Sample size for randomized trials in stroke prevention. How many patients do we need?

THE QUESTION OF SAMPLE SIZE is important, both in the planning of new investigatipns and in the evaluation of published results. Problems in the interpretation of negative trials are now widely recognized. When sample size is small, a negative result cannot be interpretated as providing evidence that the trial therapies are for all practical purposes therapeutically equivalent. In many instances the number of patients incjuded in published trials is simply too small to provide an adequate comparison of the therapies under study. It is therefore imperative that careful thought be given to sample size calculations for future investigations.