Registration of drugs for treating cancer and HIV infection: a plea to carry out phase 3 trials before admission to the market.

Drugs for cancer and HIV infection tend to be admitted to the market on the basis of results from phase 2 trials. Assessing the benefit-risk balance with phase 2 trials often is difficult--the effect of the drug is usually temporary; the correlation between response or improvement of clinical measurements and the patient's wellbeing is often poor; and the side effects of drugs for these fatal diseases are serious. Therefore, although sometimes difficult to conduct, comparative trials that use standard treatment, placebos, or best supportive care remain the cornerstone for reliably assessing the benefit-risk balance.