Many cancer trials need thousands of patients and take several years to carry out. Some trials need to be this large because they have to be able to find quite small, but important, differences between treatments.

If the difference between a new treatment and a standard treatment is expected to be quite small, many patients will be needed to show reliably which is better. Statistics help researchers make sure that trials include enough people to give useful information.

For example, to show that a treatment can give a one in ten improvement in survival for a group of patients, a trial with 1,000 patients would be needed. In order to show a one in twenty difference, 4,000 patients would be needed. However, not all cancer trials need such large numbers of people.