BACKGROUND Communication skills are considered 'core skills' in the curriculum of psychiatry but studies evaluating the effectiveness of a time-limited training course in interviewing skills in psychiatry have remained rare. The aim was to assess the effectiveness of training in patient-centred interviewing on the interview performance of psychiatric residents. METHODS Psychiatric residents (n=10) each interviewed 12 different anonymized standardized patients (SPs), eight before and another four after training. SPs simulated psychiatric out-patients who attended for a first visit to the psychiatric out-patient clinic. The consultations were videotaped, transcribed and coded with a classification scheme developed for psychiatric consultations from which an interview performance index was derived. An interrupted time-series design and a segmented regression analysis with multilevel analysis explored the performance trend within the series of consultations. RESULTS The regression model evidenced a horizontal slope at pre- and post-training, with a significant level change. These findings excluded the presence of a practice effect and indicated a significant effect of training. Performance variability between and within residents over the series of consultations increased at post-training. CONCLUSIONS The training improved patient-centred interviewing performance. More post-training exercise time and supervised practice are necessary to establish consistent performance patterns at a higher skill level.