It also important to consider that a multi-center clinical study is more complex -in terms of coordination, quality control and data management- and very expensive in terms of personnel and resources. therefore it is essential to have an efficient central coordination of all trial activities and adequate funding since the beginning. it is essential to design a common protocol that should be clear and similar for all centers, also the procedures should be completely standardized for all. To reduce variation of evaluation criteria and schemes, investigators meetings, proper personnel training (as Luisa said) and closely monitoring during trial shall be implemented. Also, as Chris said, a good design should generally aim to achieve the same distribution of subjects to treatments within each center and management must maintain this design objective.
I certainly believe that in a multi-center study, it is not about how many are considered too many or sufficient, I think that the line is drawn when a budget is allocated towards this study. It all depends on how much money or how much funding there is in order to achieve this type of study and the success of the study lies on a good clinical trial management in which communication and traveling to the centers is crucial.