In some randomised controlled trials the nature of the therapy means that subjects cannot or should not be blinded. Such studies need careful design. Particular attention needs to be given to the choice of control group and the nature of the informed consent obtained from subjects, because these affect the precise research question being addressed. A survey of published studies was carried to investigate how these issues had been tackled. The paper summarizes key findings from the survey. If the research question is about the specific effect of a therapy sometimes a good case can be made for a second control group which is 'attention-controlled'. There is a need for more detailed justifications of such design decisions in published studies.