Training military medical personnel to maintain readiness for medical emergencies and combat-related operations is a critical problem. Distance learning solutions are required for providing effective training while minimizing time away from the important peacetime duty of providing quality medical care to military personnel. Since medical emergencies are unexpected, it is important to dynamically generate customized courses to address the particular emergency. Since time is at a premium in such situations, it is important to address the precise learning needs of the medical team being trained.

We are developing an Intelligent Tutoring System (ITS), called ADAPT-MD, for military medical teams in combat and emergency procedures. Using a scenario-based approach, this system provides adaptive instruction that is customized to individual teams and their members. Team training poses challenges beyond individual training and few ITSs address this problem. Issues like student modeling, team performance evaluation, tailoring the challenge level of scenarios to student expertise, etc. take on added complexity. Adapt-MD addresses these issues by using a compositional approach to scenario generation and student model representation. A student model of a team comprises of a model of the team as a whole and models of each of the individual members. In addition to representing each members own state of expertise, the student model also represents his knowledge of the other team members' tasks and abilities. ADAPT-MD has facilities for creating scenarios that are adapted to individual team members' expertise levels. Simulations can include simulated intelligent entities to take the place of team members. The ADAPT-MD framework includes an authoring tool for specifying presentation content, domain knowledge, training scenarios, and instructional strategies. This framework is currently being applied to create an ITS for training hyperbaric treatment teams. It is, however, domain-independent and can be used to create ITSs for other medical domains.