The case of loose inclusion criteria in clinical trials

It is often believed that strict inclusion criteria must be defined before starting a comparative clinical trial of some new therapy. Strict inclusion criteria tend to make the sample of patients entered in the trial homogeneous, but they reduce the number of patients potentially available for the trial. Loose inclusion criteria, on the other hand, increase the heterogeneity of the sample, but also its size. This paper shows that loose inclusion criteria are in general preferable to strict ones, because they result in a trial of shorter duration. The implication is that unless there are good a priori reasons to exclude some subgroups of patients from a clinical trial, the broadest possible inclusion criteria should be adopted.