How to read the evidence · 4 of 4
Trial phases, randomization, crossover and accelerated approval
Phases
Clinical trials of a new drug usually move through phases. FDA describes them as follows for the typical case.
| Phase | People | Length | Main purpose |
|---|---|---|---|
| 1 | 20 to 100 healthy volunteers or people with the disease (cancer drugs are usually tested in patients) | Several months | Safety and dosage |
| 2 | Up to several hundred people with the disease | Several months to 2 years | Efficacy and side effects |
| 3 | 300 to 3,000 people with the disease | 1 to 4 years | Efficacy and monitoring of adverse reactions |
Source: FDA, Step 3: Clinical Research (see below). Many oncology trials combine phases (for example phase 1/2) or use different sizes.
Randomization and crossover
In a randomized trial, chance rather than choice decides who gets which treatment, so the groups should be alike on average apart from the treatment.
A real example on this site: in the IRIS trial, 65.6% of the control arm crossed over to imatinib (median 0.8 years on control therapy first), so the long-term analysis focuses on the imatinib arm and there is no valid OS comparison between arms. At 10 years 20.1% of the imatinib group had unknown survival status (censored at last contact), so the estimate rests on incomplete follow-up. See the trial.
Accelerated approval
FDA's accelerated approval pathway allows earlier approval of drugs for serious conditions with an unmet need, based on a surrogate endpoint. The company must then run confirmatory studies. If a confirmatory trial shows benefit, FDA grants traditional approval; if it does not, FDA has procedures that can lead to removing the drug or indication.
Sources and further reading
These pages explain general methods in plain language and were written for this site; the sources below are the references to read next. Some classic papers have no abstract online, so the explanations stay at textbook level.