The aim of medical research is to develop improved treatments or cures for diseases and medical ailments. Part of that research involves comparing the effects of alternative treatments with a view to recommending those that should be used in practice. The treatments are compared using properly controlled randomized clinical trials. In such trials the treatments are given either to healthy volunteers (in the early phase of development) or to patients (in the later phases of development). We will refer to patients, volunteers, or whoever is being compared in the trial as the subjects.
Two types of design are used in these trials: the parallel-group design and the crossover design. In order to explain these we will consider trials for comparing two treatments A and B. The latter might be different ingredients in an inhaler used to treat asthma attacks or two drugs used to relieve the pain of arthritis, for example.
In a parallel-group trial the subjects are randomly divided into two groups of equal size. Everyone in the first group gets A, and everyone in the second group gets B. The difference between the treatments is usually estimated by the difference between the group means.
In a crossover trial the subjects are also randomly divided into two groups of equal size. (In agriculture and dairy science, crossover trials often are referred to as changeover trials.) Now, however, each subject gets both ...