To the Editor: The article by Myers and Cheras1 is a good attempt to evaluate the safety of complementary therapies. However, I have one problem with it, which may be significant.

The authors rightly state at the outset that “the critical issue in assessing any therapy is its risk to benefit”. At the end of their article they consider the “wider public safety issues” and point out that the risks of complementary and alternative medicine (CAM) are minimal compared with those of conventional therapies. I think this is not quite logical. Comparing risks of therapies does not make sense, because the “critical issue” is the risk–benefit profile. Comparing the risks and benefits of, for instance, acupuncture for severe pain versus opioids for the same type of pain would, I think, favour the latter over the former, even though the risks of acupuncture are minimal compared with those of opioids. In other words, comparing absolute risks of treatments is tempting, but meaningless.

In reply: We stand by our statement that “the critical issue in assessing any therapy is its risk to benefit”. 1

Ernst appears to miss the point made in the subsequent two sentences — that this assessment is available for conventional medicine, but not yet for complementary and alternative medicine (CAM) therapies and products. In the absence of such data, society must be able to make an assessment of health practices and medicines based on their overall risk. Ernst himself has made significant contributions to this literature, and in fact three of his articles2-4 were cited in our review.

The second criterion of the Australian Health Ministers’ Advisory Council Criteria for Assessing the Need for Statutory Regulation of Unregulated Health Occupations asks the question: Do the activities of the occupation pose a significant risk of harm to the health and safety of the public? Answering this question for the currently unregulated CAM professions involves consideration of the “wider public safety issues”. This forms part of the determination about the appropriateness of occupational regulation. The two comments juxtaposed by Ernst to make his point are not mutually exclusive. While awaiting the risk–benefit analysis, society will need to make decisions about the absolute risks.