Increasing the flexibility of virtual reality-based medical simulation training through dynamically reconfigurable patient and scenario states: Virtual Manikins

Introduction: Virtual reality medical simulation training (VRMST) is a fast-growing modality for training and maintaining knowledge, skills, and abilities (KSAs) across various health professions. This growth is partly driven by several benefits over manikin-based simulation training (MBST), including lower acquisition and sustainment costs, increased portability with reduced set-up times, and unique opportunities for psychoenvironmental immersion. Recent advances in VRMST technology have resolved many prior limitations to adoption and use, but some common barriers remain. One common limitation is the provision of fixed scenarios with inflexible state flows, critical actions, and learning objectives. Though fixed scenarios can provide benefits through simplifying moderation and reduced preplanning time, they also can prevent educators from dynamically varying a simulation to adjust to the performance and training level of a specific group of trainees. Additionally, fixed scenarios can suffer from limited reusability, as trainees may recognize specific cues and features from memory on subsequent attempts rather than relying on clinical judgment. In this work, we addressed this limitation by creating a new VRMST simulation capability that enables educators to dynamically reconfigure elements of patient physiology and the global simulation state to overcome the limitations of fixed virtual scenarios.

Methods: An interprofessional research working group was formed to determine the project’s goals, objectives, and clinical requirements. This working group included clinical educators, VR simulationists, and VR engineers. Based on these findings, a set of implementation concepts was finalized using iterative refinement. A set of technical specifications was then created, describing adaptations to a commercially available VRMST system (VRMSS, SimX, Inc.). An iterative process was again used to refine the implemented adaptations until the working group assessed that the implemented functionality met project objectives.

Results and Discussion: The working group determined the following high-level plans for the capability: 1) dynamic scenarios should be focused on undifferentiated chief complaints (i.e. chest pain, community traumatic injury), 2) there should be no indicators to trainees as to the underlying diagnosis other than through clinical assessments, 3) moderators should be free during execution to select and modify lab results, vital signs, and other clinical findings in order to represent different underlying pathologies, and should be able to change findings dynamically during the scenario, 4) a variety of patients and environments should be selectable at runtime to eliminate non-clinical cues, and 5) moderators should be able to take on the role of non-player characters within the scenario at will. Based on these high-level plans, detailed specifications were created, and the required adaptations were implemented within the commercial system. After iterative refinement, the resulting capability was assessed to meet requirements by the working group. The result of the project was a novel capability for dynamic VRMST execution. Based on the project’s success, several projects for further work are now in process, including an expansion of chief complaints represented and an exploration of approaches to simplify moderation of the dynamic scenarios.