Toggle Menu

Clinical Trial Randomization Tool

  • For clinical trials, educational purposes, or just for your own interest.
  • Uses MTI randomization to generate the allocation sequence.
  • Default values are provided below. You may adjust these as you require or prefer.
  • When you click “Request Confirmation Email,” your request will be sent to a server. You will receive an email when your download is ready (typically in a few minutes).
  • For more detailed instructions, you may view the Tool Instructions page.

Basic Trial Info

A simple trial design is used by default. For more complex trials, change the trial settings below to customize arms and/or stratification.

A short identifier for your trial. When your randomized assignments are downloaded, the files will be labeled with this name.
Max: 10,000. For unstratified trials, this will be the total number of participants in the trial. For stratified trials, this will be the number of participants per stratum.
End of Randomization

  • Final count of participants in each arm will be within a tolerance controlled by the MTI value.

    Available randomization methods include the Asymptotic Maximal procedure, Chen's procedure, and the Big Stick procedure.

  • Final count of participants in each arm will be exact values determined by the arm allocation ratio. Select this option if each trial arm requires an exact targeted count of participants.

    The randomization method will be the Maximal procedure.
This choice affects which MTI randomization methods can be used for your trial.
The number of arms in this trial. Number of Trial Arms: 2
One or more values in this section are invalid. Please adjust them before submitting.
Arm:
Ratio:
Arm 1
1
The tool supports 2-arm, 3-arm, and 4-arm trials.
Arm 1
Arm 2
One or more values in this section are invalid. Please adjust them before submitting.
If you are planning to stratify your trial, select this option, then enter the strata information in the inputs that appear. Stratify Trial: False
Number of Stratification Variables: 1
One or more values in this section are invalid. Please adjust them before submitting.
Max: 4 variables. Up to 100 strata are supported.
Variable 1
Category Names

    Based on the number of categories for your 1 stratification variables, your trial will have 1 strata. For each stratum, you will receive one worksheet containing 1,000 assignments.

    Your results will use the following names for each stratum's sequence worksheet. Due to length restrictions in Microsoft Excel, category names may be truncated.

      One or more values in this section are invalid. Please adjust them before submitting.

      Algorithm Parameters

      The following parameters should work well for most trials, but can be further customized.

      Randomization Method: Asymptotic Maximal
      The difference between any two arms never exceeds the MTI. Maximum Tolerated Imbalance (MTI): 3
      At any point when the arms are imbalanced, this is the percent probability that the next enrollment goes to the arm that currently has fewer enrollments. When both arms have equal number of enrollments, the probability is 50% to each arm. If assignment to the larger arm would violate the MTI, the probability is 100% and the next enrollment goes to the smaller arm. Balance-Forcing Probability: 60.0
      One or more values in this section are invalid. Please adjust them before submitting.

      Select a method to use for randomization. All methods shown use MTI randomization and are suitable for use. For more information about these different methods, consult the Learn About Randomization page.

      Randomization Method

      • The most under-assigned arms are always favored by a balance-forcing probability that varies based on the current degree of imbalance and the number of participants remaining to be randomized. This is the only supported method for trials which require exact counts in each arm.

      • The most under-assigned arms are always favored by a balance-forcing probability that varies based on the current degree of imbalance.

      • Neither arm is favored until the MTI threshold is reached — each arm's probability of assignment is always 0%, 50%, or 100%, based on the current imbalance. The Big Stick method is a special case of Chen's procedure where the balance-forcing probability is 50%.

        This method is supported only for two-arm trials with an arm allocation ratio of 1:1.

      • The under-assigned arm is always favored by a preset balance-forcing probability, e.g., 60%. Chen's procedure is a generalized version of Big Stick where the balance-forcing probability can be greater than 50%.

        This method is supported only for two-arm trials with an arm allocation ratio of 1:1.

      Maximum Tolerated Imbalance (MTI)

      At any point in the randomization process, for any stratification group, no trial arm will be assigned more than this many participants than any other arm.

      Values are scaled with respect to the largest trial arm.

      At any point when the arms are imbalanced, this is the percent probability that the next enrollment goes to the arm that currently has fewer enrollments. When both arms have equal number of enrollments, the probability is 50% to each arm. If assignment to the larger arm would violate the MTI, the probability is 100% and the next enrollment goes to the smaller arm.

      You may enter any value between 50.0% and 99.9%, but the higher the value, the less random the procedure. We default to 60.0%.

      One or more values in this section are invalid. Please adjust them before submitting.

      Contact Info

      We will email you at this address when your results are ready!

      The Clinical Trial Randomization Tool is a web application developed by the National Cancer Institute (NCI) and the National Institutes of Health (NIH) to help researchers generate randomization sequences for their clinical trials.

      A study statistician should be consulted in conjunction with the design and implementation of your randomization scheme to ensure it is appropriate for your clinical trial. NCI and NIH are not responsible for how any randomization sequence is used in a clinical setting.

      To prove you are human, please complete the Captcha above.