Background: Quality of clinical interview is a key issue both for patient satisfaction and for diagnostic efficiency. Its adequacy relates to better clinical diagnosis treatment plans and patient compliance. Aim: To measure the quality of interviews performed by medical students in three Chilean medical schools before receiving specific training on the subject and to compare the scores obtained after introductory courses on interview. Material and methods: The interviews were videotaped and then evaluated using an objective scale, that measures 33 skills grouped in six areas: opening, problem exploration, non verbal facilitation, interpersonal. patient reaction and closing. The students were assigned to an experimental group that received an interactive workshop with roleplays, vignettes and videotape feedback, and to a non intervention group that received the usual bedside training on medical interviews. Results: Both groups shared the same skill level before the training, with better scores on nonverbal, patient reaction and problem exploration, and worse ones on closing and interpersonal skills. Comparing pre and post-test results, the overall score improved in the experimental group (from 33.2 to 38.3, p=0.002) and worsened among non intervened students. There were statistically significant changes for opening (p< 0.002), problem exploration (p< 0.05), non verbal facilitation (p< 0.0001) and closing (p< 0.0001). Conclusions: It is important to train students not only in specific knowledge contents but in the process of interview. This training should encourage the development of empathy and closing skills