Between- and within-patient designs

The simplest trial design of course is the parallel group design assigning patients to receive either treatment A or treatment B. While this is a valid and effective design 

It would be easy simply to conclude in light of the data that B is a more effective treatment than A, but is that necessarily the case One thing we have to remember is that the 50 patients in group A are a different group of patients from the 50 in group B and patients respond differently, so in fact the observed difference between the treatments could simply be caused by patient-to-patient variation. 

As we will see later, unravelling whether the observed difference is reflective of a real treatment difference or simply a chance difference caused by patient-to-patient variation with identical treatments is precisely the role of the p-value but it is not easy. 

This design is what we refer to as a between-patient design. The basis of the treatment comparison is the comparison between two independent groups of patients. 

An alternative design is the within-patient design. Such designs are not universally applicable but can be very powerful under certain circumstances. One form 

---