Control groups appropriate for surgical interventions: ethical and practical issues.

Abstract

Surgical trials differ in several critical aspects from pharmaceutic trials because they are invasive and the procedure itself may carry a significant risk. The surgical procedure cannot be fully specified in advance; although the general approach may be the same, the exact methods used during surgery may differ. Surgeons may have different levels of skill and experience with the techniques; experience is known to affect success rates. Sham surgeries are difficult to justify ethically, yet they have been successful in showing that accepted surgical procedures were not efficacious. Double masking participants in a surgical trial may be difficult; for example, we cannot mask the surgeon. There is also a need to minimize or avoid bias. Therefore, greater flexibility is needed in the design of surgical trials than in the design of pharmaceutic trials. Some of this flexibility is possible through the use of more sophisticated statistical designs. At times, it is necessary to separate the provider of the treatment (the surgeon) from the evaluator of the success of the treatment (a different physician, research nurse, or study coordinator).