In the 30 years or so since adequate clinical trials were first conducted the basic principles have scarcely changed. Randomization remains an essential ingredient. Trials have increased in size and duration, and standards of design and execution are now high. Particular attention should be paid to the effect of baseline variables, for correcting imbalance and improving precision, and also to reveal interactions between these variables and the treatments.Trials in the secondary prevention of stroke and myocardial infarction, as in other fields, frequently appear to engender controversy. This is partly due to dissenting views about aspects of the protocol on which choices had to be made (perhaps an argument for flexibility in the protocol); and partly due to termination at an inconclusive stage. The ethical dilemma prevents any easy solution, but it is suggested that early stopping should only be allowed when evidence for a difference is very strong.Different approaches to the handling of protocol departures are discussed; the pragmatic approach, with inclusion of all cases after randomization, is advocated.Factorial designs are likely to be increasingly adopted, in view of the pressure of new drugs and the need to economise in the large numbers of patients required. Some interpretative problems are discussed.
Factorial designs: an overview with applications to orthodontic clinical trials
