Event

The potential for overestimation of the treatment effect when a clinical trial stops early has been discussed extensively in the literature. However, there has been much less attention paid to the converse issue, namely, that sequentially monitored clinical trials which do not stop early tend to underestimate the treatment effect. I will discuss the nature of this underestimation, including theoretical and simulation results demonstrating that it can be substantial in some contexts. I will show how a conditional approach to estimation, in which we condition on the study reaching its final analysis, may be used to validly inflate the observed treatment difference from a sequentially monitored clinical trial and to provide valid confidence intervals. Example analyses will be discussed in which the adjustment is practically important.

Ian Marschner is Professor of Statistics at Macquarie University and also holds an appointment as Professor of Biostatistics at the NHMRC Clinical Trials Centre, University of Sydney. He has over 25 years experience as a biostatistician working on health and medical research, particularly involving clinical trials and epidemiological studies of cardiovascular disease,  cancer and HIV/AIDS. Formerly he was Director of the Asia Biometrics Centre with the pharmaceutical company Pfizer, and prior to that was Associate Professor of Biostatistics at Harvard University.