This study was deemed exempt from full review by the University of Michigan’s Institutional Review Board according to national regulations and therefore informed consent from participants was not required. This manuscript follows the modified STROBE guidelines for reporting the findings of observations studies in simulation education [18].
Program development
Content for the MR programs was defined by expert opinion (board certified emergency physician and anesthesiologists) using an iterative approach. Content scripts were defined, and mock-ups of the graphics reviewed and revised. Graphics were designed using both 2 and 3-D software. The principal graphic design programs utilized included Maya, 3D Studio Max, Adobe After Effects, Adobe Photoshop, and Macromedia Flash. Using these programs, models were developed and 3-D visuals rendered using the Unity game engine together with ArtToolKit. ArtToolKit is a powerful mixed reality application that allows for optical tracking e.g., movement of a catheter into a vein. Once the models were created, content was merged into procedure-specific interactive modules which, in concert with HoloLens technology (Microsoft Corp. Redmond WA), allows the user to view superimposed internal anatomical features on a mannequin via computer-generated images (Fig. 1). For the purposes of this program, the HoloLens was used to help the participants visualize the major arm veins superimposed on a Laerdal mannequin arm during catheter placement (Fig. 2).
Usability testing
Once developed, the mixed reality program was tested for usability by 7 healthcare individuals with varying degrees of experience (none to experienced) with PIVC placement. Each individual attempted PIVC placement using the HoloLens and Laerdal arm mannequin. Individuals used a ‘talk back” approach to elicit feedback on their experiences with the mixed reality program. Comments were then relayed back to the development team for further review and refinement.
Participants
The study population included medical students attending the University of Michigan’s Medical School together with residents, fellows, faculty, nurses, and nurse anesthetists (CRNAs) from the Departments of Anesthesiology and Pediatrics. Participants were identified through University of Michigan e-mail listings and face-to-face recruitment.
Several days prior to the start of the study, eligible participants were e-mailed a link to a short informational video of the MR technology and PIVC trainer. This video was also available to participants on the day of study who again had a chance to review and discuss with the investigators. Due to safety protocols related to the COVID-19 pandemic and minimization of disruptions to the MR calibrations, the participants did not wear the Hololens. The Hololens was secured to a stand over the mannequin arm and the participant was guided by the HoloLens projected images on to a monitor of the internal anatomy of the arm and arm veins where they would then attempt cannulation of the mannequin arm using a standard IV needle and catheter. This set-up can be seen in Fig. 2. During this procedure, participants were evaluated in real-time by one of the authors (LR, EP or AT) using a checklist approach for correct identification of the insertion site, angle of approach, and insertion and advancement of the needle and catheter. Successful placement of the PIVC on the first attempt was documented. Complications such as the development of a hematoma or extravasation were also documented together with the number of subsequent attempts required until successful placement. Following attempts at cannulation, participants completed a 28-item online survey (Qualtrics) to elicit information about their level of training, prior experience with PIVC placement, perceptions of the trainer and thoughts on the use of MR technology as a training tool for medical procedures in general. Survey items included both closed and open responses. Two weeks following the study, participants were e-mailed a link to a short 6-item online survey to elicit information about any experiences with PIVC placement in the clinical setting subsequent to their using the MR trainer and, its impact, if any, on their approach and success rates.
Statistical analysis
Data from the Qualtrics surveys were downloaded directly to SPSS version 25.0 (IBM Inc., New York, NY) for analysis. Qualitative data were described using frequency distributions. Data are described as n (%) and mean ± SD. Free-text responses to open-ended questions were evaluated qualitatively and the most common themes reported.