Introduction:
It is generally known that randomizing the patients participating in a study, improves the credibility of the study. It is of importance that the researcher does not know in which group a patient has ended so that they cannot (un)intentionally exert any influence over the result of the study. There are however risks attached to randomization.
Equal division via block randomization
If every patient separately and in random order is randomized, the chance exists that certain groups get significantly more patients than the other groups.
To combat this, block randomization is used: The desired number of patients is divided over the blocks, that exist out of a variety of groups in the research. Per block there are as many people randomized in every group.
This ensures that, even if halfway along the block no more patients are added, the risk of difference in number of patients per group is still significantly lower. But this system is not watertight. It still knows the following risk:
- It can happen that some groups have more way more patients with certain prognostic (already known) factors, such as weight, age, etc., then other groups. This can lead to skewness in the research.
By using stratification the risk of uneven distribution of prognostic factors is removed. Different stratification classes are set up, based on factors relevant to the research. For example, age, weight, medicine use or a combination of factors. Consequently, the randomization is done per stratification class - so there will be an equal number of patients in each group.
ResearchManager has built in tools to make stratifying and randomizing easy, safe and independent. You can read more about how to manage these options on our support page.