Simulation-based surgical training is now becoming mainstream. The novelty of its very existence has worn off and the medical community is increasingly integrating the few simulators on the market into their curricula. For many common surgical/procedural activities, it is no longer if it can be built, but what should be built.
While there do remain a number of difficult computational problems in representing the appropriate level of detail of real-time tool-tissue interaction, the initial problem that must be addressed before deciding what must next be solved is that there is little decomposition information on the proprioceptive and cognitive components of specific surgical activities. Creating an effective simulator requires this knowledge to focus development energy on simulating those aspects that matter most to building clinically useful skills. The means for developing this priority can be found in human factors approaches to task decomposition that have been used quite successfully in military and industrial training realms. In parallel with this is establishing objective quantitative measures of performance and associating the values from these with defining levels of proficiency. This is ultimately what the medical profession’s vested licensing organizations are most concerned with. How these measures are defined, implemented, validated, and proficiency defined relative to them is not an established process at this time.
Beyond all of the foundational content issues is the matter of the architecture and construction of the simulator itself. What was once a “roll everything yourself” undertaking can now depend upon a number of pre-existing open source and commercial implementations of technical subsystems. A key to using these is an effective understanding of the strengths & weaknesses of various solutions for the same specific technical subsection of the simulator and how this would integrate into a larger simulator architecture where other similar decisions are required. Beyond the simulator computational aspects is defining the content pipeline for creation of the teaching scenarios and integration with a learning management infrastructure. The availability of appropriate content and management of simulations within the context of the teaching process are as critical to the success of a simulator in training physicians as is getting the physics of tool-tissue interaction right.
In this talk I will discuss these issues in general and present how these were handled for two simulators in which I have been active in guiding their creation: an endovascular simulator now on the market for 6 years and an open surgery simulator currently under development.