Background: The medical curriculum of the University of Trondheim was recently changed from a traditional (2.5 years preclinical studies followed by 3.5 years clinical studies) to an integrated, problem-based approach. This study examines self-reported levels of skills in practical procedures before and after internship among doctors educated under the old vs. the new curriculum., Materials and Methods: Between 1996-99 (old curriculum) and 1999-2001 (new curriculum) all doctors who had graduated from the University of Trondheim were asked to indicate their level of skills in 88 practical clinical procedures before and after their hospital internships., Results: Of the 174 interns included, 114 replied on both occasions (66%). Before internship, graduates with the new curriculum reported a significant better level of skills than those with the old curriculum. After internship, however, no difference according to curriculum could be observed. The gender difference in the self-reports from students under the old curriculum was absent., Interpretation: The new problem-based curriculum seems to have improved the practical level of skills prior to internship. This difference does, however, disappear after hospital internship.