Application of group sequential test methods in survival analysis and power analysis of spending function

Abstract

Objective: Application of group sequential test methods in survival analysis. Material and Methods: In the study, two groups of 224 patients diagnosed as lung cancer (KHAK and KHDAK) have been followed prospectively for 1060 days. These patients have been examined using group sequential test methods step by step and two groups' survival times have been compared for each step. In addition considering varied observation numbers and survival rates a simulation study is done for comparing 5 spending functions' power analysis. Results: Patients of two different types of lung cancer have been compared with respect to their survival times for each step (every 265 days). Determination of the step in which the study terminates has been fulfilled with group sequential test methods. In addition, spending functions used in these methods have been tested for superiority using power analysis. Of all the spending functions O'Brien Fleming produced the highest power analysis score. Conclusion: In the case of two treatment methods/groups comparisons, without waiting for end of the study, which treatment methods/groups to yield better could be determined with group sequential test methods. Hence, patients' survival times could be extended directing patients to other treatment methods/groups after a particular step.