BackgroundVirtual reality (VR) offers an immersive and practical method for training medical skills, especially in emergency healthcare settings. However, it is unclear whether learning in VR will translate into real-world performance benefits. To explore these potential transfer effects, we examined the validity and fidelity of a bespoke VR environment for Basic Life Support (BLS) training, a generic skill in medical training programmes.MethodsTwenty-two medical trainees performed standardised BLS procedures within two simulation conditions: one in VR, using a Pico Neo 3 standalone system; the other in a real-world synthetic environment, which included a physical mannequin and resuscitation equipment. Patterns of task behaviour, workload, sense of presence, and visual attention were derived from user self-report questionnaires, video recordings, and eye-tracking data.ResultsData showed that the VR training environment was sufficiently high in face validity to immerse the participants, and that trainees were displaying realistic task behaviours and procedural actions. However, the fidelity of user interactions and movements in VR proved atypical, which seemed to disrupt participants’ attentional and motor responses.ConclusionsResults suggest that VR may have limitations for improving physical skills in the context of BLS training, yet be potentially valuable for developing task procedures and/or perceptual abilities.