As part of a larger study designed to improve doctor-patient communication in randomised clinical trials (RCT), we audiotaped the discussions between doctor and patient in which consent was being obtained for a RCT. This paper reports on 82 discussions conducted by 5 clinical oncologists in both District General and University Hospital outpatient departments. When introducing the subject of trials, uncertainty about treatment decisions was expressed by the doctors in the majority of cases (79, 96.3%). This was most often stated in a general sense (78, 95.1%), but some mentioned personal uncertainty (12, 14.6%), an approach which helps to maintain a trusting doctor-patient relationship. The word randomization was mentioned in 51 (62.2%) consultations, although the process itself was usually described implicitly (78, 95.1%), e.g. by telling the patient that they would be allocated either one or other treatment. Analogies were used in 28 (34.1%) cases to describe the randomisation process. In addition, although treatments and side-effects were described frequently, (68, 82.9%) and (72, 87.8%) respectively, information leaflets about the trials were not given to 23 (28%) patients. The study shows that U.K. clinicians adopt individual methods when providing information and eliciting consent to trials.