Journal of Rare Cardiovascular Diseases
ISSN: 2299-3711 (Print)
e-ISSN: 2300-5505 (Online)
Menu
AI-Driven Anomaly Detection in Healthcare Claims Data: A Business Intelligence Perspective
Bindu Madhavi Mangalampalli
Show More
1
Sr. BI Developer ORCID ID: 0009-0001-1070-3856.
Received: 2023-11-08
Revised: 2023-11-23
Accepted: 2023-12-13
Published: 2023-12-28
Full Text
Abstract
Background: Artificial intelligence (AI) is the biggest technological innovation and disruption for today’s organizations and society. Organizations are investing increasingly more time, budget, and focus on integrating AI tools into their business processes to support the changing customer demands and needs. In the Business Intelligence (BI) tandem, these tools can assist organizations in detecting data anomalies in real time, presenting the results for analysis and decision-making, and issuing alerts dynamically. In the area of Healthcare organizations, claims data requires careful oversight to identify and reduce fraudulent attempts. These BI-oriented integrations address the issue by implementing an Anomaly Detection engine in the BI solution for claims data. The underlying open-source code has broad generalizability, and two proof-of-concept case studies are presented. The first case study demonstrates how the integration successfully detects several fraudulent attempts to alter claims data entered into the system. The second case study highlights the detection of erroneous duplication of benefits for the same patient, presented to one of the main stakeholders as a pre-gated intelligence report. Anomaly detection aims to identify instances that differ significantly from the majority of other observations. The common process includes an anomaly detection approach, monitoring the system, an alert-generating engine, and a dashboard that embeds the alerts. The Anomaly Detection engine monitors a feature dataset for values that are numerically clustered together and flags data points significantly different from the neighbors in a negative way. The integration with the BI solution enables visualization of the flagged data points, and events can be pushed to any external system through a Simple Queue Service (SQS) notification. Dashboards can be designed from BI for the end-user business area to monitor any flagged data points, enabling timely preventive action.
Keywords
AI-Driven Business Intelligence (BI), Healthcare Claims Fraud Detection, Anomaly Detection Engines, Real-Time Data Monitoring Systems, Intelligent Alert Generation, Claims Data Analytics, Open-Source AI Integration, Fraudulent Transaction Identification, Duplicate Benefit Detection, Predictive Healthcare Risk Analytics, BI Dashboard Visualization, Feature Clustering Algorithms, Data Outlier Detection Techniques, Automated Fraud Prevention Systems, Queue-Based Notification Services (SQS), Data Integrity Monitoring, Intelligent Decision Support in Healthcare, AI-Enabled Compliance Oversight, Proof-of-Concept Fraud Case Studies, Preventive Analytics in Health Insurance Systems.
Journal Help
Indexation
User
Keywords
Classification of Rare Cardiovascular Diseases
anticoagulation
atrial fibrillation
atrial septal defect
cardiomyopathy
computed tomography
congenital heart disease
echocardiography
electrocardiogram
electrocardiography
heart failure
implantable cardioverter‑defibrillator
magnetic resonance imaging
pregnancy
pulmonary arterial hypertension
pulmonary hypertension
rare cardiovascular disease
rare disease
right heart catheterization
right ventricular failure
Instruction for authors
Submit an article