Abstract Rationale Intercostal chest drain (ICD) insertion is a core procedural competency across Respiratory, Emergency and Acute Care medicine. Traditional instructor-led simulation is limited by faculty availability, resources and infrequent practice opportunities, impacting trainee confidence and patient safety. With increasing demand for scalable and efficient training, Augmented Reality (AR)-assisted simulation offers a self-directed, repeatable and standardized alternative. By providing accessible practice opportunities, AR can enhance procedural competence, confidence, and patient safety. Its adaptability encourages its use as a sustainable cross-specialty training tool. This study evaluated whether AR-assisted simulation, using a HoloLens headset, is non-inferior to traditional instructor-led training for ICD insertion. Trainee perceptions of AR as a learning tool were also explored. The primary endpoint was skill acquisition, with secondary endpoints including knowledge and skill retention, and trainee confidence. Methods Twenty-one interns and residents, novices in pleural procedures and not on a respiratory training programme, were randomized into two groups for ICD insertion training: 1) Control: instructor-led, or 2) Intervention: AR-assisted simulation, via HoloLens, using a pre-recorded tutorial. At baseline, participants completed a Knowledge Assessment post-teaching, with skill acquisition assessed using an Objective Skills Assessment. In the following six-week interval, participants were encouraged to arrange practice using their assigned learning method, with assessments repeated at the end to evaluate skill retention. Baseline and interval questionnaires captured participant perceptions and confidence. Results Post-teaching performance was equivalent between groups (mean Skills Assessment score 21/22, p 0.05) (Table 1). The Control group demonstrated significantly greater pre-teaching confidence, compared to the intervention group (p = 0.045). Despite this, post-teaching confidence improved in both groups. At the six-week interval, the intervention group achieved a comparable mean Skills Assessment score (20/22), with no statistically significant difference. Both groups showed interval decline in knowledge retention. Absolute scores in Knowledge and Skills Assessments were slightly higher in the control group, but no statistically significant differences were observed. Conclusion AR-assisted simulation was non-inferior to instructor-led training for skill acquisition in ICD insertion. Although baseline confidence was lower, AR enabled post-teaching confidence sufficient for comparable skill outcomes. Despite modest declines in knowledge retention, AR provides an effective, scalable and accessible alternative training platform. By supporting self-paced, flexible and standardized learning, it overcomes barriers of instructor availability and resource intensity; thereby promoting repeated practice and enhancing patient safety. Incorporating interactive features and real-time feedback could further optimise efficacy. Readily transferable across simulation-based procedures and specialties, AR offers a sustainable, learner-centred model for procedural training. This abstract is funded by: None
Wong et al. (Fri,) studied this question.