Create your randomization application in just minutes
Simple randomization is the simplest form of treatment allocation. When the subjects are randomized, the system randomly selects a treatment for each one.. This is much like flipping a coin for every randomization. In the short term, this may result in a slightly different number of subjects in each arm.
For example, a simple randomization with two treatment arms (A,B) is:
ABABAABAABABBABBABA
Permuted-block randomization, or blocking, is used to balance treatment arms within a block so that there are the same number of subjects in each treatment arm. A block contains the same number of each treatment. Blocks of different sizes are combined to make up the randomization list.
For example, a simple block of size 4 with two treatment arms (A,B) is:
ABBA
The randomization list is created by combining blocks. Randomize.net can combine random block sizes.
An example of a randomization list with block sizes of 4 and 6, with two treatment arms (A,B), is:
ABBA BABAAB AABBAB BBAA ...
When Randomize.net generates a randomization list with random block sizes, it will randomly choose between the block sizes with equal probability.
Stratified randomization allows the configuration stratification variables to balance treatment arms between prognostic characteristics. It uses permuted-block randomization within each stratification level when building the blocks.
For example, you may wish to stratify based on clinical site and gender. Each combination of stratification levels will result in a separate randomization list:
CLINICAL SITE 1:
MALE:
ABBA BABAAB AABBAB BBAA ...
FEMALE:
ABABBA BABA AABB BABBAA ...
CLINICAL SITE 2:
MALE:
ABAB AAABBB AABBAB ABAABB ...
FEMALE:
BBAA BABA BBAA BABAAB ...
Minimization is a dynamic randomization algorithm designed to minimize imbalance between treatments, taking prognostic variables into account. Based on the prognostic variables of the current and previously randomized subjects, an imbalance score is computed for each treatment. The treatment with the lowest imbalance score is assigned to the current subject.
More information on minimization can be found here: Minimization: Testing A Dynamic Randomization Algorithm.
Please note: Minimization often involves customization to your Randomize.net trial and may result in additional fees.
Randomize.net, which is owned by Interrand Inc., was established in 2000 and developed under the supervision of Andrew Willan, PhD, Professor of Biostatistics, University of Toronto, and Professor Emeritus, Department of Clinical Epidemiology and Biostatistics, McMaster University.
We strive to provide all of our clients with top-quality service and support.
© 2015 Interrand Inc. Privacy Policy