Fundamental Issues of Choosing the Right Type of Trial

BACKGROUND: Relevant advantages and disadvantages of crossover/self-controlled and parallel group studies are discussed. RESULTS: For ethicists crossovers are frequently less of a problem. Also, they may be less costly, and, in addition, statistically more powerful. If the treatments in a trial are curative, however, or if the variables are subjective, a parallel group design is the right type of trial. In case of unstable disease, or a negative correlation in a treatment comparison, this is so. CONCLUSIONS: Routinely including a set of rules prevents procedural biases and wrong study designs. The present paper is a preliminary effort for the very purpose of that.