Neuroepidemiology
-
Migraine trials often have to be performed at many centres, which creates a number of methodological problems. The present paper discusses single-centre versus multi-centre designs and also comments on the design of multiple-independent trials (MIT). From a methodological standpoint, the single-centre design should virtually always be preferred. ⋯ The major advantage of multi-centre trials, apart from recruitment, is that they may yield a more representative basis for conclusions regarding the whole population. Multiple-independent trials give an opportunity to summarize information from different sources. It has not the same methodological strength as the single- or multi-centre trial, but could be valuable, especially for evaluation of side effects.