Assessing treatment-time interaction in clinical trials with time to eventdata: A meta-analysis of hypertension trials

Exploration of the variation of treatment effect over time in randomized cl inical trials with low event rates is limited by lack of power. A meta-anal ysis on individual patient data from such trials can partly solve the probl em, but brings other computational difficulties. Using an example in hypert ension, we describe appropriate methods for graphical description and stati stical modelling of treatment-time interactions in large data sets. Also, a method is developed for determining the total number of events required to detect treatment-period interactions of plausible magnitude. We conclude t hat trialists tend to overinterpret the observed data when looking for pote ntial treatment-time interactions by visual comparisons of survival curves, failing to realize the substantial amounts of data that are needed for the ir detection and estimation. (C) 1998 John Wiley & Sons, Ltd.