This study provides an analysis of the structure of the initial cancer consultation, the consultation styles of medical and radiation oncologists, and their effect on patient outcomes. One hundred and fifty-five cancer patients attending their first consultation with either a medical or radiation oncologist were audiotaped and the transcripts were analysed using the Cancode computer interaction analysis system. Findings revealed that medical oncologists allowed patients and their families more input into the consultation and were rated as warmer and more patient-centred compared with radiation oncologists. However, radiation oncologists spent a longer period discussing, and were more likely to bring up, social support issues with patients. Both medical and radiation oncologists varied their consultation style according to the patient's gender, age, anxiety levels, prognosis, and education. Patients seeing an oncologist who was rated as warmer and discussed a greater number of psychosocial issues had better psychological adjustment and reduced anxiety after consultation. These findings provide current evidence that may be used to inform improvements of communication skills training for oncologists and highlight the need for future communication research to separately consider oncologists from different disciplines. The most widely recommended model of medical interactions in clinical practise is patient-centred care. A patient-centred approach is one in which the doctor listens to patients attentively and sympathetically, talks about psychosocial and non-medical issues (Arora, 2003), appears warm and caring towards the patient rather than hurried, and allows the patient to have input into the consultation (Butow et al, 1995). Researchers have identified the importance of doctors varying their consultation style in response to differing patient characteristics (Butow et al, 1995). For example, researchers suggest that doctors should match their style to patient preferences for involvement in decision-making (Keisler and Auerbach, 2006), and recommend responding flexibly to patients' emotional and informational cues (Butow et al, 2002). Others have found that male and female patients may benefit from different communication strategies (Butow et al, 1997;Parker et al, 2001).Many studies have shown an association between patients receiving patient-centred care in their consultation and subsequent positive patient outcomes (Fogarty et al, 1999). However, previous studies have suggested that patient-centred care and flexibility in consultation style are used inconsistently in medical consultations. Oncologists have been shown to be poor judges of patient preferences for participation, significantly underestimating cancer patients' preference for a shared approach to decision-making (Bruera et al, 2002) and desire for information. They also tend to overestimate the amount of information they believe they have given (Chaitchik et al, 1992) and cancer patients' understanding of this information (Gattellari et al, 1999). O...