Providing Case-Based Retrieval as a Decision Support Strategy in Time Dependent Medical Domains - Computational Intelligence in Healthcare: Advanced Methodologies - page 201

Analysing long and complex time series of measurements at a screen or on
paper can be tedious and prone to errors. Physicians may be asked to recognize
small or rare irregularities in the signal, whose identification may be extremely
relevant for an accurate diagnosis or for the patient monitoring activity. A man-
ual identification of these irregularities requires a certain amount of expertise
in the specific field [37]; an automated decision support strategy is therefore
strongly needed in these domains.
Case-based Reasoning (CBR) [1], a reasoning paradigm that exploits the
knowledge collected on previously experienced situations, known as cases ,is
recently being recognized as a valuable knowledge management and decision
support methodology in time dependent applications (see e.g. [32]).
In CBR, problem solving experience is explicitly taken into account by storing
past cases in a library (called the case base ), and by suitably retrieving them
when a new problem has to be tackled.
CBR can be summarized by the following four basic steps, known as the CBR
cycle , or as the four “res” [1]:
- retrieve the most similar case(s) with respect to the input situation from the
case base;
- reuse them, and more precisely their solutions, to solve the new problem;
- revise the proposed new solution (if needed);
- retain the current case for future problem solving.
However, in many application domains it is common to find CBR tools able to
extract relevant knowledge, but that leave to the user the responsibility of pro-
viding its interpretation and of formulating the final decision: reuse and revise are
therefore not implemented. Nevertheless, even retrieval alone may significantly
support the human decision making process [54].
This consideration especially holds for the medical domain, where automated
revision/adaptation strategies can hardly be identified [29], while experiential
knowledge representation as cases and case-based retrieval can be particularly
helpful, for the following reasons:
- case retrieval resembles human reasoning in general, and medical decision
making in particular. As a matter of fact, physicians are used to reason
by recalling past situations similar to the current one. The process is often
biased by the tendency to recall only more recent or dicult cases, or only
the positively solved ones. Case-based retrieval can enable to store and recall
older, simpler or negative examples as well. Storing and recalling practical
cases comes out to be very useful also for sharing other clinicians' skills,
and for training un-experienced personnel, which are key objectives of every
health care organization;
- cases allow unformalized knowledge storage and maintenance . Properly
maintaining knowledge is a relevant issue to be addressed in the medical
domain, where large amounts of operative and unformalized information are
generally available. Every past case implicitly embeds an amount of domain
knowledge (i.e. the problem-action pattern adopted on that occasion), which