Clinical trial ethics reconsidered?

drj writes "Ethical design of randomized controlled clinical trials has required that there be substantial uncertainty about which treatment is most likely to benefit the patient, a principle known as 'equipoise'. Fries and Krishnan examined all 45 trials reported at a professional rheumatology meeting and found that all trials gave results favorable to the industry sponsor, demonstrating a significant deviation from equipoise because the results were predictable. (Nearly all patients would be concerned with even a 70:30 outcome distribution.) The authors suggest that extensive preliminary findings produced a beneficial 'design bias'. This is good, they believe, because true equipoise is inefficient, exposes more patients to risk, and conflicts with other major ethical principles. They propose a principle of "positive expected outcomes". Is this a refreshing acknowledgement of reality and useful guide or just a rewording of accepted clinical practice where treatment group size already varies based on expected outcomes?"
Fries & Krishnan, Arthr. Res. & Ther. 6(3), 2004. James F Fries and Eswar Krishnan "Equipoise, design bias, and randomized controlled trials: the elusive ethics of new drug development" Arthritis Research & Therapy Vol 6 No 3