[en] Owing to the establishment of a procedure aiming at a reduction of the medical supply (numerus clausus), the French Community of Belgium had decided to introduce, at the end of the third year of medical studies (third candidature), a procedure of selective attribution of the certificate allowing for a unlimited medical practice. We assessed the predictive value of the results obtained during and at the end of the first candidature (first year of medicine) for the attribution of this certificate. We studied the results of the 225 students enrolled in the first candidature in medicine at the University of Liege during the academic year 1997-1998, and we correlated them to the subsequent attribution (1999-2000- and 2000-2001) of the certificate. The success of the first candidature in the first or second session is an important predicting factor towards the reception of the certificate. A non-authorization to continue the studies in medicine in case of failure in the first session or a non-authorization to repeat the first candidature in case of failure in the second session would have, respectively, a negative predictive value (NPV) of 91.9% and 97.4% and would allow to restream 111 students at the end of the first session or 39 students at the end of the first candidature. The results of the preliminary assessments mode during the month of January are also highly predictive of the chances to get the certificate. Among the 160 students who have not successfully passed a minimum of 3 exams, only 14 students obtained the certificate. The restreaming of those students would correspond to a NPV of 91.2%. Students who were not successful in a minimum of 2 exams would have less than 5% chance to obtain the certificate. In this case, 109 students would be restreamed (NPV = 95.4%). In conclusion, the selection of the students who will obtain, at the end of the third year of medicine, the certificate allowing them for an unlimited medical practice is obtained, in a large proportion, during the first candidature and especially after the session of January. In view of those results and considering the decree abolishing the selection at the end of the third year of medical studies, one could argue about the appropriateness of the current procedure, chosen by the French Community of Belgium, compared to other solutions prioritising an earlier selection of the future doctors.