SEQPWR AND SEQOPR - COMPUTER-PROGRAMS FOR DESIGN OF MAXIMUM INFORMATION TRIALS BASED ON GROUP SEQUENTIAL LOGRANK TESTS

The maximum information trial paradigm for clinical trials with failur e time data was recognized and has been investigated. With the maximum information trial paradigm, a study is concluded when a prespecified maximum number of events of interest, thus maximum information, has be en accrued if there was no early stopping due to treatment difference or lack thereof. We present two interactive FORTRAN programs for use i n designing maximum information trials based on group sequential logra nk tests. The program SEQPWR computes the attainable power of group se quential logrank tests given the combinations of the accrual and follo w-up durations. The program SEQOPR allows the users to investigate the operating characteristics of the maximum information trial given the information fractions of interim analyses. A clinical trial from the E astern Cooperative Oncology Group is provided to illustrate the usage and features of the programs.