Faculty: Mark Riedl, Scott Robertson, Rob Solomon
Students: Amelia Lambeth
This study employs gaming technologies and techniques to create an intelligent encapsulated conversational agent (ECA) to act as a virtual coach who will lower the cognitive effort required by prostate cancer patients to understand key aspects of decision-making, provide more appropriate reference points from which patients more accurately interpret personal risk, and frame information to optimize the patient’s chances of applying his own preferences and values to the decision at hand. A stylized, animated ECA will have a brief, focused conversation with a patient in order to explain, in layman's terms, the various treatment options and their risks and benefits and ask questions to assess the patient's medical literacy and values preferences, for example, the patient may value interventions with lower risk of side effects over being cancer-free. The eCoach ECA is being developed with the Unity3D game engine and uses gaming AI tools such as behavior trees to model a dialog and ECA behavior. The patient will respond to each ECA question by selecting from among several predetermined answers and the history of patient answers will determine how the conversation unfolds. For example, if the ECA determines that the patient is not sure about the risks and benefits of the various treatment options, it will spend more time explaining what these are as well as ask questions to assess knowledge of them afterward. This study represents a multidisciplinary collaboration between Emory University’s School of Medicine, the College of Computing and the Interactive Media Technology Center (IMTC) at the Georgia Institute of Technology.