Agent Technology and Healthcare: Possibilities, Challenges and Examples of Application - Computational Intelligence in Healthcare: Advanced Methodologies - page 30

the services they can provide ( i.e. when a SA enters the system, it sends a message
detailing its services to the associated MCA or DA).
In addition, each department contains a Guideline Agent (GA) that performs all ac-
tions related to guidelines ( e.g. , it can retrieve the CG associated to a specific dis-
ease). This GA contains only CGs associated to the department where it is located,
allowing versioning of CGs authored by the practitioners.
The Ontology Agent (OA) provides information about the concepts that appear in a
guideline - medical and/or organisational terms - (know-what) and their relations
(know-how). The OA uses the medical ontology designed to represent all relation-
ships between all kind of actors, their roles, and allowed activities. There is one OA
per department in order to locate the information close to the clients, but at the same
time, different OAs can use different ontologies, for instance, to represent different
roles or responsibilities of the medical staff, or different available resources in a
department.
At the bottom of the architecture, a Medical Record Agent (MRA) controls the ac-
cesses to a database that stores all medical records of the patients of the medical cen-
tre. Appropriate security measures have been taken to ensure that only properly
authenticated and authorised agents may access and update the medical records.
Agent interactions are implemented over communication protocols and sharing a
common knowledge representation. Mainly, agents use well-known protocols such as
FIPA-Query, FIPA-Request and FIPA-Contract Net [17]. The MAS has been imple-
mented using JADE, which is compliant with the FIPA specification for agent develop-
ment [5]. Moreover, HeCaSe2 has been designed following an agent-oriented software
engineering methodology (INGENIAS [42]) in order to provide a more general and
reusable framework, documenting all the services, and improving further updates [25].
4.2 Task Coordination during the Treatment
One of the main advantages of using agents in healthcare is the possibility of coordi-
nating the activities or tasks that cannot be achieved by one entity due to their inher-
ent complexity [30]. In the case of healthcare, a practitioner may use the facilities
provided by the HeCaSe2 system during the treatment of a patient, which consists on
the enactment of a CG. A CG incorporates the declarative knowledge related to the
treatment of the patient, including the steps to be followed: actions, enquiries and
decisions. Depending on the case, the practitioner may require the support of another
colleague or an external service.
In addition, HeCaSe2 guides this search taking into account preferences of the user
(patient) about a set of criteria such as the medical centres, the days and the period of
the day in which a medical test should be performed. This negotiation-based proce-
dure balances the load among centres (agents send different proposals of free slots)
and, at the same time, permits the user selecting the most appropriate alternative ac-
cording to his preferences. This is a flexible approach that allows the managers of the
medical centres the analysis and supervision of the waiting lists of patients in particu-
lar services.
As said previously, one of the possibilities that a practitioner can find during the
execution of a treatment is that the patient requires an action . This specifies a medical
procedure that should be performed over a patient. At this point the MAS searches