Design for decision in het clinische experiment

Design for decision in the clinical trial. Whereas an engineer would refuse to undertake the construction of an impossible bridge, the doctor is often compelled to act without a sufficient theoretical basis. Although it is common practice to describe this type of activity as the art of medicine, it should never be forgotten that this so‐called art is not a desirable state but a deplorable necessity. The ideal is medical activity derived entirely from scientific knowledge. It is the object of scientific medicine to reduce the necessity of using „impressions” as premises in medical reasoning. The only valid type of reasoning in medicine is the statistical reasoning. The nature of statistical reasoning is described in non‐technical language. The author stresses the extreme difficulty of obtaining certain knowledge in the field of therapy. The main reasons for this are (1) the variability of the material, which is irreducible, because no two individuals are exactly alike, (2) the uncertainty of our knowledge, which prevents us from exactly defining classes, (3) the placebo effect of any medical intervention, and (4) the limitation of experimentation by ethical considerations. These difficulties are analyzed and illustrated with examples. The development and limitations of modern medical statistical methods are described.