The effects of patients' and practitioners' preferences on randomized clinical trials