters and wishful thinkers. Such interventions waste valuable development time of autistic children, waste family and government funds, and in some cases, actually cause direct harm. Unfortunately there have also been those who are opposed to curing or treating autistic persons, those who wish to promote alternative interventions, or who wish to avoid expending public resources to provide treatment, who attack Applied Behavior Analysis (ABA) despite the quality and quantity of research supporting the efficacy of ABA as a health and education intervention for autistic children.

CAIRN, the Canadian Autism Intervention Network, defines "evidence based":

By evidence-based, we mean the best available information based on scientifically rigorous research that produces consistent findings no matter how many times the study is repeated.

CAIRN also notes that there are substantial differences in the quality of evidence. It lists on its web site the characteristics of studies that provide high quality evidence.

Systematic reviews
A systematic review uses a clear and systematic method of finding and appraising relevant, high quality research studies whose combined results are used to answer a pre-determined question about treatment.

The strength of the systematic review is in its ability to combine data from studies of different populations in different settings to show if a treatment is widely applicable. By pooling data from patients in a number of studies, thus increasing the sample size, a systematic review can increase the reliability of the findings, showing whether a treatment actually works or may be useless or even harmful.

Randomized controlled trials (RCTs)
In RCTs, research subjects are randomly placed in one of two groups. One group receives the intervention that the study has been designed to measure; the other group (known as the control group) does not. The control group may receive a placebo, no treatment, or another therapy. Study participants (and ideally, the researchers) do not know to which group they have been assigned.

RCTs are considered the second strongest level of evidence for the effectiveness of a treatment. The strength of the RCT is that it helps ensure that the two groups under study do not have any important differences between them that could influence whether or not the treatment they receive works.

Controlled clinical trials (CCTs)
In a controlled clinical trial, one group receives a therapy and the other (control group) does not. As in RCTs, the control group may receive a placebo, no treatment, or another therapy. Unlike RCTs, however, the participants are not randomly assigned to each group.

With some populations, or in some situations, it is not possible to randomly assign participants to one group or another. CCTs still use a comparison group and efforts are made to ensure that the two groups do not have serious differences between them that could influence the results of the study.

Multiple and single case studies
In multiple and single case studies, subjects are tested to establish a baseline. They are then given an intervention, after which they are re-tested to determine what change, if any, has occurred.

Sometimes, especially when there is great variability in a condition, or small numbers of people with a condition, it is not possible to conduct research with two groups. In multiple and single case studies, the participants are used as their own comparison when they are given alternating treatments, or a treatment and then no treatment. Although not providing the same strength of evidence as an RCT or CCT, when done rigorously, these studies can yield valuable information about treatments.

In its February 2000 (Rev.) Autism Task Force Report the Maine Administrators of Services for Children with Disabilities (MADSEC) reported the results of its thorough review of the professional literature on educational interventions for autism. The MADSEC Task Force also interviewed leading practicioners of various autism interventions. It concluded that only one intervention met the standard of being an evidence based effective autism intervention - ABA.

• Substantiated as effective based upon the scope and quality of research:
Applied behavior analysis. In addition, applied behavior analysis’ evaluative procedures are effective not only with behaviorally-based interventions, but also for the systematic evaluation of the efficacy of any intervention intended to affect individual learning and behavior. ABA’s emphasis on functional assessment and positive behavioral support will help meet heightened standards of IDEA ‘97. Its emphasis on measurable goals and reliable data collection will substantiate the child’s progress in the event of due process.


There is a wealth of validated and peer-reviewed studies supporting the efficacy of ABA methods to improve and sustain socially significant behaviors in every domain, in individuals with autism. Importantly, results reported include “meaningful” outcomes such as increased social skills, communication skills academic performance, and overall cognitive functioning. These reflect clinically-significant quality of life improvements. While studies varied as to the magnitude of gains, all have demonstrated long term retention of gains made. Other major contributions of ABA to the education and treatment of individuals with autism include:

• a large number of empirically-based systematic instruction methods that lead to the
acquisition of skills, and to the decrease/elimination of aberrant behaviors;
• a technology for systematically evaluating the efficacy of interventions intended to affect individual learning and behavior; and
• substantial cost/benefit.
Over 30 years of rigorous research and peer review of applied behavior analysis’ effectiveness for individuals with autism demonstrate ABA has been objectively substantiated as effective based upon the scope and quality of science.

Since the publication of the MADSEC Autism Task Force Report there have been more studies confirming ABA as the most effective autism intervention. In "A comparison of intensive behavior analytic and eclectic treatments for young children with autism" published in Research in Developmental Disabilities, 26, (2005), pp. 359-383, the authors, Jane S. Howard, Coleen R. Sparkman, Howard G. Cohen, Gina Green, Harold Stanislaw reported the results of a non randomized comparison trial with three intervention groups of children with autism spectrum disorders. 29 children received one-on-one intensive behavioral therapy for 25 to 40 hours per week (IBT group); 16 received eclectic therapy (multiple treatment methods) with a teacher to student ratio of 1:1 or 1:2 for 30 hours per week (AP group); and 16 received a non-intensive, eclectic, small group, public early intervention program for 15 hours per week (GP group.) The study's authors reported that 14 months after initiation of the interventions, the intensive behavioral therapy group scored significantly higher on all measures, with the exception of motor skills, where there was no difference among groups.

Alan Harchik, senior Vice President with the May Institute, has written an excellent article on the topic of evidence based interventions. Mr. Harchik describes evidence based standards and cautions against the use of the many treatments and interventions for autism that do not meet that standard including sensory integration, facilitated communication, auditory integration, chelation and hyperbaric oxygen treatments.

The term evidence-based practice is frequently used in the fields of science and medicine. It refers to procedures that have been tested using scientific research methods and shown to be most likely to produce positive results. This means that objective studies have been conducted using reliable data collection methods, consistent implementation of the treatments and the careful control of the research conditions.

The research findings are then published in professional journals after review by a panel of experts in the field.

Mr. Harchik, like the MADSEC Autism Task Force Report, concludes that ABA is the intervention with the most evidence supporting its effectiveness.

"For children with autism, the procedures that have the most evidence supporting their effectiveness are those that use applied behavior analysis. .... applied behavior analysis methods include an assessment of the factors that are impeding learning or maintaining behavior problems, using positive rewards, teaching in small steps, using prompts and guidance, and collecting data to monitor progress. "

The studies and expertise which provide evidence of the efficacy of ABA in autism intervention have not persuaded some who oppose the use of ABA interventions with autistic persons. Some will continue to insist on an unattainable standard being met before they will acknowledge the efficacy of ABA. For others though, for parents, professionals and government decision makers seeking to help autistic children now ABA, their only agenda is to find out what works, what really works, based on the best available evidence. To date, ABA is the only intervention for which claims of efficacy meet the high quality evidence based standard. That is why parents and sincere autism advocates try to obtain passage of legislation in Canada and the United States to ensure the availability of ABA treatment for autistic children.


Unknown said...

I agree, wholeheartedly...

Andrew Livanis


My only argument with ABA's standing as to efficacy is that it is so statistically based that is a 'darling' of those seeking quantifyable evidence. I don't doubt that it works (we use Carbone's angle on ABA with Liv) I just think that other therapies might get short schrift because they are less statistically based...

Anonymous said...

Is Michelle Dawson wrong, when she writes in "The misbehaivour of behaviourists":
"Apart from objecting to aversives, autism-ABA opponents have been, it seems, too busy to take on ethics. What they have taken on would fill a phone book:
threats to internal and external validity, randomized assignment, differential sex proportions (inadequate chi-square analysis), selection bias, statistical
regression, outcome measures, the problem of residual autism and how to detect it, the problem of replication, the perpetual how-many-hours-per-week question,
etc., and this just for Dr Lovaas' 1987 study and its 1993 follow-up."

She seems to conclude that the research conducted to show evidence of the effectivenss of ABA, is suspect. Now I know that she opposes ABA, bu tno-one has ever criticized her on the content of her article so I assume she's accurate.

Also, how can we be sure that no treatment exists - or will be developed - that is more effective, or has more evidence of being so, than ABA, if ABA is presumed to be medically necessary and, hence, we run a risk of not being able to do research on other treaments cause they aren't presumed to be effective in the firs tplace (by the ABA industry at least).

Laslty, why is it that ABA is not commonly used in some states, like Michigan (where floor time / Greenspan's treatment is widely used) or North Carolina (which uses TEACCH), while ABA is supposedly the *only* effective treatment for autism?

Unknown said...


I don't know if other therapies "get short shrift". They either haven't been subjected to study in accordance with evidence based standards or they have been studied and found wanting, even harmful in some instances, as is the case with Facilitated Communication.


Unknown said...


The point of this comment is to suggest that it is important to understand the use of the expression evidence based as a standard for evaluating autism treatments. MD's Ph Ds in pyschology, psychiatrity and education have assessed ABA as the treatment/therapy which meets the evidence based standard of demonstrated effectiveness based on the quantity and quality of research supporting the use of ABA. There have literally been hundreds of studies which point to the efficacy of ABA as an autism intervention.

Although I can think of many reason why these professionals would not want to respond to Ms. Dawson I will not speculate about that on this forum.

We can not be sure that no treatment exists or will be developed that will be more effective than ABA. What we can be sure about is that AT PRESENT no other treatment has been subject to as much study as ABA and no other treatment has as much rigorous study based evidence of its effectiveness as ABA.

Most North American jurisdictions in fact implement as in my home province of New Brunswick, Ontario, Alberta, British Columbia, or endorse ABA. ABA has been endorsed by the US Surgeon General's office, and state agencies in New York, California and Maine amongst many others.

Parents with newly diagnosed autistic children have to make real life decisions which will impact the quality of life of their children. They deserve to know which treatments or therapies for autism are supported by the best available evidence as recognized by credible sources so that they can make responsible, informed choices for their children.

That is exactly what ABA offers and parents should know that when they make their choices.