The objective of this project was to enable close-to-real life medical simulation experiences through the integration of natural language processing and advanced AI technologies with medical manikins and virtual patient simulators. The team defined, developed, tested, and demonstrated a robust patient simulation software architecture usable for building intelligent patient simulators that provide realistic medical training experiences.
The solution provides the ability for students and instructors to ask questions and receive responses from the simulator(s) in natural language and execute voice commands from the instructor to change discrete simulated patient physiological parameters on the fly. Over time, and with continued use, the AI and natural language processing becomes more precise in its ability to return realistic responses. The students and instructor are able to carry out simulation dialog without having to speak in pre-defined manner. The software is designed to integrate seamlessly with both manikins and screen-based patient simulators widely in use across the DOD. It also integrates with existing intelligence modules such as physiology engines, behavior models, and adaptive testing. The methods facilitate multimodal dialog and feedback between the user and system, working with the user to understand, clarify, and effectuate the user's intent.
It is routine for instructors to be overwhelmed running simulation components while also observing, evaluating, and recording learner performance. The solution also reduces cognitive load for the instructor/facilitator through the integration of language and A.I.-based intelligence models that also become more precise over time with repeated use.
The intelligent patient simulation software API (application programming interface) resulting from this project includes test modules related to Tactical Combat Casualty Care and enables use across several languages. It also provides a value added improvement for commercial manikin and virtual patient simulation products supporting medical training of learners with different levels of clinical knowledge and with degrees of difficulty.