INTRODUCTION: In Flanders an important part of training to become a general practitioner (GP) is undertaken within a general practice. This requires a GP trainer to know how to facilitate learning processes. This paper reports a study focused on the research question: Does short but individualised training of GP trainers contribute to their teaching competence? METHOD: A total of 47 GP trainers were randomly divided into 2 groups of a pretest/post-test control group design. After a multiple-station teaching assessment test (MSTAT), the experimental group received a short but individualised training to improve teaching competence while the control group were given no specific programme regarding teaching competence. One year later, all participants were invited to sit the post-test. RESULTS: In all, 61 GP trainers participated in the pretest and 51 in the post-test; 44 GP trainers participated in both tests. Despite the large attrition, the internal validity of the experiment was preserved. Beginners gained a median score of 2.04 (on a scale of 0-5) on the first MSTAT. In the post-test, the experimental group (median = 3.12) scored significantly better on the entire test (Mann-Whitney U = 166.5, P < 0.05) compared with the control group (median = 2.84). However, the GP trainers in the control group also appeared to have made progress. CONCLUSION: General practitioners who start as GP trainers have insufficient teaching competence to guarantee good coaching of students. A personal programme leads to progress in teaching competence; however, it would seem that more time and support are necessary to allow GP trainers to gain full teaching competence. The fact that the control group made progress as well suggests that the test had an important learning effect.