499 - A Randomized Trial Comparing a Novel Web-Based Ambient Augmented Intelligent Cognitive Aid Vs. Standard Cognitive Aid in Simulated Pediatric Resuscitation
Assistant Professor University of Michigan Medical School Ann Arbor, Michigan, United States
Background: Extended reality applications that enhance physical environments with superimposed digital images are increasingly being used in medical care. However, efficacy, safety, and acceptability of novel technologies compared to current care delivery is unclear. We developed a novel web-based platform, Ambient Augmented Intelligence in Medicine (AAIM) an interactive, multimodal, electronic cognitive aid to superimpose digital Pediatric Advanced Life Support (PALS) resuscitation algorithms on the physical world to aid decision making during pediatric critical care. Objective: Randomized control trial (RCT) comparing the efficacy and acceptability of PALS algorithms delivered via AAIM to currently used cognitive aids in improving adherence to resuscitation guidelines in simulated settings. Design/Methods: We designed a multicenter simulation-based RCT. Emergency medicine residents were randomized to 1 of 2 study groups: an intervention group where PALS algorithms were delivered by AAIM available on multiple devices (cell phones, laptops, and augmented reality device (HoloLens); and a control group that used the PALS pocket card. Each participant managed the same standardized simulated scenario of pediatric cardiac arrest. Participant’s ability to complete each of the critical PALS algorithm steps were assessed using a 7-item, dichotomously scored, checklist by an independent provider. We compared the groups’ time (in seconds) between resident’s recognition of cardiac arrest to first treatment (defibrillation) (primary outcome) and frequencies of successful completion of the PALS algorithm checklist, measured by summed score, using Kruskal-Wallis test (secondary outcome). We measured acceptability by Modified Technology Acceptance survey. Results: Twenty-one participants were analyzed, 12 in the AAIM Group and 9 in the PALS card. There was no difference between groups in time to defibrillation between the AAIM and PALS groups, having mean times of 84.58 (SD=60.43) and 75.67 (46.59) seconds. There was also no difference in PALS algorithm compliance between groups when comparing checklist scores. The AAIM Group had a mean sum score of 6.42 (0.90) while the PALS Group had a mean sum score of 6.22 (1.40). The group that used the device had a slightly higher acceptance of AAIM and desire for future use in patient care (0.75 to 0.33).
Conclusion(s): Extended reality that augments clinician’s physical environment with digital images was found to be safe and acceptable in pediatric simulated resuscitation but its application in real world contexts need exploration.
