• J Lubsen and J G Tijssen.
• Center for Clinical Decision Analysis, Erasmus University, Rotterdam, The Netherlands.
• Control Clin Trials. 1989 Dec 1; 10 (4 Suppl): 151S-160S.

AbstractThe design of a clinical trial derives from its objective, which in its turn is dictated by the needs of clinical practice. Therefore, a common condition, a simple treatment, and a simple outcome measure are opportunities rather than design options. The argument that the direction of the net treatment effect can be expected to be similar across patient subgroups is inconsistent with biologic reality, and may misguide interpretation of both individual and pooled trial results. If no data on patient characteristics are recorded, scientific generalization is hampered and there is limited opportunity to learn about mechanisms of action and about specific indications and contraindications. If the treatment influences an intermediary outcome, which is assumed to have a causal role in the disease process that the treatment eventually seeks to influence, the effect on the intermediary outcome is of great importance--for clinical practice is very much centered around this outcome and the effects on intermediary outcomes may serve as an aid in explaining why there is an effect or why not. In conclusion, there are some research questions that may warrant large and simple trials, but many clinically very relevant research questions require a more complex trial design.

10 keywords...

hide…