Читать книгу Randomised Clinical Trials - David Machin - Страница 51

The particular interventions to be compared in a clinical trial will be determined by the choice of the specific research question concerned. We shall see in Chapter 9 that the size of a clinical trial, that is the number of subjects that need to be recruited, depends critically on the effect size. In loose terms, the effect size is the anticipated difference between (say two) interventions. If the effect size is small – trials have to be large, whereas if it is large – trials may be small. Now, setting all other aspects aside, small trials are preferable to large. They require fewer patients and therefore can be completed in a timely manner. Following on from this, large effects are preferable to small! Of course, before the trial is conducted we do not know the actual size of the effect (indeed that is what we are trying to determine from the trial itself once completed) but we can choose to make comparisons between interventions which at the onset are as different in their anticipated effect as reasonably possible. Thus, the way in which the comparator (often termed control and test) interventions are chosen is critical.