Information Technology Reference
In-Depth Information
and adolescents' health. It is also noteworthy that due to the rareness of the RDs,
it is not financially reasonable for pharmaceutical companies to produce medicines
to treat them; therefore, the relevant research and production are quite limited. Be-
cause of the limited number of cases it is quite difficult to establish the efficiency
of the therapeutic interventions by employing the classical methods. This fact in
its turn indicates the advisability of elaborating new methods and means of treat-
ment and assessing their efficacy [15]. The current situation indicates the gravity of
the problem and the necessity to find the ways to solve it, not only in a particular
country, but globally as well.
25.2
Description of the Approach
The main goal of our investigation is to elaborate a methodological approach that
can be used in computer-assisted diagnosing to help a physician to suspect the RD.
This can be a part of an algorithm, to be used as a basis for a computer-assisted med-
ical decision support system. For our purposes we choose the fuzzy logic framework
(the reason will be explained later in this section).
We do not consider rare disease's cases when the outstanding symptoms directly
exhibit a disease, rather we deal with cases when the RD is masked by a common
disease.
We assume, that a process to suspect the RD in a given patient is as follows.
In daily practice, it is very seldom, that a patient will be given an exact diagnosis
about the presence of some RD during his/her first visit to a practitioner: “a rare
disease has such a low prevalence in a population that a physician in a busy general
practice would not expect to see more than one case a year” [27]. Moreover, quite
frequently the description of the RD is related to very expensive and invasive lab
tests, investigations, etc., and it is not realistic to expect all the patient to undergo
such non-standard procedures.
In our approach we assume, that a patient was given a preliminary diagnosis (that
could be wrong), but some hesitation about “normality” of the proposed diagnosis
was evident. Then a physician would ask a computer system, if there are any devia-
tions from the “normal” (typical) case. If deviations were confirmed by the system,
it could be considered as a sign of possible RD.
Thus, a physician (who uses the computer decision-support system) should be
alarmed by a computer program about a possible rare disease. Our task is to describe
an appropriate algorithm for such computer program, i.e, the program should warn
the user about the possible abnormal cases, which are “outside” of the scope of the
possible diagnoses of the system. It can be considered as a first step, and the next
step might be monitoring incl. performing of additional investigations, using the
special search machines in order to decide finally whether we have only atypical
case of the common disease or some other diseases, including a rare one.
A search of the RD can be continued in the specific databases. For instance, one
can found a description of RDs in the database like ORPHANET [23], some other
web portals [5], [22]. A search for an appropriate case can be continued there.
