Kick-off
Intro, objectives and expectations

The Introduction to the problem of Medical Insurance Fraud

• Medical Insurance Fraud and its implications for society and economies
• Major Types of Fraud, waste, and abuse in healthcare insurance industry
• Most common fraud schemes

Basic Machine Learning Applications in Health Care Fraud Detection – EDA, Reporting & Dashboards

• The Evidence of Healthcare Fraud and How to Collect it / Discuss what data to analyze
• What kind of data is available for the health care fraud detection?
• What kind of EDA is useful in health care fraud detection?
• What is Data Analytics?
  • Descriptive, Diagnostic, Predictive and Prescriptive
  • Importance of Descriptive & Diagnostic Analytics
  • When we need Predictive & Prescriptive Analytics
• Fundamentals of Descriptive Analytics – Exploratory Data Analysis (EDA) / BI Reports / Visualization

Descriptive / Diagnostic Analytics in Action to combat Health Care Fraud

• Dashboards
  • How to do efficient EDA?
  • How to develop effective and outcome driven dashboards?
• What KPIs to report?
• How to convert your FWA expertise into business rules that could be run automatically every day / hour / week? – Case studies and Roundtable discussion
• Examples of visualization dashboards / tools to help fraud investigation
• Possible demo of some of these visualization tools / dashboards

Deep Dive on Machine Learning in Health Care Fraud Detection - Case Studies and Round Table Discussions

• Review of previous day learning sessions and data analytics discussed
• Fundamentals of Machine Learning - Predictive & Prescriptive Analytics
  • Major types of Machine Learning (e.g., classification, regression, association rules, clustering, text analytics, anomaly / outlier detection, etc) and relevant examples
  • Common use cases where Machine Learning is used but you may not know it
  • Common health care use cases where AI / machine learning is used today
• Effective feature construction for successful fraud detection using ML / AI
• How basic AI and machine learning could be used to fight health care FWA?
  • Business rules – quick review from yesterday
  • Classification models – detecting variations of know health care fraud
• Common challenges and risks using classification predictive models

Predictive / Prescriptive Analytics in Action to Combat Health Care Fraud

• How advanced unsupervised AI and machine learning could be used to fight health care FWA?
  • Feature generation for unsupervised learning
  • Clustering approaches for fraud detection
  • Anomaly / Outlier Detection – detecting novel types of health care fraud
  • Link Analysis – detecting fraud rings and nation-wide organized fraud groups
• Doing your homework – assets, marital and financial status, bankruptcies/divorces/substance abusedata
• Common challenges and risks using unsupervised ML models for fraud detection
• AI and machine learning for Post-pay vs. Pre-Pay Fraud Detection
• Reviewing actual use cases of successful investigations from AI and machine learning
• Round table discussion on possible fraud use cases and how AI and machine learning could help

Fraud Use Cases and Open Discussions on how to apply ML and AI to address those

• Open discussion of the course
• Each participant will come prepared with their own fraud use case and we will work together to analyze how to address it and how to apply machine learning and AI
• Quiz and discussion of responses
• Review of course and wrap up