Some thoughts on the planning of survival studies

Hans Ulrich Burger  – Biostatistics, Hoffmann - La Roche Basel

Survival type endpoints play a major role in the clinical development of many new treatments in a number of therapeutic areas. In oncology, so called progression-free survival and overall survival are usually the main endpoints required to achieve regulatory approval for a new therapy. Also in metabolism, e.g. for cardiovascular or diabetes therapies, time to first cardiovascular event plays a major role, either as a safety or as an efficacy endpoint. Study designs and conduct for such an endpoint are however not straightforward. Already standard sample size calculations can become difficult when the assumed affect is close to a minimally clinically relevant effect. Crossover and non-proportional hazard functions could be other challenging criteria. The talk will highlight some of the issues and will provide some recommendation for the planning and conduct of time to event studies.