The efficacy of a communication skills training programme was shown through a randomised trial. Oncologists (N=160) from 34 cancer centres were allocated to written feedback plus course; course alone; written feedback alone or control. Each clinician had 6 - 10 interviews with patients videotaped at baseline and 3 months postintervention. Analysis of videotapes revealed improvements in the communication skills of clinicians randomised to training (n=80) compared with others (n=80). A 12-month follow-up assessment is reported here. Robust Poisson conditional analyses of counts of changes in communication behaviours revealed no demonstrable attrition in those who had shown improvement previously, including fewer leading questions, appropriate use of focused and open-ended questions and responses to patient cues. Additional skills, not apparent at 3 months, were now evident; the estimated effect sizes corresponded to 81% fewer interruptions (P=0.001) and increased summarising of information to 38% (P=0.038). However, expressions of empathy (54%, P=0.001) declined. The overall results show that 12 - 15 months postintervention, clinicians had integrated key communication skills into clinical practice and were applying others. This is the first RCT to show an enduring effect of communication skills training with transfer into the clinic.