SMALL CLINICAL-TRIALS - ARE THEY ALL BAD

Statisticians have long argued that randomized controlled trials shoul d be sufficiently large to achieve their purpose, and for common disea ses with major public health implications this has brought many benefi ts. However, there are many instances where it is unrealistic to expec t clinicians to provide the information on which sample sizes are calc ulated and undue emphasis on trial size can be counterproductive. Exam ples of such trials are given and some general issues discussed, inclu ding the unhelpful contribution of hypothesis testing, the need for re plication and the role of the statistician.