Premium
Designing complex group sequential survival trials
Author(s) -
Lakatos Edward
Publication year - 2002
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.1193
Subject(s) - interim , computer science , schedule , statistic , survival function , interim analysis , sample size determination , transformation (genetics) , log rank test , realization (probability) , piecewise , rank (graph theory) , function (biology) , statistics , survival analysis , mathematical optimization , econometrics , mathematics , clinical trial , medicine , mathematical analysis , biochemistry , chemistry , archaeology , pathology , combinatorics , gene , history , operating system , evolutionary biology , biology
This paper presents methodology for designing complex group sequential survival trials when the survival curves will be compared using the logrank statistic. The method can be applied to any treatment and control survival curves as long as each hazard function can be approximated by a piecewise linear function. The approach allows arbitrary accrual patterns and permits adjustment for varying rates of non‐compliance, drop‐in and loss to follow‐up. The calendar‐time–information‐time transformation is derived under these complex assumptions. This permits the exploration of the operating characteristics of various interim analysis plans, including sample size and power. By using the calendar‐time–information‐time transformation, information fractions corresponding to desired calendar times can be determined. In this way, the interim analyses can be scheduled in information time, assuring the desired power and realization of the spending function, while the interim analyses will take place according to the desired calendar schedule. Copyright © 2002 John Wiley & Sons, Ltd.