Geneva - Every year malaria, tuberculosis and AIDS kill around 6 million people, almost all of them in the developing world. These premature deaths are a reproach to us all. They are also a huge blow to countries' hopes for development. Urgently, more needs to be done to save the lives of millions of poor people. Part of the problem is poor countries' lack of access to drugs. The poor cannot afford expensive medicines. Keeping an AIDS patient alive for a year can cost up to $15,000 - 24 times the average annual income in Zimbabwe, where one in four adults is HIV-positive.

Critics of the World Trade Organization say its agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) makes matters worse. They argue that by requiring developing countries to enforce pharmaceutical patents, the agreement enables drug companies to charge exorbitant prices that the poor cannot afford. Clearly, we need to find new ways of improving access to existing drugs in developing countries. But we also need to give pharmaceutical companies an incentive to develop new drugs. Industry puts the average cost of developing a new drug at around $500 million. Were it not for a patent system that rewards companies for risking millions on research, anti-AIDS drugs would not exist.

That is why the TRIPS agreement tries to strike a healthy balance between the short-term need to make vital drugs available to those who need them and the long-term, equally vital need to encourage research into new drugs. To reward research, the agreement protects patents for 20 years (although, since it usually takes years to test and approve new drugs, a patent's effective life is much shorter). To improve access to drugs, it imposes some conditions and allows certain restrictions on patent rights.

For one thing, patent holders have to disclose their invention. This allows others to use information about a patented drug to research new drugs during the patent's life, and ensures that it is truly in the public domain once the patent expires. Second, if a patent holder refuses to license a patented drug on reasonable commercial terms, a government is allowed to license it to other companies or use it itself without the patent holder's authorization, so long as adequate compensation is paid. Third, as a recent WTO panel has concluded, governments can facilitate the "early working" of patented pharmaceuticals by generic competitors. Fourth, if governments authorize parallel imports of a patented drug from countries where it is sold more cheaply, this cannot be challenged at the WTO.

Developing countries have generally had to enforce the TRIPS agreement since Jan. 1, 2000, when a five-year transition period ended, but those which do not already provide patent protection for pharmaceuticals have until 2005 to introduce it. The transition period for least-developed countries ends in 2006 (with the possibility of an extension). But most developing and least-developed countries already grant patent protection for pharmaceuticals. For them, the real significance of the TRIPS agreement may be less the requirement to protect new drugs and more the explicit enshrining in international law of the flexibility I have described. One promising idea is differential pricing; pharmaceutical companies would charge less for drugs in poor countries than in rich ones.

This is consistent with the TRIPS agreement and is backed by, among others, the World Health Organization, the European Commission, Médecins sans Frontières and some industrialists. It is already starting to happen. The WTO and WHO secretariats are organizing a workshop to explore how it could become more widespread. Important issues are how to prevent low-priced drugs from leaking back from poor countries to rich ones and how to convince rich-country consumers and taxpayers of the fairness of lower prices in poor countries. Of the 300-odd drugs deemed essential by the WHO for basic health care in developing countries, fewer than 20 are under patent protection anywhere.

There are no effective treatments for some ills that affect people in poor countries only, because developing them is not commercially viable. Only 10 percent of global research funds target diseases that affect the poorest 90 percent of the world's population. Donors urgently need to stump up funds to finance research, for instance, into vaccines against malaria and AIDS. Funds are also needed to ensure that companies have a credible market for new drugs and vaccines that they develop for such diseases, as well as to help pay for existing essential drugs, insecticides and anti-malaria bed nets.

Nor should providing the basics, such as clean water, good sanitation, better nutrition and more condom use, be neglected. Lower tariffs and taxes and more efficient distribution channels are also important. Most of this is outside the WTO's remit. But by promoting free trade we can make a difference. Openness is essential for economic growth, which can help pay for better health care and sanitation. The writer is director-general of the World Trade Organization. He contributed this comment to the International Herald Tribune.