For those unfamiliar with the Autism Research Institute (ARI), despite its reputable-sounding name, its research seems anything but. I consider myself an autism advocate: I advocate for myself and I would like to advocate for those on the spectrum to live happy fulfilling lives (this does not mean though that they necessarily need to be "normal" in order to do so).Therefore, it comes as little surprise that I consider such an organization as ARI harmful in that it treats autism as a demon personified, trapping small children in lonely worlds; however, also, it professes that autism is a separate entity from the child and therefore worthy of loathing (Defeat Autism Now!). Autism is the perpetrator and the person with autism is the victim. Such an attitude teaches a person that they are broken and diseased. And while it is easier to separate a diseased leg from one's personality, try separating the brain.

ARI instead is better at propagating desperation (usually that of the parents) and spreading false hope. What is wrong with some false hope you ask? Perhaps you should ask those people who fell victim to various televangelists like Peter Popoff
or Jim Bakker.

It comes as a great shock that the Institute's original founder, Bernard Rimland, was a vocal proponent of the Mercury-Autism Hypothesis, particularly given his strong beginnings in turning the tide of autism theory from the Refrigerator Mother to acceptance of autism as a neurobiological developmental condition (see his 1964 book, Infantile Autism: The Syndrome and its Implications for a Neural Theory of Behavior).

Nevertheless, the Institute is not held in particularly high scientific regard, and given its publications (one of which I will critique below) that comes as little surprise.

Critique on the ARI article, Preliminary Results of 3rd DMSA Study by Adams et al.
http://www.autism.com/danwebcast/presentations/anaheim/adams.pdf

I would like to say first off that I do not believe autism is caused by mercury poisoning. However, given newer research on immunology and oxidative stress in autism, I do feel that something is going on, perhaps involving vaccination as one agent of effect. However, given postmortem evidence of brain anatomy in autism, it is clear that the development of autism begins in early utero, far predating vaccinations.

That said, my critique point by point:

Methodology:

1) The paper states that this is a double-blind study, implying that there is a control group. Yet in this article, they report no results on the controls. Therefore, the reader is unable to contrast and compare results between the two groups.

2) There is no mention as to how the participants were recruited and how they were chosen to take part. Medical establishments? Autism Speaks? CAN? Local churches? Where they all given the ADOS and any child who tested in the autism range was included? This is important because, without knowing where the participants came from, there's no way to accurately judge that this was an unbiased sample and an unbiased sample is vital to do the statistical analyses these researchers performed.

3) In this study, the ADOS was given to the participants following treatment 1 and then treatment 2. I have a big problem that the researchers gathered no ADOS scores prior to the children beginning the study, thereby giving no baseline score for comparison.

4) The researchers give no background information on any of the participants. There is no indication as per the ages of the children nor a listing of other medicinal and nonmedicinal treatments these children may or may not be receiving. Various medications do have effects on the body's oxidation. Also, was the control group required to be medication-free? If so, that has the potential for a huge confounding effect. And if any of the experimental kids were receiving behavioral treatments, this went unanswered. Any of these factors could be playing a part: natural development with age, medication, or other treatment plans which are taking place concurrently or prior to the study.

Results:

5) 80 began the study, only 40 finished. Even though this is enough to resemble some sort of statistically normal distribution, 40 is still a very small number for these researchers to recommend glutathione as a treatment for autism.

6) The results of the study were compared to both parent impressions (general interview as well as ATEC) as well as the ADOS. I have two issues with this: a) yes, parents should be asked of their opinions; however, in such a study, more objective measures should be used and biased opinions avoided. The ADOS, I have personally learned, is an adequate tool for diagnosis of Autistic Disorder and PDD-NOS in young children. However, it is a poor tool for differentiating within the spectrum. Therefore, it should not have been used to detect a change in severity, especially when the shifts in severity were so subtle. For measuring such severity, it is an inadequate tool.

7) With the ADOS, I would've liked to know who performed the first and the second ADOS. Were they the same people, potentially introducing bias? Or different people who might score the same child slightly differently? There is no mention as to who performed these tests except that they were "qualified".

8) As for the correlation between excretion levels and ADOS, this to me spoke mountains. The most significant correlation between excretion levels and ADOS scores following the first dose was 0.36. For a little background on correlations, here are the ratings of correlations:

0.0-0.2 = no correlation
0.2-0.4 = low correlation
0.5-0.7 = moderate correlation
0.8-1.0 = high correlation

These occur in both positive and negative ranges, with negative numbers indicating a negative (or inverse) correlation. Most researchers view correlations that aren't moderate or higher with due suspicion that the result was potentially coincident. Following the 9th dose, the highest correlation is 0.35 (another low correlation). Yet these results are espoused as being significant. Therefore, all scores ranged from low correlation to no correlation at all.

9) I take issue with a study that uses a quote to make its conclusions-- a quote nonetheless originated by the leading researcher in this study:

Finding lead or mercury in an autism victim is like finding a bullet in a homicide victim-- further investigation needed for 100% certainty, but in both cases it is highly likely that one caused the other (Adams, 2007).

Why is this a problem? Well, first off, it's clear that Adams has already made up his mind before even performing this study, which makes him a prime candidate for accidental or intentional experimenter bias. He also proposes that finding lead or mercury in autism is the guilty bullet having caused the conditions; and if that is his logic, then every child in their study should've been a "homicide victim" since every person, autistic or not, has mercury and lead in his/her body. On the same note of logic, that would also imply every person is autistic.

10) Before performing further replication studies, the researchers of this study are already recommending longer treatments for those whose lead levels were still "high". I.e., they're recommending a treatment before having tested whether it's sound and relatively risk-free without having studied the effects of longer durations, but at the same time they continue to state that more research is needed before the treatment should be recommended. Double-talk.

11) From a single study, they have concluded there are no adverse side effects on general health.

12) Despite the low correlations, they even bolded their conclusions that higher excretion decreased severity as measured by the ADOS, using techniques like bolding or highlighting to make their point stick despite that it's a weak one. An advertising technique. In science, if the results are significant enough, no amount of bolding or highlighting is ever necessary.

13) Their final conclusions stated, "Glutathione, lead, and mercury are linked to severity in autism, and to improvement due to DMSA." This statement is prematurely false. They proffered no evidence in this study directly linking higher mercury and lead levels to their autism participants. The authors have automatically jumped that extra step and assumed correlation is equivalent to causation (and low correlations at that). For instance, if indeed there are higher levels, it could be very well that autism is the cause of higher levels of mercury and lead, rather than the reverse.

14) Despite that the paper is called the 3rd preliminary study, there is no review of previous results. Did the current results hold true to previous ones? Do they differ? Did the methodology differ? There's no indication.

15) The article gives no indication as to having been published in a peer-reviewed journal. I assume this means it hasn't been published except online by ARI who was funding the research.

I don't mean to say that the only place for publication is in established journals, no. What is important is the peer-reviewed process. And there is no such indication for this article.

My Conclusions:

This was a poor study, with researchers who seem to have made up their minds before performing the research and whose bias, either accidentally or intentionally, affected the design of the study. The methodology was poor, the reporting of said methodology was poor, making replication impossible. This study has been designed to only be replicated by the ARI researchers and therefore impervious to true scientific scrutiny.