Page 214 - Special Topic Session (STS) - Volume 4

STS580 Vassilis P. P.
More specific, supervised learning approaches have a wide application in
the domain of health care fraud detection. Indicatively, several studies have
been proposed supervised learning based models for to healthcare fraud
detection including classification schemes such as Neural Networks [8],
decision trees [9,10] and Support Vector Machines [11, 12]. Concerning the
unsupervised learning perspective, new types of fraud can be uncovered
through the application of this category. usually the relative approaches are
based on clustering methods. The first choice of selecting an unsupervised
approach is data clustering, while fraud can be detected through samples that
are nor members in a dense cluster or samples that are too far from the center
of cluster. That means that this sample has less shared features among other
samples and should be further evaluated. In recent literature there is a
plethora of clustering approaches for fraud detection [13–17].
Also, an effective way is to search for samples outliers, which are potential
fraud samples since they do not follow the behavior of the other samples [15].
Approaches using association rules is also an efficient manner for detecting
healthcare fraud [18]. Furthermore, some studies have been integrated
supervised and unsupervised methods proposing hybrid approaches for
healthcare fraud detection. An indicative example is the study [19], where the
authors examined an electronic fraud detection program that compared
individual provider characteristics to their peers in identifying unusual provider
behavior.

3. Analysis - Methodology
The data studied in this paper are from the National Organization for the
Provision of Health Services (EOPYY), the main public purchaser of health
services in Greece. EOPYY founded in early two thousand twelve, so it is still
taking its first steps as i) a buyer of Health Care Services for Greek citizens and
their families, ii) an assessor of Quality and Safety Services, by establishing
rules in healthcare market, iii) an Health Technology Analyst of healthcare
products and iv) a negotiator with healthcare stakeholders. More specific,
purchasing enough and effective healthcare services for the insured citizens,
the pensioners and the protected members of their families, of the insurance
agencies that have been integrated with EOPYY, according to what is being
foreseen in the regulatory framework of healthcare services as every time in
effect. Also, the establishment of rules in designing procedures, in quality, in
development, in assessing the efficiency and effectiveness of healthcare
services market, the auditing of the funding process along with the
rationalization in the use of public funding. The establishment of the criteria
in the terms of the contracts with the providers along with the amendments
of the contracts terms whenever needed. It is worth mentioning that the
negotiation process with the providers regarding their remuneration, the

203 | I S I W S C 2 0 1 9