Health Telematics (AIM) Final Report


	Updated: Feb 1, 97
GAMES-II

A General Architecture for Medical Knowledge-Based System

Project Code: A2034

Project value: 4955 KECU

EC contribution: 2499 KECU

No of partners: 7

No of countries: 6

Duration: 36 months

Contact:
Dr. Francesco Bruno
SAGO S.p.A.
Viale A. Gramsci 22
I-50132 Firenze, Italy
Tel.: +39-55-247.78.31
Fax: +39-55-247.78.34

Overview

Knowledge-based systems (KBS) development must be based on research contribution in health care at an international level. In order to achieve this goal, the project team has defined an epistemological model of medical reasoning which may also be considered a state of the art contribution for an exchange of information in future KBS networks.

By offering faster production and service development for coping with complex and costly medical situations, the prototypes should attract industries operating in the informatics market. The project team has been active since the beginning in involving industrial representatives in project orientation. Physicians are mainly interested in visible and user-friendly tools. The environment is using development tools (UNIX, LISP KEE) which are considered best suited for achieving this goal. GAMES II will be implemented in key pilots (UCL, Ulm, Amsterdam and Pavia Universities, Geneva Cantonal Hospital).

Project purpose

The ultimate goal of GAMES II were the development of a comprehensive methodology for medical KBS construction. The main goals of the project were the following: 1) Developing an epistemological model of medical reasoning specifying the desired problem solving behaviour for a KBS through a sound epistemological analysis of the different types of knowledge involved, 2) Building a computational model of a target KBS, based on the epistemological model previously defined. 3) Implementing a set of frameworks allowing a user to represent chunks of medical knowledge adopting formalisms which exploit sophisticated and innovative methodologies.

Methods used

The GAMES II epistemological model of medical reasoning splits medical reasoning in two subsequent steps. The first one is aimed at selecting a set of hypotheses representing plausible solutions to the problem at hand, the second one is aimed at testing them. The model, dubbed Select and Test Model (ST-Model), cam represent the three generic Fundamental Medical Tasks, (diagnosis, therapy planing and monitoring). The ST-Model provides the basic model for KBSs development and considers two different kinds of knowledge taking part m the process: the application ontology and the inference model. The former organises domain categories and their relationships, while the latter describes the problem solving process m terms of the basic inference types (i.e. abstraction, abduction, deduction, and eliminative induction), together with the mappings to the ontological categories of the domain. Hence, the first requirement for GAMES II is that of making available tools for acquiring both the application ontology and the inference model underlying the behaviour of the target KBS .

Results

The principal products of the GAMES-II project are: M-KAT, QCMF, GAMES, and CUE. I (M-KAT is the nucleus of GAMES-II; it can be seen as am instrument for construction of GAMES-II based Decision Support System (G-DSS) or an environment of execution of G-DSS. 2) QCMF and GAMES are decision support products, each based on a particular methodology; like M-KAT, they enable both the design of new applications and the solution of new cases with already predisposed applications 3) CUE is a product which supports the construction methodology of the epistemological model; its output is the knowledge base and it allows the definition of ontologies in the standard language ONTO LINGUA.

List of Deliverables

Year 1

Epistemological model 1.0 (3/R/I)
Computational model 1.0 (6/R/I)
Preliminary classification of medical domains using GAMES (6/R/L)
Epistemological model 2.0 (9/R/P)
System models formulation and generation 1.0 (9/P/L)
State or event state chain models formulation and generation 1.0 (9/P/L)
Representation of temporal primitives (9/R/L)
Computational model 2.0 (12/R/L)
Identification of medical problems for evaluation of GAMES (12/R/P)
Identification of proposed architecture and adopted standards 1.0 (12/R/P)
Methods for temporal relations and trends (12/P/L)
Knowledge modelling editor 1.0 (12/P/L)
Decision tree induction tool 1.0 (12/P/L)

Year 2

Knowledge-base editor 1.0 (3/P/L)
Knowledge modelling advisor 1.0 (6/P/L)
KA tools: Background + user manual 1.0 (6/R/L)
System models formulation and generation 2.0 (6/P/P)
State or event state chain models formulation and generation 2.0 (6/P/P)
Representation of temporal uncertainty (6/P/L)
Model-base learning tool 1.0 (9/P/L)
System models outcome analysis n1.0 (9/P/L)
State or event or chain models outcome analysis 1.0 (9/P/L)
Knowledge modelling editor 2.0 (12/T/P)
Knowledge-base refinement tool 1.0 (12/P/L)
Computational model 3.0 (12/R/P)
Prototype knowledge-based systems in the exemplar areas 1.0 (2/P/L)
Framework for medical evaluation (2/R/P)
Identification of proposed architecture and adopted standards 2.0 (12/R/I)

Year 3

Knowledge-base editor 2.0 (3/I/P)
Decision tree induction tool 2.0 (3/T/P)
Knowledge modelling advisor 2.0 (6/T/P)
System models outcome analysis 2.0 (6/P/P)
State or event or chain models outcome analysis 2.0 (6/P/P)
Temporal features representation tool (6/T/P)
Proposed architecture and adopted standards (6/T/L)
Prototype knowledge-based systems in the exemplar Areas 2 (9/P/P)
The application of GAMES methodology to medical exemplar areas (9/R/P)
Results of medical evaluation of GAMES : recommendations (9/R/P)
Knowledge-base refinement tool 2.0 (12/T/P)
Model-based learning tool 2.0 (12/T/P)
KA Tools: Background + user Manual 2.0 (12/R/P)
System models tool (12/T/P)
State or event chain models tool (12/T/P)
GAMES final report (12/R/P)

List of Participants

Dr. Vassilis Moustakis Foundation of Research and Technology Forth Institute Of Computer Science Daidalou 36 PO Box 1385 GR-71110 Heraklion, Greece Tel.: +30-81-22.93.46 Fax: +30-81-22.93.42 / 3 E-mail: moustaki@csi.forth.gr	Dr. J.R. Scherrer Hopital Universitaire de Geneve Centre d'Informatique Hospitaliere Rue Micheli-du-Crest 24 CH-1211 Geneva 4, Switzerland Tel.: +41-22-22.62.01 Fax: +41-22-47.64.86 E-mail: scherrer@cih.hcuge.ch
Dr. A.Th. Schreiber University Of Amsterdam Dept. Social Science Informatics Roeterstraat 15 NL-1018 WB Amsterdam, The Netherlands Tel.: +31-20-525.67.89 Fax: +31-20-525.68.96 E-mail: schreiber@swi.psy.uva.nl	Dr. Mark Leaning University College of London Clinical Operational Research Unit Gower Street London WC1E 6BT, U.K. Tel.: +44-71-387.61.46 Fax: +44-71-383.47.03 E-mail: mark@uk.ac.ucl.stats
Prof. Mario Stefanelli Universita' di Pavia Dip. Informatica & Sistemistica Via Abbiategrasso 209 I-27100 Pavia, Italy Tel.: +39-38-250.53.50 Fax: +39-38-250.53.73 E-mail: stefa@ipveenbitnet	Dr. Hauke Kindler University of Ulm - Klinikum Inst. Occup. & Social Medicine FAW Albert Einstein Allee D-7900 Ulm, Germany Tel.: +49-731-50.14.80 Fax: +49-731-50.19.99 E-mail: kindler@dulfaw1a.bitnet



	Copyright 1997 © EHTO All rights reserved This server is the only official EHTO WWW knowledge repository. Mail suggestions to: webmaster@ehto.org