Open or uncontrolled studies have suggested that providing cancer patients with audiotapes of their clinical interviews can improve information recall and reduce psychological distress. We tested these hypotheses in a 'clinician-blind', prospective, randomised controlled trial. A total of 117 patients newly referred to a medical oncology clinic who were to be given 'bad news' had their consultations audiotaped. Blind to the clinician, patients were randomly allocated to receive a copy of the tape to play at home or not (control group). At 6 months follow-up, tape group patients reported positive attitudes to the audiotape and were shown to recall significantly more information about their illness than did controls. Overall improvement in psychological distress at 1 and 6 months follow-up, as measured with the 30-item General Health Questionnaire and the Hospital Anxiety and Depression Scale was no different in the two groups. However, a second-order interaction suggested that poor-prognosis patients were disadvantaged specifically by access to the audiotape, with less improvement in psychological distress at 6 months follow-up than non-tape controls. Patient access to audiotapes of clinical interviews promotes factual retention but does not reliably reduce psychological distress and may be actively unhelpful in some subgroups of patients.