Mediated
agent interaction
Centre for
Health Informatics
University
of New South Wales,
UNSW NSW
2055, Australia
ewc@pobox.com
Abstract: This
paper presents a framework for agent communication and its mediation by
technological systems. The goal of the framework is to provide quantitative
mechanisms that will allow principled decisions to be made about the use and
construction of mediating technological systems. Beginning with a simple model
of interaction between agents, a model of communication influenced by bounded
knowledge of others (or common ground) is developed. This leads to predictions
that agent interactions generate equilibrium phenomena where optimal levels of
grounding emerge over time between agents.
Enrico Coiera is a
Professor and Foundation Chair in Medical Informatics, in the Faculty of
Medicine, co-Director of the Centre for Health Informatics, and an adjunct
Professor, in the School of Computer Science and Engineering, University of NSW,
Sydney. His main interests are medical informatics with particular interest in
the merging of communication and information technologies in healthcare; and
Artificial intelligence , in particular machine learning , ILP and qualitative
reasoning.
After graduating from Sydney
University Medical School in Australia, he worked as a resident medical officer
at the Royal North Shore Hospital for two years. He then undertook a Masters
degree in Biomedical Engineering, followed by a PhD in artificial intelligence
at the School of Computer Science, University of New South Wales . His thesis
explored the design of an intelligent monitor system for use in Intensive Care.
He developed an interest in qualitative modelling and machine learning. He moved
on to be the Head of the Clinical Computing Department back at the Royal North
Shore Hospital, before moving to the UK. There he joined Hewlett Packard's
Bristol Laboratories in 1990, where he initiated and lead projects focussing on
the application of information and communication technologies. The projects have
included: Intelligent signal interpretation for patient monitors, User-interface
design for an anaesthesia workstation, Protocol-support systems for anaesthesia,
Protocol-support systems for home insulin management, Study of the role of
communication in healthcare and
Advanced mobile voice and data systems for healthcare workers.
In 1999 he returned to Australia
where I have set up a new research Centre for Health Informatics at the
University of New South Wales, and hold the Foundation Chair in Medical
Informatics.
On
Articulation and Localization
- Some Sociotechnical Issues of Design, Implementation and Evaluation of
Knowledge
Based Systems
Department
of Health Policy and Management
Erasmus
University Rotterdam
P.O. Box
1738, 3000 DR Rotterdam, The Netherlands
E-mail:
j.aarts@bmg.eur.nl
Abstract: The standard model
of clinical work is a fixed sequence of tasks covering diagnosis and treatment
of patients. Knowledge based systems have been designed according to this
sequence. This ideal typical approach accounts for the relative modest success
of knowledge based systems in healthcare practice. In reality however, clinical
work is highly contingent, ad-hoc and idiosyncratic and therefore hard to fit
into in a formal model. A physician is said to manage complex and diverse patient
trajectories. Therefore the concept of a trajectory should not only relate to
the course of the disease of a patient, but to all the organizational work
during that course as well. We will highlight two aspects of this 'messy' view
of clinical work and examine the consequences for the design, implementation and
evaluation of knowledge based systems. Articulation refers to the fact that a
lot of invisible work is being done in order to complete a visible task of a
physician. A physician may see a patient, but before she can do that a lot of
work has been done to assure that she actually sees the patient. Localization
refers to the fact that clinical work is being adapted to local and situational
circumstances. This is not primarily related to the variance in medical work as
a result of uncertain knowledge about the true clinical state of a patient, but
to the constant negotiating with colleagues, local opportunities and restraints
and the possibilities of protocols and technologies. In short, the way how
patient trajectories are being shaped by human and non-human elements. A
knowledge based system that has the potential of adaptability to patient
trajectories seems to offer new opportunities. Such an approach would place the
user in the centerfold of the design, implementation and evaluation of such
systems.
Jos Aarts is an Assistant
Professor of social medical sciences at the department of Health Policy and
Management of Erasmus University Rotterdam in the Netherlands and visiting
lecturer of health informatics at the University of Surrey at Guildford in the
UK. Until recently he held a position as senior lecturer of health informatics
at Fontys University in Eindhoven. He hold a BSc and MSc in experimental physics
from the Katholieke Universiteit Nijmegen in the Netherlands. His research
interests focus on the design and implementation process of clinical information
systems in an organizational context.
He is director of the health
information management specialization program of the four year MSc course of
health policy and management. He also coordinates the Dutch partnership of the
European MSc Health Informatics course, a joint collaboration of the University
of Surrey at Guildford, the University of Manchester and Erasmus University
Rotterdam. From 2002 this course, under the name of Master of Health Information
Management, will be continued at Erasmus University Rotterdam. He is the
chairman of Working Group 9 ‘Human and Organizational Issues of Medical
Informatics’ of the European Federation of Medical Informatics (EFMI). He has
published on educational and organizational issues in health informatics, and
presented papers at various conferences such as the AMIA Fall Symposium, Medical
Informatics Europe and MedInfo.