Dr. Mansfield is preparing to test a new anticancer drug. Based on preclinical and preliminary clinical data, he has reason to believe that this drug will slow metastasis in patients who have melanoma. His proposal to the Institutional Review Board (IRB) calls for a double-blinded, treatment protocol in which half of the patients would receive the drug and half would receive placebo, but all would receive standard therapy.

Dr. Pembroke, a clinical oncologist and member of the IRB, reminds the other members of the IRB that the current treatments for melanoma have relatively little impact on disease progression. For this reason, he argues that the investigators face two problems: (a) Although half of the patients will receive the placebo, many will enroll in the study because of the belief that they will be receiving something new and effective. Under this circumstance, is it fair to call this "informed consent"? (b) The lack of an existing, effective treatment for a typically fatal illness means that all patients should be given the opportunity to receive the drug. Therefore, to assess the effectiveness of this drug, Dr. Pembroke proposes that comparisons can be made to historical controls.

Dr. Mansfield counters that changing treatment standards, variations in monitoring of patients, and variable expertise among practicing oncologists make the validity of a study with historical controls highly questionable. He is particularly concerned that such a study might detect a substantial qualitative improvement, but would miss detection of what he believes will be a small, but important, improvement.