Healthcare Benefit Optimization and Fraud Detection using Adaptive Learning and Data Driven Modeling

Abstract

With the exponential rise in government and private health-supported schemes, the number of fraudulent billing cases is also increasing. Detection of fraudulent transactions in healthcare systems is an exigent task due to intricate relationships among dynamic elements, including doctors, patients and services. Hence in order to introduce transparency in health support programs, there is a need to develop intelligent fraud detection models for tracing the loopholes in existing procedures, so that the fraudulent medical billing cases can be accurately identified. There is also a need of methodology to optimize both the cost burden for the service provider and medical benefits for the client. The thesis presents an actor level fraud detection framework based on three entities (patient, doctor, service). The proposed framework initially computes association scores for the underlying healthcare ecosystem entities and filters out the identified (anomalous) cases using association scores. The filtered cases are forwarded for the evaluation by the Rule engine which is designed by applying G-means clustering and from the generated clusters, confidence values are computed for each service in each specialty. Rules are generated based on the confidence values of services for each specialty. Then, the framework classifies cases into fraudulent activities or normal activities based on the similarity bit’s value. The thesis also presents a novel process based fraud detection methodology to detect insurance claim-related frauds in the healthcare system using sequence mining and sequence prediction concepts. Recent literature focuses on the amount-based analysis or medication versus disease sequential analysis rather than detecting frauds using sequence generation of services within each specialty. The proposed methodology divides into modules: the ”Sequence rule engine and the Prediction-based engine.” The sequence rule engine generates frequent sequences and ii probabilities of rare sequences for each hospital’s specialty. Comparing these sequences with the actual patient ones, leads to identify anomalies as both sequences would not be compliant with the rule engine’s sequences. The system performs further in detailed analysis on all non-compliant sequences using the prediction-based engine. The validation of the actor level fraud detection framework and process based fraud detection methodology is performed from the last five years of a local hospital’s transactional data which includes many reported cases of fraudulent activities. Also, in order to make the validation more robust, we introduced data anomalies of our own. The results demonstrate that the methodology is pertinent to detect healthcare frauds. Apart from this, the thesis also proposes the design of need-based insurance packages in healthcare. By designing these packages, medical benefit optimization, which is the core goal of our proposed methodology, is effectively achieved. Therefore, the optimization gives relief to the enterprises as well as employees. Also, the conventional computing mechanisms for insurance packages and premium methods are time-consuming, designation-specific, and not cost-effective. Our proposed methodology derives insurance packages that are need-based and optimal according to our defined criteria. We achieved this by first applying the k means clustering technique to historical employee’s medical records. Subsequently, medical benefit optimization is achieved from these packages by using a probability distribution model on five years employees’ insurance data. The proposed design provides 25% optimization on medical benefits amount compared to current medical benefits amount, therefore, gives better healthcare services to all the employees. The proposed solutions improve the delivery of healthcare services by identifying anomalies and analyzing the healthcare need of the people. The designed knowledge base evaluates each transaction for the identification of actor-level anomalies or frauds. Secondly, the solution resolves the misutilization of healthcare services from irrelevant specialties. Moreover, another outcome is the generation of need-based insurance packages to optimize the premium amount and the sum insured amount within the existing insurance policies. These solutions can inculcate the transperancy in Sehat card scheme recently introduced by government of Pakistan.