ADDPLAN provides the comprehensive toolbox for the design, simulation, and analysis (including sample size reassessment) for group sequential studies. Specifically, designs and sample size planning facilities include the following designs:
- one-sided and two-sided testing of means (paired and unpaired samples)
- one-sided and two-sided testing of proportions (paired and unpaired samples)
- noninferiority and equivalence (TOST) testing of means (difference and ratio specified)
- noninferiority and equivalence (TOST) testing of proportions (difference or relative risk specified)
- survival analysis
- non-inferiority and equivalence (TOST) for survival analysis
- general design, which is used as a synonym for a (group sequential) test situation, where there is no (assumed or observed) effect available, thus, testing and sample size, or power calculation is solely based on p-value/z-value, or sample size, respectively
Designs are based on the specification of the following parameters:
- the type I error rate
- the number of planned interim analyses
- the type of stopping boundaries within the Wang and Tsiatis family of tests , within the class of alpha-spending designs, or the class of Pampallona-Tsiatis designs, respectively
- the stopping for futility bound, overall or by stage, binding or non-binding, as required
- the planned time points of analysis
- the assumed effect and the outcome's variability depending on the type of endpoint.
ADDPLAN allows the users to choose between the on-hand computation options depending on the type of endpoint (variance known/unknown in case of normal endpoints, with/without continuity correction in case of binary endpoints, Schoenfeld/Friedman formula in case of survival endpoints) , as well as whether the sample size will be calculated at a defined power , or power will be calculated at a given sample size. In addition, the sample size allocation ratio can be changed from the default setting '1' to another value , or optionally the optimum sample size allocation ratio, which minimizes the maximum and the average sample size for the design chosen, will be calculated.
Output comprises the total sample size in a comparable fixed sample design, total maximum sample size according to the group sequential plan, sample sizes per stage, ASN under the null hypothesis, ASN under the alternative hypothesis, and ASN under a value midway between null and alternative hypothesis, as well as power spend, alpha spend, and one-sided significance level.
Graphical output comprises plots of stopping boundaries, power, ASN and maximum sample size.