Abstract:
Planning and analysing of the one-sample log-rank test is investigated. The one-sample log-rank test might be the method of choice if the survival curve of patients under a new treatment shall be compared to a historic reference curve. Such settings arise, for example, in clinical phase IIa trials if the response to a new treatment is measured by a survival endpoint in order to account for potential loss to follow-up. Present sample size formulas for the one-sample log-rank test are based on the number of events to be observed in order to achieve a certain power. Here, we propose a new criterion to be followed in order to achieve a certain power. Both approaches are asymptotically equivalent. However, simulations indicate that the new criterion might be preferred for small sample size.