Researchers, practitioners and drug companies around the world are engaged in a complex, tentative dance over the best way to tap into the unknown potential of traditional Chinese medicine. The scientific community and the drug industry both tend to be sniffy about 'traditional' cures; yet there is a strong sense that millennia of practice in China — much of it barely documented — is likely to have yielded at least some treatments that work.

Pharmaceutical companies are understandably eager to enter a Chinese domestic market that was estimated by the Boston Consulting Group to be worth US$13 billion last year, and growing fast. But they are tantalized by one opportunity above all: the prospect that the nation's traditional medicine might contain a number of potentially profitable compounds hidden somewhere in its arcane array of potions and herbal mixtures.

The task of finding these elusive gems has been approached in a typically reductionist manner, with researchers seeking single compounds that might have a role in treating specific diseases. Sometimes this has been successful: artemisinin, for example, which is currently the most effective treatment for malaria, was fished out of a herbal treatment for fevers. But such success stories have been few and far between.

So if traditional Chinese medicine is so great, why hasn't the qualitative study of its outcomes opened the door to a flood of cures? The most obvious answer is that it actually has little to offer: it is largely just pseudoscience, with no rational mechanism of action for most of its therapies. Advocates respond by claiming that researchers are missing aspects of the art, notably the interactions between different ingredients in traditional therapies.

Nevertheless, the drug industry is not exactly awash with promising new medicines at the moment. Perhaps as a result, the global regulatory process has become increasingly receptive to traditional approaches. In 2004, for example, the US Food and Drug Administration issued new guidelines on botanical drugs that made it much easier to get extracts into clinical trials if there was some history of prior use, and that obviated the need to characterize all compounds in an extract.

Some researchers in China and elsewhere, meanwhile, are advocating systems biology — the study of the interactions between proteins, genes, metabolites and components of cells or organisms — as a way to assess the usefulness of traditional medicines (see page 126). Constructive approaches to divining the potential usefulness of traditional therapies are to be welcomed. But it seems problematic to apply a brand new technique, largely untested in the clinic, to test the veracity of traditional Chinese medicine, when the field is so fraught with pseudoscience. In the meantime, claims made on behalf of an uncharted body of knowledge should be treated with the customary scepticism that is the bedrock of both science and medicine.