|M.Sc Student||Gershov Sapir|
|Subject||Autonomous Medical Simulators|
|Department||Department of Autonomous Systems and Robotics||Supervisor||DR. Shlomi Laufer|
In critical and complex life-and-death situations, such as during complex resuscitation procedure, anesthesiologists’ decision-making is of utmost importance. Thus, clinical decision support systems (CDSS) have been deployed to assist the medical staff by enhancing clinical decisions. In a field where seconds can make the difference between life and death, integrating an autonomous CDSS framework capable of predicting medical treatment planning and assisting accordingly, can save lives. This requires the framework to accommodate a certain level of awareness and understanding which will not affect clinicians’ work except in cases it is required. In addition, the system must withstand diagnostic ambiguity and chaotic environment.
In this paper, we describe a technique for mining speech uttered during medical simulations to automatically create plans of resuscitation procedures, which leverage graph networks and language models. Furthermore, during complex resuscitation, we describe a technique for recognizing and monitoring medical treatment plans and predicting physician next action. This can be used to save time by prepping the required instrument in advance. This autonomous CDSS can be used to assist anesthesiologists during medical emergencies, and our simulations shows it can save precious minutes in prepping adrenaline dosage, which is crucial for a successful resuscitation.