Intelligent clinical training systems.

Clinical medicine is one of the most chal lenging areas for education. The develop ment of clinical competence requires the assimilation of large amounts of knowl edge combined with acquisition of clinical skills and clinical problem-solving ability. Clinical skills include the technical skill in implementing a procedure as well as skill in patient consultation and physical exami nation. Clinical problem solving requires the ability to synthesize the information contained in a clinical case and to integrate it with the physician’s knowledge and experience in order to diagnose and manage the patient’s problem. It also requires the ability to work in teams and the ability to transfer one’s knowledge to unfamiliar situations such as rare problems, disasters and emergencies. Currently, training toward clinical competence follows an apprenticeship approach, which consists of close expert supervision while interacting with patients. This method of training can subject patients to discomfort, risk of complications, and prolonged procedure times, creating a clinical governance dilemma. At the same time, there may be limited access to apprenticeship training in more complex sce narios with corresponding difficulty training in a time-effective manner. Intelligent clinical training systems hold the promise to address many of these issues. A facilitating technological environment has emerged in recent years through the maturation of research in intelligent tutoring systems, medical simulation, and virtual reality (VR) techniques and the develop ment of Web 2.0 collaborative authoring and social networking tools. The field of intelligent tutoring systems has come a long way since its start in the 1980s. There is now a well accepted standard architecture for such systems [1] and a number of well developed and tested user modeling techniques such as Bayesian networks [2]. The field has matured to the extent that Carnegie Mellon University is now using intelligent tutoring as a key technology in its ambitious Open Learning Initiative [3]. Recent work on incorporating medical ontologies into intelligent tutoring systems [4] and on leveraging existing large-scale medical ontologies like UMLS [5] hold promise to increase the domain coverage and quality of interaction and to decrease the cost of producing such systems. Clinical training during the past decade has witnessed a significant increase in the use of simulation technology for teaching and assessment [6]. Medical simulations, in general, aim to imitate real patients, ana tomic regions, or clinical tasks, and /or to mirror the real-life circumstances in which medical services are rendered. The simulator response will vary according to user actions (for example, heart rate and blood pressure will change appropriately depending on the dose of a particular drug administered intravenously [7]). Training and assessment using these simulators can focus on individual skills (e.g., ability of a resident to intubate [8]) or the effectiveness of teams [9, 10]. The use of virtual and augmented reality techniques to create realistic simulations of the physical aspects of the clinical environment is attracting increasing attention due to the promise of creating high-quality training environments, and to the rapid development and decreasing cost of software and hardware, driven in part by develop ments in the computer game industry. Building upon successful VR simulations in specific areas [11–13], a stream of work has emerged to build generic opensource software toolkits for medical VR Methods Inf Med 2010; 49: 388–389