Achieving research competences through medical education

Recently, Van Schravendijk and colleagues published the results of a European survey on research components in medical curricula. The emerging picture shows wide variation between medical schools and between countries: 3/4 of the schools offer research courses and in 2/3 students can do research themselves. However, in most schools, less than 10 % of students choose this option. Only half the medical schools require the writing of any thesis or literature review for graduation [1]. In contrast, within the Netherlands there seems to be large agreement: all eight university medical centres spend a considerable amount of time (14–30 ECTS) on mandatory research courses, in addition to a full-time research project for at least 4 months that all medical students do before obtaining a Master’s degree [2]. Obviously, the Dutch 2009 Framework for Undergraduate Medical Education, specifying learning outcomes and competences and thereby almost serving as a blueprint for medical curricula in the Netherlands, has been indispensable in reaching this consensus [3]. However, there are many more considerations why this amount of time and effort is spent on academic and scientific training in medical curricula, as has been nicely put forward by De Beaufort and De Goeij [4] in this issue. Interestingly, as described by Marz et al. [5] also in Europe consensus in opinion about learning outcomes regarding scientific education seems to increase, although implementation in medical curricula across Europe is a challenging next step. Some new approaches on implementing scientific training in medical education are presented in this issue. First, De Groot et al. [6] argue that it is useful to make a distinction between using evidence for individual patient care and using scientific knowledge for the development of protocols or guidelines for groups of patients or professionals. Making this distinction facilitates teaching of the accompanying competences and different CanMEDS roles. Secondly, Vereijken et al. [7] present an