A virtual simulator created by the Center for Safety, Simulation & Advanced Learning Technologies (CSSALT) has been enhanced with artificial intelligence to help students learn the fundamentals of surgical site skin preparation, all from the comfort of their own environment.
With the goal of improving skin preparation training, David Lizdas, a CSSALT simulation engineer, built the original skin preparation simulator in Macromedia Director and later in Unity, with industry funding obtained by Nik Gravenstein, M.D., the Jerome H. Modell, M.D., Professor of Anesthesiology.
Nurses and nursing students had the opportunity to experience the AI-based simulator during the UF College of Nursing’s 2022 Dorothy M. Smith Nursing Leadership Conference on Feb. 18. CSSALT developed the AI-based simulator training during the pandemic along with Ira Harmon, a UF Ph.D. student in computer science in the laboratory of Daisy Wang, Ph.D., who created the AI component.
“An innovative aspect during the workshop is that we actually transfer control of the simulator to someone in another location,” said Samsun Lampotang, Ph.D., FSSH, FAIMBE, the Joachim S. Gravenstein, M.D., Professor of Anesthesiology and Director of CSSALT. “This is an exciting way to allow learners to practice this technique in a safe environment while complying with social distancing requirements.”
Surgical site skin preparation is a key operating room skill that can have important implications for patient safety.
CSSALT engineers rose to the challenge of teaching during a virtual conference, setting the simulator so that the training could be conducted entirely via Zoom. Lampotang ran the simulator on his computer but transferred control of the cursor to an attendee. The participant was then able to perform a surgical site skin preparation by dragging their cursor on a simulated patient on their screen.
AI provided virtual coaching in real time, displaying feedback at the top of the screen. It detected whether the learner used a concentric or spiral approach to the skin preparation, which is no longer considered the gold standard, rather than a back-and-forth method. It also provided feedback on performance aspects such as whether the learner selected the proper drying time and provided adequate prepping in at least two directions. At the conclusion of the activity, the learner was scored and debriefed on performance.
“Before, a teacher would need to be present in person to provide that feedback, but this way the AI picks up the pattern,” Lampotang said. “One of the main learning objectives is to ensure that everyone is updated on the correct way to perform skin preparation. Another objective is to see AI in action in terms of virtual coaching, meaning that a human instructor does not need to be there to identify whether the preparation pattern is appropriate or outdated.”
Virtual coaching is increasingly important not only in pandemic situations, but also as a way to effectively use instructors’ limited and valuable time.
Going forward, a team at UF plans to conduct a study to determine whether training with the simulator can decrease surgical site infections in the hospital, which is a key metric of patient care. These infections can be caused by poor patient hygiene, as well as improper skin preparation.