Metoda vrednotenja uspešnosti sistemov za obvladovanje prevar na področju zdravstvenih zavarovanj

• Main Author: KUMER, ALEŠ
• Other Authors: Bajec, Marko
• Format: Master Thesis
• Language: Slovenian
• Published: 2016
• Subjects: obvladovanje prevar; obvladovanje prevar na področju zdravstvenih zavarovanj; metoda vrednotenja uspešnosti sistemov za obvladovanja prevar; neposredni učinki obvladovanja prevar; posredni učinki obvladovanja prevar; PyMC; fraud management; fraud management in health insurance; performance evaluation of the fraud management system; direct effects of fraud management; indirect effects of fraud management; sami
• Online Access: https://repozitorij.uni-lj.si/IzpisGradiva.php?id=85738; https://repozitorij.uni-lj.si/Dokument.php?id=88877&dn=

Učinkovito obvladovanje zavarovalniških prevar ima lahko velik vpliv na konkurenčnost zavarovalnic na trgu. Potencialni prihranki, ki se akumulirajo v prevarah, segajo namreč tudi do 10% vseh odhodkov, ki jih imajo zavarovalnice s poplačilom škod, kar v svetovnem merilu znaša več sto milijard evrov. Možnosti za odkrivanje zavarovalniških prevar je seveda več, najenostavnejša je ročno pregledovanje majhnega dela škodnih zahtevkov, ki pa je že v sami osnovi neučinkovita in v veliki meri odvisna od sreče preiskovalca, da izbere ravno primere, ki jih kasneje prepozna kot prevaro. Boljše je, če si preiskovalci pomagajo z informacijsko tehnologijo, ki jim pomaga sistematično označiti vse primere, ki so po vnaprej določenih kriterijih prepoznani kot sumljivi. Na takšen način so preiskovalci osredotočeni zgolj na tiste primere, ki so potencialno zanimivi. Vpeljava informacijskega sistema za obvladovanje prevar v poslovni sistem zavarovalnice prinese tudi potrebo po merjenju učinkov takšnega sistema na njeno poslovno uspešnost. Omenjene učinke lahko v grobem razdelimo v dve skupini, in sicer na neposredne in posredne učinke. Neposredni učinki se nanašajo na število, vrednost in relevantnost odkritih prevar, posredni pa na oceno uspešnosti odvračanja ključnih akterjev oz. prevarantov od izvajanja prevar. Neposredne učinke je mogoče meriti in jih kot take neposredno upoštevati pri izračunu poslovne uspešnosti zavarovalnice, posrednih pa pri večini zavarovalnih vrst ne. Pri prevarah gre namreč za veliko stopnjo razpršenosti in tudi naključnosti, ki jo težko uokvirimo v napovedne modele, še posebej takšne, ki bi bili zmožni dajati ocene z zadostno stopnjo zaupanja, da bi jih lahko vključili v ocene prihrankov zavarovalnice. Pri določenih zavarovalnih vrstah pa je ocena posrednih prihrankov do neke mere izvedljiva, to so tiste zavarovalne vrste, kjer je število potencialnih prevarantov omejeno in relativno majhno. Za takšne sisteme lahko predpostavimo, da se posamezen tip prevare po odkritju in sankcijah ne pojavlja več oz. se pojavlja v manjšem obsegu. V takšnih sistemih je mogoče oceno dobiti na podlagi razlik v dinamiki pojavljanja sumljivih škodnih primerov. Primer takšne zavarovalne vrste so zdravstvena zavarovanja. Rezultat magistrskega dela je metoda, ki je namenjena razširjenemu vrednotenju uspešnosti sistemov za obvladovanje prevar na področju zdravstvenih zavarovanj, ki poleg neposrednih prihrankov upošteva tudi oceno posrednih učinkov, ki nastanejo kot posledica sistematičnega boja proti prevaram. Zdravstvene zavarovalnice lahko s predlagano metodo pridobijo celovitejši pregled nad uspešnostjo in učinkovitostjo procesov za obvladovanje prevar ter oceno vpliva le-tega na poslovno uspešnost celotne zavarovalnice. Efficient insurance fraud management can have significant effect on insurance companies’ competitive market position. Potential savings accumulated in fraudulent activities can add up to 10% of all expenses insurance companies pay for damage claims, which globally add up to several 100 billion Euros. There are various available methods to detect insurance fraud. The simplest one, that is to manually review a small number of insurance claims, is highly inefficient and its success rate is largely dependent on investigator’s luck to select the exact cases that would turn out fraudulent after investigation. It is more efficient for investigators to use information technology that systematically highlights all cases which are identified as suspicious by predetermined criteria. Thereby, the investigators can focus solely on the potentially interesting cases. The implementation of fraud management information system into the business system of an insurance company also demands measuring the effects of such a system on its business performance. These effects can be divided into two categories: direct and indirect. The direct effects relate to quantity, value and the relevance of detected frauds, while the indirect effects give an estimation of fraud prevention effectiveness among individuals who are most likely to commit fraud. Direct effects can be measured and directly taken into account when calculating insurance company’s business success. Indirect effects, however, cannot be measured this way for most insurance lines. A high level of fraud diversification and randomness is hard to encapsulate into prediction models, especially those capable of forming estimates with an sufficient degree of confidence to be included in the insurance company’s savings estimates. However, the assessment of indirect savings is to a certain extent feasible in some insurance lines where the number of potential fraudsters is limited and relatively small. For such cases, it can be presumed that once fraudulent activity is detected and sanctioned, it will no longer occur or it will occur to a lesser extent. In fore mentioned systems assessment can be made on the basis of the differences in the dynamics of occurring suspicious fraud cases. An example of such insurance line is health insurance. The result of the master’s thesis is an extended assessment method which can be used to evaluate the performance of health insurance fraud management systems. The method takes into consideration both the direct savings and the assessment of indirect effects which occur as a result in the systematic fight against fraud. With the help of the suggested method the health insurance companies gain not only a full overview of the success and efficiency of the fraud management processes but also an assessment of how these influence the business performance of an entire insurance company.