Methods of evaluation of performance of adaptive designs on treatment effect intervals and methods of designing two-stage winner designs with survival outcomes
PDF
PDF format is widely accepted and good for printing.
Fang, Fang. Methods of evaluation of performance of adaptive designs on treatment effect intervals and methods of designing two-stage winner designs with survival outcomes. Retrieved from https://doi.org/doi:10.7282/T3PV6K41
TitleMethods of evaluation of performance of adaptive designs on treatment effect intervals and methods of designing two-stage winner designs with survival outcomes
DescriptionThe accuracy of the treatment effect estimation is crucial to the success of phase 3 studies. The calculation of fixed sample size relies on the estimation of the treatment effect and cannot be changed during the trial. Oftentimes, with limited efficacy data available from early phase studies and relevant historical studies, the sample size estimation may not accurately reflect the true treatment effect. Several adaptive designs have been proposed to address this uncertainty in the sample size calculation. These adaptive designs provide the flexibility of sample size adjustment during the trial by allowing early trial stopping or sample size re-estimation at the interim look(s). The use of adaptive designs can optimize the trial performance when the treatment effect is an assumed constant value. However in practice, the treatment effect is more reasonable to be considered within an interval rather than as a point estimate. Proper selection of adaptive designs will decrease the failure rate of phase 3 clinical trials and increase the chance for new drug approval. This dissertation proposes an optimal design based on an interval using the "regret concept". A mathematical framework is developed to evaluate the adaptability of different designs. In addition, this dissertation identifies the factors that may affect the performance of adaptive design and derives the expected sample size for two-stage sample size re-estimation designs. In drug development, a phase 2 trial may not be feasible due to long follow-up or lack of resources. So it may necessary to evaluate several promising regimens in the confirmatory phase 3 trial. In this case, an interim analysis is often used to drop the inferior arms and to avoid the high cost, long term trial conduction, and exposure to ineffective treatment. This approach is considered as combining the two phases into one study: phase 2 portion will be carried out by the interim analysis. When appropriate surrogate endpoints exist, such as progression free survival in oncology trials, they can be used at the interim analysis to accelerate the drug development process. The statistical frameworks are available in the literature for designs with continuous endpoints. However, it is very challenging to derive the correlation between log-rank statistics at interim and final analysis when survival endpoints are used. An asymptotic correlation of log-rank statistics is developed and the features for a two-stage design survival trial using same or different endpoint at interim analysis is explored.