East® Adapt
Validated Adaptive Trial Methodologies Expand the Design and Simulation Capabilities of East®

Adaptive designs effectively "tune" a trial to maximize the probability of a successful outcome. Cytel's East® Adapt incorporates the most up-to-date adaptive statistical methodologies for later stage, confirmatory studies.

Midcourse sample size corrections are becoming increasingly popular since they avoid the risk that a trial will fail because it is underpowered.

East® Adapt makes practical design alterations to the sample size based on unblinded looks at the interim data without inflating the false positive rate.

Additionally, trial planners use East® Adapt to design and monitor trials with adaptive changes in the number and spacing of future looks at the accruing study data. By scheduling interim examinations of the study data, sponsors can revise the assumed primary and nuisance parameters – all while preserving the study's statistical validity and integrity.

New East® Adapt Validated Methodologies and Tools

Latest additions - many found only in East® Adapt

Trial Simulation Tools
· evaluate power and expected sample size of the Cui, Hung and Wang (Biometrics, 1999) method
· evaluate power and expected sample size of the Chen, DeMets and Lan extension (Statistics in Medicine, 2004)
· evaluate power and expected sample size of the Müller and Schäfer method (Biometrics, 2001)

Conditional Power Calculators
· for use at the planning stage
· for use at the interim monitoring stage
· for computing conditional power if trial is stopped at any future look

An Interim Monitoring Worksheet specifically for group sequential designs with one sample size reassessment at each interim look:
· computes the Cui, Hung and Wang weighted statistic at each look
· permits use of the unweighted statistic if the Chen, DeMets and Lan criteria are met
· computes repeated p-values and repeated confidence intervals, adjusted for adaptive changes

Median Unbiased Point Estimates, Confidence Intervals and p-values based on a new publication by Brannath, Mehta and Posch (Biometrics, June 2009). No other software package provides these median unbiased point estimates and their associated confidence intervals.

Possible trial adaptations include:

• Increasing or decreasing the sample size
• Modifying the error spending function
• Changing the number and spacing of interim looks
• Refining the primary hypothesis

Contact us for more information, including licensing and training options.