Health Insurance Fraud Detection and Characterization

在加拿大，医疗保健成本为每年2150亿美元。欺诈和滥用占这个**金额的2%到10%，相当于每年大约20亿美元的损失。欺诈的成本增加了处方药和疾病的成本。尽管欺诈性索赔的成本巨大，但大多数保险公司实施的欺诈检测流程**在很大程度上是一个手动过程，领域专家可能会标记**超出一定阈值的索赔以进行进一步审查。这些审查既昂贵又低效。领域专家可能花了很多时间，但通常只能识别一小部分欺诈性声明。随着良性和欺诈性索赔数量的持续增长，迫切需要创新的**解决方案，以解决健康保险行业中这一重要但具有挑战性的问题。**Segic是一家总部位于加拿大的公司，旨在提供创新的软件平台，以满足**集体保险行业的需求。这项研究项目是Segic利用人工智能(AI)来增强其提供的软件服务能力的战略计划的一部分。具体地说，这个为期6个月的**项目旨在开发一种有效的机器学习方法来识别和表征欺诈性索赔。**Segic将贡献他们在识别和表征欺诈性案例方面的专业知识。麦吉尔团队将**贡献他们在机器学习和网络取证领域的知识。这两个团队的联合将产生一个有效的医疗保险欺诈检测引擎原型。研究结果将**提高医疗保险行业欺诈检测的效率和有效性，进而**有助于降低加拿大的医疗成本。

In Canada, the cost of healthcare is $215 billion annually. Fraud and abuse account for 2% to 10% of this**amount, translating into losses of approximately $2 billion per year. The cost of fraud increases the costs of**prescription drugs and illnesses. Despite the huge cost of fraudulent claims, the fraud detection process**implemented by most insurance companies is very much a manual process, where domain experts might flag**claims that go beyond a certain threshold for further review. These reviews are costly and inefficient. Domain**experts may have spent a lot of time, but often can identify only a small fraction of fraudulent claims. As the**volume of both benign and fraudulent claims continues to grow, there is a pressing need for innovative**solutions that can tackle this important yet challenging problem in the health insurance industry.**Segic is a Canadian-based company that aims at providing innovative software platforms to meet their needs of**the collective insurance industry. This research project is part of Segic's strategic plan to utilize artificial**intelligence (AI) to enhance the capabilities of their offered software services. Specifically, this 6-month**project aims at developing an effective machine learning method to identify and characterize fraudulent claims.**Segic will contribute their expertise of identifying and characterizing fraudulent cases. The McGill team will**contribute their knowledge in the areas of machine learning and cyber forensics. The joint force of the two**teams will result in an effective health insurance fraud detection engine prototype. The research result will**enhance the efficiency and effectiveness of the fraud detection in the health insurance industry, which in turn**contributes to the reduction of healthcare cost in Canada.