Michelle Dawson writes a letter

Evidence-based medicine is the idea that all medical decisions should be based on the best scientific evidence that is available.  The concept is really very straightforward. You take the results that research has provided, rank them according to the quality of the information, and use that ranked evidence to decide what the best course of treatment is.

The important thing to note here is that a treatment doesn't have to be proven beyond all possible doubts in every possible scenario before it can be used - it just has to be shown to be applicable to the case at hand and have enough evidence that says it is more likely than not to work.

In a way, the name says it all. You base your medical decision on what has been shown to work, in other words evidence-based.

As easy as that concept is, some people seem not to understand it, especially when it comes to autism. As it stands now, there are very few treatments that have been shown to be effective in helping people with autism. There are many different things that are used in and attempt to help people (especially children) with autism  but most of these things do not have a solid evidence base behind them. I am not saying that these treatments can't or don't work but rather that they fall outside what is considered evidence-based medicine - they would be considered alternative medicine.

One of the few treatments for autism that does have a solid evidence base is ABA (Applied Behavior Analysis).   While it is not guaranteed to work for everyone, the available evidence shows that it can be an effective tool to help teach children with autism and is almost universally recommended.

That is, with the exception of the universe of Michelle Dawson.  As I have pointed out before, Ms Dawson has a real problem with ABA. She seems to have an almost irrational obsession with proving that ABA is somehow unethical or immoral to use on children with autism. She would tell you that she has ethical concerns and that there is very little evidence that ABA works.  However, Ms Dawson is almost universally alone in her opinion.

With that in mind, consider a letter that Ms Dawson wrote to The Lancet in rebuttal to a study¹ published late last year. In this letter², Ms Dawson attacks the evidence cited in the study to support the assertion that ABA is "highly effective". She dismisses three of the cited works as not studying ABA and says that another didn't study what it says it did.  This leaves, in her opinion, only one small study as the entire evidence base for ABA which leads to the conclusion that -

"the claims made by Levy and colleagues, with respect to intensive ABA-based programmes for autistic children, have no basis—either in the review they cite or in any other published study."

Open and shut case - ABA isn't evidence based - so says Michelle Dawson. We should all stop using it now and rejoice that we have seen the error of our ways, right?

Wrong.

The authors of the study responded³ to the letter. They say that Ms Dawson is correct about the one cited reference - it did not do what it said it did. However, they "respectfully disagree" with her assertions about the other three studies and point out that they did in-fact study ABA. They go out to point out that there were five additional studies that they did not cite directly because of space limitations and that brings the total to "only" nine randomized trials that support the effectiveness of ABA.

Remember when I mentioned above the quality of the information? Randomized trials (when done correctly) are considered some of the best evidence available. To put this into perspective, most issues are considered "proven" when they have two or three good randomized trials.  Here there are nine supporting the conclusion that ABA is effective. And, as the authors go on to point out -

In addition to these trials, a host of other studies that used rigorous and sometimes not-so-rigorous quasi-experimental designs point to the efficacy of ABA-based methods, such as discrete trial training, pivotal response training, and teaching in functional routines, either alone or in combination, in improving adaptive behaviour, language, and in some cases socialisation of children with autism.

In other words, the response to Michelle Dawson's letter was an unequivocal "what do you mean there is no evidence?"  Kinda makes you wonder what Ms Dawson is talking about. Perhaps she should reconsider the evidence.

But don't take my word for it, the text of the letters are freely available on The Lancet's site (free registration required), go read them for yourself.


References

1: Levy SE, Mandell DS, Schultz RT. Autism. Lancet. 2009 Nov 7;374(9701):1627-38.
Epub 2009 Oct 12. Review. PubMed PMID: 19819542.
doi:10.1016/S0140-6736(09)61376-3

2: Dawson M, Gernsbacher MA. Effectiveness of intensive autism programmes.
Lancet. 2010 Feb 27;375(9716):722-3. PubMed PMID: 20189018.
doi:10.1016/S0140-6736(10)60299-1

3: Mandell DS, Levy SE, Schultz RT. Effectiveness of intensive autism programmes
- Authors' reply. Lancet. 2010 Feb 27;375(9716):723. PubMed PMID: 20189020.
doi:10.1016/S0140-6736(10)60300-5