Testing alternative therapies

The journal Science is today reporting on a controversial plan by the US National Institute of Mental Health (NIMH) to test an alternative treatment for autism on children. The treatment, known as chelation therapy, involves the use of drugs that remove heavy metals from the blood. It's based on a the theory - unsupported by conventional science - that mercury in vaccines triggers autism.

Chelation therapy is widely used, but its benefits and effects are not well understood. The NIMH have therefore argued that there is a "public health imperative" to test the drug. But opponents claim that any such study would be unethical, since the quality of the trial is likely to be poor, and any results - especially negative ones - would be unlikely to alter the behaviour of parents who support the therapy.

As the science article notes, there are historical reasons to doubt these concerns. In the 1990s, national authorities felt compelled to test another popular alternative therapy for autism - a drug known as Secretin - and as a result of the studies, the drug quickly fell into disuse. This suggests that parents may be more sensitive to scientific evidence than opponents of the chelation trial fear.

There is, however, another reason to be concerned about the proposed trial. Ordinarily, drug trials on humans are only allowed to proceed when there is 'genuine equipoise' - that is, when there is genuine uncertainty regarding whether the treatment to be tested is superior to the existing 'gold standard' treatment. On one interpretation of this requirement, the scientific investigator should be indifferent between the two treatments. On another interpretation, the community of expert clinicians should be in genuine disagreement about which treatment is superior.

Now while it is true that there is some uncertainty about whether the addition of chelation therapy would be an improvement on current best practice treatment for autism, any trial of the treatment is unlikely to pass either the 'investigator indifference' test, or the 'clinician disagreement' test. That's because conventional clinicians and scientists generally believe there is no preliminary evidence or theoretical rationale for the view that chelation therapy is likely to be successful treatment: testing it would therefore be on a par with testing a randomly selected drug.

What the NIMH appears to be faced with, then, is a conflict between the equipoise requirement, which suggests that any trial of chelation therapy would be unethical, and the requirements of protecting the public health, which suggest that a trial may be warranted. The quandary is particularly vexing since any trial would have to be conducted on children. Arguably, the equipoise requirement should be regarded as particularly important for paediatric research, since the research participants are not in a position to give their informed consent. On the other hand, the ethical costs of continued widespread use of chelation therapy might also be thought particularly high since it is being used on children who are not in a position to consent.

All of this raises the interesting question of whether and when public health reasons can justify jettisoning the equipoise requirement. It certainly seems possible to imagine scenarios in which the public health costs of not performing a trial would be so great that conducting a trial where there is no genuine equipoise could be justified. (Suppose the world was in the grip of a bird flu epidemic that threatened to wipe out are large proportion of the population. Surely, in such circumstances there would be good grounds to test bird flu treatments which lacked a sound theoretical or empirical rationale.) However, it's far from clear that the current situation with regard to chelation therapy is such a scenario. There would, after all, be other ways to limit the use of chelation therapy - for example, its over-the-counter availability could be restricted.

References:

Erik Stokstad, “Stalled Trial for Autism Highlights Dilemma of Alternative Treatments”, Science 321, no. 5887 (July 18, 2008): 326, doi:10.1126/science.321.5887.326. Available at: http://www.sciencemag.org/cgi/content/full/321/5887/326?rss=1

Erik Stokstad, “Desperate Parents Spark Search for New Treatment”, Science 294, no. 5540 (October 5, 2001): 37, doi:10.1126/science.294.5540.37. Available at: http://www.sciencemag.org/cgi/content/full/294/5540/37

Kimball C. Atwood, IV et al., “Why the NIH Trial to Assess Chelation Therapy (TACT) Should Be Abandoned”, The Medscape Journal of Medicine 10, no. 5 (2008): 115. Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2438277

Comments

[Dominic]

Tom,

Equipoise is all about convincing doctors to do what they are not immediately inclined to do - to allow their patients to be randomised, to give over control over their treatment.

One difference between chelation and other alternative therapies for autism, and say a new drug therapy, is that alternative therapies are readily available to patients in the absence of a trial. There is no need for a physician to 'prescribe' chelation, and a large number of families access it already despite the lack of equipoise about its efficacy in the medical community.

The real question is whether there is a significant risk of harm from chelation.

If there is no risk of harm, then there is no real ethical problem with physicians enrolling their child patients in the trial (assuming that parents are fully informed and consent). It does not matter that the doctor does not believe that it will do any good. The worst that it could do is to have no effect.

If there is a risk of harm, then the current ethical dilemma is that hundreds of parents are subjecting their children to a risk of potential harm in the absence of any evidence that the treatment will be beneficial. In the face of that, it would be far preferable to subject only half of a group of children who might otherwise have chelation treatment to those risks.

Equipoise is a convenient fiction that allows doctors to salve their conscience. There are plenty of examples where equipoise exists, but trials would nevertheless be unethical. As in this case, there may be situations where no equipoise exists, but yet trials should be conducted.

cheers
Dom

[Tom Douglas]

Thanks Dom.
I agree with a lot of that, but I'm not convinced that trying to ascertain the risk of harm will help. At least as I understand the situation, one of the problems is that no-one has a very good idea what the risk of harm is - part of the NIMH's justification for doing the study was to get safety data. So perhaps we have to just assume there is a significant risk of harm. In that case, I'm now inclined to agree that the important question is just whether "it would be... preferable to subject only half of a group of children who might otherwise have chelation treatment to those risks".