Agent Technology and Healthcare: Possibilities, Challenges and Examples of Application - Computational Intelligence in Healthcare: Advanced Methodologies - page 28

separate the knowledge source from its use. Examples of this kind of systems are the
following projects: the National electronic Library for Health (NeLH) [34], the Vir-
tual Electronic Patient Record (VEPR) [14], the Context-aware Hospital Information
System (CHIS) [46], and the proposal made by Choe and Yoo [9]. These systems
crawl proactively the sources in order to maintain an up-to-date repository of knowl-
edge, and at the same time, agents coordinate their activities in order to improve the
retrieving and processing of data.
Decision support systems . Approaches aimed to assist the professional in the execu-
tion of healthcare treatments. In this case, agents are used to retrieve, monitor and
decide which information is transmitted to the user. Users are usually represented in
the system with his/her particular agent and individual preferences. These systems
use a knowledge base to support the inference process. Case-based reasoning and
domain ontologies are two of the most used techniques to represent the medical
knowledge. Examples of this kind of applications are the following: the proposal
made by Godó et al. [22], the Singh's et al. [49] system, the HealthAgents project
[23], and the HeCaSe2 system [24]. From the analysis of those systems, the reader
can observe the flexibility that agents provide implementing three different topolo-
gies. At first, one problem is replicated in parallel among different sources that solve
the problem locally and send a (partial) result that is integrated. In a second case, a
complex problem is divided into several parts which are solved by individual agents
in a grid-like manner, and then, an agent aggregates the received solutions. In last
case, the problem is solved co-ordinately by agents exchanging the appropriate medi-
cal information.
Planning and resource allocation . Systems centred on the coordination and schedul-
ing of human and material resources. Communication and coordination, which are
basic characteristics of agents as stated previously, are extensively exploited in this
kind of systems due to the required negotiation among different partners, taking into
account different constraints, variables and features (which may introduce potential
contractions between them). Good examples of this category are: the Agent.Hospital
infrastructure [32], CARREL [13], the Medical Information Agents project [6], and
the Operations Management in Healthcare project [35]. The agents of these systems
replicate real world behaviours in order to realistically automate processes. Mainly,
they decompose a problem in smaller units which are easier to deal with and which
are assigned to individual agents [31].
Composite systems . Systems which offer complete and integrated solutions for
healthcare management for a concrete organization. These kinds of systems combine
different AI techniques with a particular purpose under the umbrella of e-Health.
Assistance to elder or disabled citizens (such as the projects SHARE-IT [12], K4Care
platform [8], and the Geriatric Ambient Intelligence [10]), and community care (such
as the INCA project [3]), are two examples of this area of application. These systems
can be considered as final applications meant to substitute traditional and ad hoc solu-
tions currently running in medical organizations. The maturity of some of these sys-
tems also shows the feasibility that agent technology can offer to the healthcare
domain.