Ethics of clinical trials from a bayesian and decision analytic perspective: whose equipoise is it anyway?

Abstract

Best treatment—a question of values Treatments typically influence competing objectives. Radical prostatectomy versus more conservative methods, for example, involves a trade-off between cure and side effects. Because different people value outcomes differently, the best treatment can only be determined after an individual has been consulted in such a way as to invite careful consideration of the issue. Donovan et al advocate that men be informed unequivocally that all treatments for early prostate cancer are equally suitable, but this does not encourage individuals to explore their values in such a way that these can be reflected in decisions taken. A man with early prostate cancer who wants a child may place a higher value on preservation of fertility than someone who has no such aspirations. So, given a typical decision involving trade-offs, a recruiter cannot legitimately be uncertain about the best treatment until the individual concerned has been consulted. Stating unequivocally that the best treatment is uncertain, as advocated by Donovan et al, forecloses on further discussion about which treatment may best suit an individual. Indeed these authors go on to say that “if recruiters gave any indication that they were not completely committed to these aspects, patients would question randomisation, often using subtle and sophisticated reasoning.” So, confidently stating that the best treatment was unknown would suppress sophisticated questioning. Provided no one spots what is going on, trials are likely to flourish but at the cost that patients will not be put in the best position to choose their care. The conclusion that the best treatment is unknown is a possible result of a patient's decision, not an input to that decision.