Thursday, July 23, 2009

An anomaly in autism intervention research
Promotion first, science later, if ever. This pattern is near universal when it comes to autism interventions. In the absence of good quality research, autism interventions are loudly claimed to be effective.

For those promoting ABA-based autism interventions, claims of effectiveness unfounded in good quality research were only the first step. The real triumph has been widespread agreement that fair tests of ABA-based interventions are unethical and bad for autistics. As a result, any experimental design carrying the risk of being informative about the benefits and harms of ABA-based interventions has, for a long time now, been considered unethical.

The practice of claiming effectiveness for an autism intervention which has not been fairly tested, then using these claims of effectiveness to deem fair tests unethical, has clear benefits to service providers. And this practice has received wall-to-wall support from autism advocates, who have in turn imposed it on autistics through lobbying and litigation.

Meanwhile, this practice is not admired outside the realm of autism advocacy. Premature claims of effectiveness in themselves make fair tests of interventions more difficult--particularly, as is the case with many autism interventions, when blinding cannot be fully achieved. But those promoting ABA-based autism interventions go further and demand that autistics be entirely denied the benefit and protection of good experimental design.

In my view, this practice--its longstanding and widespread support by autism advocates, its more recent support by ASAN (example here), and its widespread imposition on autistics--continues to greatly harm autistics. Any group so denied such basic standards of science and ethics would be harmed.

Interestingly, not all autism researchers have stooped to prevailing autism advocacy standards.

Some years ago, a pilot RCT of an early autism intervention was published (Aldred et al., 2004). The intervention was manualized, of relatively low intensity, and targeted solely at the parents of preschool autistic children. Through a true experimental design, the intervention was found to have several significant positive effects, most markedly in increasing the expressive language of very young autistic children.

Successful RCTs of early autism interventions are virtually non-existent. But the researchers involved in Aldred et al. (2004) totally failed to go forth and proclaim the effectiveness of their manualized intervention. After all, they only had a small pilot RCT--the same size as the only published RCT, not a notably successful one (Smith et al., 2000, 2001), in the entire 48-year history of research into ABA-based autism interventions.

Instead of issuing premature claims of effectiveness, the authors of Aldred et al. (2004) and many other collaborating UK researchers went on to conduct a large multi-site RCT, the Preschool Autism Communication Trial. They received MRC funding for this 4-year project in 2005.

Some information about the PACT's design has been reported in a recent paper (Aldred & Green, 2009).

The trial began in early 2006 and will finish late this year. The researchers hoped to recruit 144 preschool autistic children across three sites, but exceeded their expectations and recruited 152. These children were randomized into two arms, one receiving the PACT intervention, while children in both arms continued to receive whatever other interventions their parents chose ("treatment as usual").

So far as I know, the PACT is larger than any existing published RCT in the history of autism research. The PACT's design means that it carries a good risk of being informative about the effectiveness of the studied intervention. There is no good reason for the PACT to be unprecedented, but it is.

In one of the PACT's recent newsletters, the PACT's Chief Investigator, Jonathan Green, notes that there has been a lot of interest in the PACT manual and in "training opportunities" with respect to this intervention. Dr Green responds:

Until we have the results of the trial we are not able to disseminate the intervention manual or to undertake any training; but be assured that after the results are in (and depending of course on the outcome in terms of effectiveness!) we do have potential plans in place for such dissemination.
In other words, and keeping in mind there is a successful published pilot RCT, there are no plans to disseminate information about the intervention--not until there is good quality evidence as to its effectiveness. This is the recognized scientific standard in non-autism areas.

Aldred et al. (2004) and the PACT shouldn't in any way be above scrutiny or criticism. For example, I can see problems in how Aldred et al. (2004) was designed (I mention one of them here), which may or may not be carried over to the PACT. But the pilot RCT and the resulting PACT generally demonstrate the only science- and ethics-based sequence when it comes to interventions and treatments: good quality research first, then think about spreading the word.


Aldred, C.R., & Green, J. (2009). Early social communication
interventions for autism British Journal of Hospital Medicine, 70 (3), 143-145

Aldred, C., Green, J., & Adams, C. (2004). A new social communication intervention for children with autism: pilot randomised controlled treatment study suggesting effectiveness Journal of Child Psychology and Psychiatry, 45 (8), 1420-1430 DOI: 10.1111/j.1469-7610.2004.00338.x

Smith, T., Groen, A.D., and Wynn, J.W. (2000, 2001). Randomized trial of intensive early intervention for children with pervasive developmental disorder. American Journal on Mental Retardation, 105, 269-85. Erratum in American Journal on Mental Retardation, 105, 508. Erratum in American Journal on Mental Retardation, 106, 208.

Addendum: This post is included in the 116th Meeting of the Skeptics' Circle hosted by Beyond the Short Coat.