[Predictability of the results of the 2004 French national-ranking exam]

To assess the individual predictability of success for undergraduate medical students taking the French national-ranking examination (examen classant national [NRE]) in June 2004.

In total, 86 students in Rouen were classified according to RNE scores. We used the Spearman rank test [rs] to correlate to the RNE score each student's previous test results during the medical curriculum: written exams (first two years of medical classes), clinical skill tests (third- and fourth-year) and the clinical and therapeutic synthesis certificate (CSCT), and the northwest region official RNE practice test. Each student's ECN rank was then predicted by a regression formula.

Written exam (r5 = 0.35), third- and fourth-year clinical skill (r(s) = 0.62), CSCT (r(s) = 0.49) and regional practice test (r(s) = 0.56) scores were strongly and significantly associated with the RNE classification. Of the 59 students who took all tests, 12 students (20.3%) diverged from the regression model: three students obtained a national classification substantially better than expected, nine ranked substantially lower than expected.

Test results during the undergraduate medical curriculum in a cohort of undergraduate medical students were strongly predictive of the RNE results. An identifiable group of students with unstable results could benefit from targeted teaching and individual support. The effect of knowledge of their individual position before the RNE remains to be assessed.