MEDICAL DIAGNOSIS BY INTERVIEWING AVATARS

The Multi-Lingual Virtual Simulated Patient Project (MVSPP) involves the creation of a virtual simulated patient that simulates a real human patient presenting several symptoms for medical students and health care trainees in the field of primary health care. Students will be able to interview the patient to diagnose his/her diseases as a doctor would do in a real situation. The virtual patient can communicate using natural language and express different moods that depend on the diseases he/she suffers from and on the student's behaviour.
MVSPP builds upon a previous project run by IAVANTE (Spain) and the University of Granada (Spain), which developed an embodied conversational agent (ECA) that simulates a real human patient presenting several symptoms for medical students and health care professionals in the field of primary health care. Students interview the ECA to diagnose his/her diseases as a doctor would do in a real situation. This virtual patient can communicate using natural language and express different moods that depend on the diseases he suffers from and the student's behaviour. The ECA's behaviour is done by means of the coordination of several modules devoted to different tasks: natural language understanding, dialogue management, emotional state control and natural language generation. An ontology that gathers the domain knowledge of the agent specifies a semantic language that the modules use to communicate themselves.
Currently, this sort of simulation is usually done through the use of actors, a practice which is widespread as it does effectively permit the detection of errors in the clinical interview with the patient. However, the process is very costly, both in time for the training of the actors, as well as the need to physically bring the actors and the trainee doctors to the same geographical location, which apart from logistical issues, also disallows the use if e-learning as a mode of provision for such courses.
The MVSPP project is taking the work done on the ECA forward in two different ways. First it leads to new simulated patients developed in six new languages, namely English, Italian, Portuguese, German, Hungarian and Bulgarian. Secondly, these language versions are being adapted to produce so-called 'non-native' versions, i.e. virtual patients that speak like a member from an ethnic minority in the respective country, e.g. a Turkish patient speaking in German at a German clinic.
Arguably the most ambitious aim of the project is however in fact promoting the use of VSPs, aiming at the mainstreaming of their use within educational systems. The project partners all believe that VSPs can become an important and effective adjunct to the current suite of training tools used in medical education, and in this belief will organise pilot courses and dissemination seminars in each of their countries, so as to help test the VSPs created, but also, and more importantly to demonstrate the use of them in practical settings and to discuss the advantages of them both with this practitioners as well as with stakeholders in the system.