Objectives: The purpose of this study is to explore the relation between basic science knowledge and the ability to understand and make use of basic science in explaining a clinical scenario in the final year of medical school. Methods: A sample of senior medical students was re-assessed using the same test they had taken 3 years earlier. This was followed by an in-depth interview on one of the topics taken from the test. Their respective level of knowledge was compared with their performance in the interview. The test was analysed according to the revised Bloom's Taxonomy, and the interviews carried out according to the phenomenographic approach. Results: Performance was around 60(n=19) of the original performance, with no significant correlation between original test and re-test (r = 0.258, p = 0.29) and large interpersonal variation. A high performance in the original test did not predict a good performance; rather, the reverse seemed probable. None of the students who achieved high grades in the original test displayed a stable long term understanding that was measured in the interview. The test comprised questions of a generally low taxonomical level, but could not explain the mismatch between test-result and level of understanding. Conclusions: Findings suggest substantial loss of basic science knowledge during medical training. Attention should be directed to designing examinations that are purposeful, when it comes to what kind of knowledge is desirable in medical graduates as well as how that knowledge should be acquired. Further larger-scale research is needed to corroborate these findings.