Clinical Trial Conduct (CTC) web site capabilities


The Clinical Trial Conduct (CTC) web site supports two dose-finding methods: (BMA-)CRM and EffTox. The standard CRM method is a special case of the new Bayesian Model Averaging (BMA) CRM. The CTC can conduct any trial designed using CRM, BMA-CRM, or EffTox applications. For more details regarding the capabilities of these methods, see the user's guides for these programs. The user's guides are installed with the software and are also available directly from the web pages for downloading the software.


The CTC supports two basic kinds of randomization: Pocock-Simon and adaptive randomization.

The Pocock-Simon software implements "Method 2" from the paper of Pocock and Simon, Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial (Corr: V32 p954-955), Biometrics, 31:103-115, 1975. To set up a trial using the Pocock-Simon method, please contact an analyst who will fill out the request form.

The adaptive randomization software can implement any trial designed by the Adaptive Randomization simulation software. See the user's guide for the software for details of what options are supported. As with the dose-finding software, the user's guide is installed with the software and is also available directly from the web pages for downloading the software.

Note that the "adaptive" randomization trials need not actually adapt. You can use the software to design equally randomized trial with a Bayesian monitoring rule.

Single-arm trials

Single-arm trials can be designed using Multc Lean for binary outcomes and TTEDesigner for time-to-event outcomes. There is an alternative program to TTEDesigner called OneArmTTE which is available internally. OneArmTTE has more simulation features but is more difficult to use.

Single-arm trials, both binary and time-to-event, can also be designed using a method based on Bayesian hypothesis testing using the BFDesigner software. This method has slightly better operating characteristics than Multc Lean for binary trials but much better operating characteristics than TTEDesigner for time-to-event trials.

Binary outcome trials, whether designed using Multc Lean or BFDesigner, are conducted by referring to a table of stopping boundaries. No software is required for trial conduct.

Time-to-event trials can also be conducted by using a table of stopping boundaries. However, this requires calculating total-time-on-test and so is more work than conducting a binary trial.

The CTC currently supports only the method in TTEDesigner (inequalities using a conjugate exponential/inverse gamma model).