BackgroundAugmented reality (AR) technology involving head-mounted displays (HMD) represents a significant innovation in medical education, particularly for training in guided invasive procedures. Novice physicians often face challenges in simultaneously identifying anatomical landmarks and performing procedures when learning point-of-care ultrasound (POCUS). Our primary objective was to determine the effectiveness of AR in enhancing physician training for ultrasound-guided interventions using AR visual overlays. The secondary objective was to compare cognitive load between traditional ultrasound training settings and AR-assisted training settings.MethodsThis randomized crossover study, conducted from 2021 to 2022, compared performance and cognitive load of trainees attempting accurate central venous catheter (CVC) placement using an AR-HMD to display ultrasound images (AR-US), compared with standard ultrasound without visual overlay (S-US). We enrolled 47 trainees, including 22 final-year undergraduate medical students and 25 postgraduate physicians (years 1–4) from three hospitals in Taiwan. All participants had basic training in US-guided CVC placement but lacked AR-US experience. Performance and cognitive load were assessed using time measurements and NASA Task Load Index (NASA-TLX), covering the dimensions of Mental-, Physical- and Temporal Demand, and Performance, Effort and Frustration.ResultsWe found AR technology reduced the time required for critical steps in CVC placement, while minimizing users’ neck strain. Female and junior trainees using AR-US identified anatomy and localized veins faster than those using S-US. Using AR-US, female trainees significantly outperformed males in anatomical identification [mean difference (MD): -10.79 s (95% CI: -2.37 to -19.21)]. The NASA-TLX cognitive load assessment showed mental workload trending lower in all dimensions except performance while using AR-US, compared to S-US. Similarly, junior trainees’ effort- and frustration scores were lower [MD: -2.73 (95% CI: -5.04 to -0.41) and -2.41 (95% CI: -4.51 to -0.31), respectively], as were female trainees’ effort scores [MD: -3.07 (95% CI: -6.10 to -0.03)] when using AR-US, compared to S-US, whereby these differences were statistically significant.ConclusionsAR technology helped improve trainee performance and reduced cognitive load during ultrasound-guided CVC placement. These findings support the application of AR technology to enhance physician training in ultrasound-guided interventional procedures, suggesting that AR-HMDs could be a valuable tool in medical education.Trial registrationThe trial was registered with Clinicaltrials.gov on 20/09/2023 as a clinical trial, under the identifier NCT 06055400.