In a previous study, urologists, radiation oncologists and medical oncologists in Britain, Canada and the United States were asked to state how they would wish to be treated if they had urologic cancer as described in six clinical scenarios and whether they would agree to be entered in randomised clinical trials. This study disclosed major controversy regarding treatment options for each scenario and reluctance by these experts to enter randomised clinical trials. In the present study a second questionnaire which included a summary of the treatments selected initially was sent to the same 227 oncologists. Respondents were asked, in view of these additional information, how they would wish to be treated and whether they would enter themselves (or their patients) on randomised trials comparing the two treatment options most favoured by their colleagues. Most respondents did not modify their treatment preference. There was still poor agreement to enter themselves on trials (29%), but a higher proportion would offer such trials to their patients (45%). Thus the demonstration of controversy about optimum treatment did not influence personal bias, but could facilitate the entry of patients into trials that address major controversies. We conclude that treatment strategies of urologic oncologists are influenced minimally by opinions of their colleagues, but that the method of using surrogate questionnaires is a valuable aid to the design of clinical trials.