The results indicate that medical students benefit from endovascular simulator training as well as shadowing and increase their self-assessed IR knowledge. Simulator training led to a significant increase in perceived knowledge in the students' self-assessment regarding all items of basic IR skills. Shadowing showed a significant increase in self-assessed knowledge of instruments and Seldinger technique. The gain in self-assessed knowledge about the steps of the Seldinger technique was significantly greater after simulator-based training compared with shadowing.

Several studies investigated medical students’ receptiveness for IR revealing that curricular under-representation resulted in lack of knowledge and interest [1,2,3, 6, 7]. In a systematic review from 2019, Emin et al. examined the international representation of IR in medical schools and described poor IR knowledge in more than half of the students [6]. In response to poor knowledge and under-representation in medical school curricula, the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) published the first "Curriculum of Interventional Radiology for Medical Students" in its second edition in 2019 [8].

Referring to the CIRSE publication, Shaikh et al. studied the impact of a ten-hour lecture series based on the CIRSE curriculum on medical students. The effects of the intervention were tested through a quiz on IR before and after the 2-day theoretical teaching. After the lectures only 17% rated their IR knowledge as poor or non-existent compared with other disciplines after the teaching format compared with 62% before the intervention [9].

In contrast to theoretical teaching, other studies looked at the effects of hands-on training on students [4, 10,11,12]. In 2020, Stoehr et al. compared the effect of theoretical teaching with endovascular simulation-based training on fourth-year medical students regarding the change in IR knowledge and enthusiasm. Students participated in a 90-min training using an endovascular simulator. Knowledge and interest in IR were documented on a Likert scale. Both teaching methods led to a significant increase, whereby the endovascular simulator training was significantly more effective compared with the theoretical seminar. Stoehr et al. summarized the increase in knowledge as "knowledge of endovascular techniques" [4]. In comparison, in our study, we tried to obtain more precise information about which knowledge the simulator training conducted. In this way, we were able to shed more light on the effects on knowledge regarding instrument and material, the steps of the Seldinger technique and the steps of aortography, as basics of IR.

Some studies outside IR have already dealt with the effects of shadowing on medical students [5, 13, 14]. Evans et al. investigated the impact of a 5-day program mostly consisting of shadowing on the clinical skills of newly qualified doctors. Participants underwent pre- and post-program assessment using the Objective Structured Clinical Examination (OSCE) format, which tested basic clinical skills. The program led to a significant improvement in clinical skills and an improvement in confidence [15].

Teaching IR to medical students can be challenging considering curricular under-representation and limited teaching options. The analysis of the individual interventions’ sequence can be beneficial to the organization of training and an efficient transfer of knowledge. We found that greater overall knowledge gains were achieved in the group that received simulation-based training before shadowing. However, only regarding the steps of aortography a significant difference in gain of the self-assessed knowledge compared with group with an inverted teaching method sequence was found.

The small number of participants is one of the main limitations of our study. Moreover, in contrast to simulator training, shadowing elective live cases does not provide a reproductive structure. Self-assessment on Likert scale to measure the medical students’ knowledge is an established way to get an impression of a teaching effect [8, 14]. Yet response bias must be considered when using a questionnaire-based approach. It can lead to respondents answering questions untruthfully because they think they are expected to answer in a certain way. We used an anonymous study design to counteract this pitfall. For objective quantification of knowledge, other forms of evaluation such as a pre- and post-quiz, evaluation of performance on the endovascular simulator through an expert and measurement of time practicing on the simulator should be implemented in further studies. Moreover, we only analyzed the practical parts of the course and not the theoretical seminar on the 1st day.