Opportunities for building learning activities around real patients have decreased. Therefore, various forms of representative simulation have become an increasingly common alternative. The use of virtual patients is one such simulation developed to support the delivery of clinical teaching. Virtual patient scenarios offer opportunities for 'game-informed learning'. This is due to their experiential and problem-based learning approaches as prime pedagogic drivers. A region has been developed in Second Life that aims to deliver game-based learning activities for delivery of virtual patients that can drive experiential, diagnostic, and role-play learning activities concerning diagnoses, and selection of, investigations and treatment. The game-based learning activities for virtual patients were designed based on the four-dimensional framework developed by De Freitas and Martin, as well as other design considerations that look at emergent narratives and modes of representation. This paper will present the three-component architecture for the delivery of Virtual Patients following game-based learning activities in a multi-user virtual environment (MUVE), such as Second Life. The proposed architecture consists of different virtual patient components that provides the personal and clinical data relevant to the clinical scenario, a data availability model that enables the sequencing and progressive disclosure of a virtual patient identifying triggers and scaffolding information, and an activity model which encodes the activities available and how the learner will be able to engage with the virtual patient. It is anticipated that this three-component architecture will accommodate many different deployment and delivery options for virtual patients in a MUVE. An overview of recent trials is also provided. The trials aim to explore the experience of computer and videogame play among medical students, and to identify any gender-related differences and social propensities that might exist between high gamers (frequently play computer games) and low gamers (rarely play computer games) in their approaches to game-based learning in Second Life.