Skip to main content

Randomization, Stratification, Blinding

​What else is important to know about Phase III Trials?

The results of a phase III study may change the standard approach to treatment in a group of patients. Because of this, there are often rules which are present to limit the chance that wrong conclusions will be drawn from the study. In phase III studies, the goal is to compare one treatment to another. Ideally, the groups of patients being treated with the old and new treatments will be as similar as possible to each other, so that only the treatment is different between the groups. The special rules of a phase III trial help to make the groups of patients as similar as possible. The special rules, or steps, are called "randomization", "stratification", and "blinding".


Randomization means that, for each patient who agrees to participate in the study, one of the study treatments is selected for them randomly, like flipping a coin, or like drawing a name from a hat. This is done to avoid having different numbers of patients receive one of the treatments, and to avoid doctors and patients biasing the results by choosing which treatment they think is better. Usually randomization is done by a computer or person at a central location who does not know the patient, but knows that they are eligible for the study.


Often there are types of differences between cancers and patients which would be predicted to affect the results of treatment. For example, within a group of patients with breast cancer being offered participation in a trial, there may be patients with different age group, different menopausal status, or different tumour stages, for example. Stratification means that the person or computer randomizing patients to each group tries to assign roughly equal numbers of patients with similar health or tumour characteristics to each type of treatment. The factors to be stratified for are identified from the outset, as part of the study rules. Stratification is again done to try to make sure that true conclusions can be made about the usefulness and safety of each treatment being studied. It also may help investigators to find out if there are specific features of a patient or their cancer which would indicate a higher or lower chance of benefit from a treatment.


"Blinding" means that the patient or doctor or both are not aware of which treatment the patient is getting. This is done so that the measurements being taken of the response to the treatment or the side effects of the treatment are not biased by the doctors' or patients' expectations. Blinding again helps to make sure that the conclusions that can be made about the treatments at the end of the trial are as correct as possible. If blinding is done, the treatments being given are designed to look the same, and to be given on the same schedule. This may require the use of a placebo treatment. 


SOURCE: Randomization, Stratification, Blinding ( )
Page printed: . Unofficial document if printed. Please refer to SOURCE for latest information.

Copyright © BC Cancer. All Rights Reserved.

    Copyright © 2024 Provincial Health Services Authority