In total, N = 157 students enrolled and completed the serious game course. Of them, 77.7% (n = 122) were female and 22.3% (n = 35) were male. Characteristics of the cohort are summarized in Table 1.

Around a third of students (30%, n = 46) indicated that they use books as their primary learning medium, and 21% (n = 33), respective 17% (n = 26) that they already used apps or videos as a learning medium. Although only some participants (n = 31) play online games in their leisure time, participants stated pre-game experience that they were open to online learning games to deepen and consolidate their knowledge (mean 4.8, SD 0.6).

There was high agreement of a mean 4.2 (SD 0.8) on a 5-point Likert scale, that ADRs were realistic and transferable to clinic. The usability of the serious game was rated as good, with a SUS mean of 77.7 (SD 15.3). See Table 2 for an overview of the self-assessments pre- and post-serious game play.

All items ranked anonymously on a 5-point Likert scale from 1 “don’t agree at all” to 5 “agree totally.” Item 3 was only assessed once after using the serious game. SD standard deviation.

Confidence to handle medications and to prescribe them was low across time (pre-, post-game play) but still improved significantly after using SeeMe from mean 1.8 (SD 0.8) to 2.4 (SD 0.9), V = 790.5, p < 0.001, r =  − 0.36. Along with this, a certain anxiety to harm patients with prescribing a medication reduced slightly, yet significantly from 3.8 (SD 0.9) to 3.5 (SD 0.9), V = 3078, p = 0.001, r = 0.14. The wish for more support and training in the field of pharmacology was underlined by participants with an overall agreement of a mean of 4.4 (SD 0.7) before and 4.3 (SD 0.7) after playing the game (V = 1231.5, p = 0.02, r = 0.12). Students stated that their knowledge improved modestly after using the serious game with a mean 3.8 (SD 0.9), although it was still considered to be low. Yet, students highly agreed that the use of the serious game made the topic more accessible to them (mean 4.2, SD 1.0).

Due to a technical error, the exam results were only saved for the years 2021 to 2022 (n = 110). Test results in the formative exam during these years improved significantly from 31% (9.4, SD 4.9) correct answers before playing the game to 55% (16.6, SD 5.1) correct answers after game play (Fig. 2), i.e., a significant main effect of time (pre, post) was found, F(1,98) = 324.97, p < 0.001. Notably, the marginal R2 for the fixed effects was 0.56, and the conditional R2 for both fixed and random effects was 0.75. Further, there was a significant effect of study year, and students in higher semesters reached better exam results, F(5,98) = 14.63, p < 0.001. However, the interaction between study year and time (pre-, post-game play) was not significant, indicating that the year of study did not influence the increase in knowledge, F(5,98) = 1.25, p = 0.29. Similarly, no significant gender differences were observed in exam results, F(1,98) = 0.55, p = 0.46.

Boxplots showing correctly answered questions of the exam before and after using the serious game

More than half of the students, 56.1% (n = 88) played the game at least twice a week. Students started 16,939 times a patient case, of which 98.1% (n = 16,615) patient cases were finished. Each student finished between 30 and 47 cases. On average, in 63.6% (n = 7408) of game sessions, students agreed that patient cases were realistic, and in only 31.2% (n = 3652) of game sessions, students agreed that patient cases were easy. Students used the options to ask questions to the patient between 0 and 27 times and requested help from the fictional doctor within the game between 0 to 42 times. In 70% (n = 11,865 out of 16,939) of game sessions, students identified the presence or absence of ADRs correctly.