Mediclaim Fraud Detection and Management Using Predictive Analytics

Abstract

Medical insurance fraud detection has become very significant and a demanding task. Fraudulent activities occurring in insurance claims has resulted a huge loss to the Health insurance companies. The goal of this proposed system is to collect and analyze insurance data including insurance claims, hospital records, past insurance claim data, patient’s data and to provide a single platform for checking and providing a list of suspicious claims using predictive analytics. Predictive analytics helps to assess risk associated with certain conditions, and guides in decision making. The medical insurance dataset is viewed as a collection of data of hospitals, insurance company, doctors, patients, pharmacies, and other fields. These data can be huge with many doctors, hospitals and millions of patients details involved. In this work, a mediclaim fraud detection system is designed using predictive analytics, which detects fraudulent medical claims resulting from substantial monetary loss in healthcare systems. An analytical approach to detect medical claim insurance fraud is done using Logistic regression. The fraudulent claims are to be detected using the data from various sectors like insurance company, hospitals, pharmacy, and insurance policyholder. So a multi criteria decision support system is developed to predict if a claim is fraudulent or legitimate. The experimental results demonstrates that proposed model using logistic regression and multi criteria decision analysis performs better than the existing system.