ABSTRACT
Most diseases follow a pattern that we call the ‘natural history’ of disease. When this pattern is entirely predictable it is possible to determine the effects a new treatment has on a disease just by giving the treatment to a sample of patients and comparing the pattern seen to the natural history of the disease. In this situation no control group is necessary because the historical control is sufficiently reliable. There are a few conditions that follow such a predictable course and in some of these it is not necessary to carry out a controlled study, particularly if the condition has a very high mortality, where it might be unethical to have a control group.
