Shouldn't the control group get placebo treatment? Simply by knowing they have been "treated" could already make them feel better. When the effect you are looking for is a psychological effect, it is even more serious a research design problem.