Choice GP

An ABC of complementary medicine: a new dawn

We are excited. For years we have been planning and preparing our ABC of complementary medicine, and, finally, it's arrived (p 693). The uninitiated might imagine that an ABCall pictures and short sentencescan be dashed off in a few evenings. Regularly we hear from specialists who have seen some terrible mistake in a patient and decided that an ABC is needed to educate generalists. We explain the process. They begin. Years later they admit defeat. Clarity and simplicity need hard work.

Catherine Zollman and Andrew Vickers begin their ABC with the question "What is complementary medicine?" The answer is a heterogeneous group of practices and ideas, butcontrary to mythit is available on the NHS, regulated (in part), taught in some medical schools, and not necessarily "harmless" or more "natural" or "holistic" than conventional medicine. Nor is complementary medicine "unproved": increasing evidence shows the effectiveness of some treatments in some conditions.

Complementary treatments must face all the same tests as conventional treatments, described by Brian Haynes as "Can it work? Does it work? Is it worth it?" (p 652). "Can it work?" is answered by an efficacy study, which is conducted in optimum circumstances. But an effectiveness study is needed to see whether an efficacious treatment can work in the real world, where resources are limited, misdiagnoses are common, patients don't take their treatments, and follow up is difficultthe world where most BMJ readers work.

Medical journals have tended to include many more efficacy than effectiveness studies, not least because effectiveness studies tend to be messier. That's why we applaud the study of a group from Sydney to see if a multifaceted shared care intervention could work in elderly depressed patients in residential care (p 676). This is exactly the sort of research we need more of, even though our statisticians question the study in a commentary (p 682).

The latest "wonder drug," zanamivir, will need to go through the same tests (p 655). Created by computer assisted design, it inhibits the neuraminidase activity of both influenza type A and B viruses and has been shownin efficacy trialsto reduce the duration of major symptoms. But will the drug be as useful in the real world? It may not be, not least because "intention to treat analyses" may show it not to be so beneficial because it will be given to people who have respiratory illnesses not caused by influenza. Sadly, these less impressive results may be obscured because journals are poor at reporting intention to treat analyses (p 670).

Footnotes

To receive Editor's choice by email each week subscribe via our website: www.bmj.com/cgi/customalert