The Evolution of Use of Machine Intelligence for Fraud Detection by US Insurers
This block of instruction will track the actual history and development of data analytics through early use of computer technology, data-driven analytics, to supervised learning (edits), neural networks and finally the use of AI solutions.
Case studies from ten years’ experience with a large payer’s experimentation with vendors, data scientists and “home-grown” applications will be shared.

Emerging Trends in Healthcare Fraud: An Evolving International Landscape
Case studies from the most recent adjudicated criminal cases, with emphasis on transnational criminal organizations. Emphasis on the newest schemes, which involve fraudsters’ use of AI for manipulation of Electronic Health Records systems to drive inappropriate utilization of Medicare-covered products and services.

Beneficiary Fraud: Detection and Mitigation
Cases regarding beneficiary insurance “card sharing”, cloned medical records, the use of AI in medical imaging, false record creation, as well as misrepresented cosmetic treatments, falsification of diagnoses and kickbacks related to drugs, medical devices, durable medical equipment, and other products paid for by federal healthcare programs. Drug, device or biologics pricing, including arrangements for discounts, rebates, service fees, and formulary placement and price reporting. This final block will discuss the evolving and growing need for hospice care in the US, and the ensuing fraud and abuse by providers eroding this important benefit. Discussion of Home Health and Nursing Home abuse perpetuated by private equity firms. Use of data detection will be shared in case studies.
 

Emerging Trends & The Future of Healthcare Fraud Detection
Group exercise with the use of data utilization trending for identification of fraudulent “beta testing” in claims systems. Real-life data graphs will be presented for the participants to analyze and determine.
 

Hospice, Home Health Visits and Skilled Nursing Home Fraud and Abuse
This final block will discuss the evolving and growing need for hospice care in the US, and the ensuing fraud and abuse by providers eroding this important benefit. Discussion of Home Health and Nursing Home abuse perpetuated by private
equity firms. Use of data detection will be shared in case studies.
Quiz

The Evolving Fraud Landscape

• Briefly recap key insights from first two days on emerging fraud trends and the historical context of machine intelligence in fraud detection.
• Set the stage for a hands-on exploration of AI's practical application.

The Power of Data Visualization – From Dashboards to Generative AI for EDA

• What is Data Visualization?
• The journey from traditional static reports to interactive dashboards.
• Leveraging Generative AI for Exploratory Data Analysis (EDA) and Visualization:
• Hands-on Exercise: Generative AI for Visualizing Healthcare Claims Data
  • Task 1: Basic Claims Overview
  • Task 2: Spotting Suspicious Patterns (without coding): Guided prompts to help identify basic fraud schemes
  • Outcome: Experience how quickly you can generate insightful visualizations and spot "interesting patterns" without writing a single line of code.

Understanding Analytics: From Descriptive to Generative
The Spectrum of Analytics: Descriptive Analytics, Diagnostic Analytics, Predictive Analytics, Prescriptive Analytics, Generative Analytics

Basics of Predictive/Prescriptive Analytics (Machine Learning)

• Introduction to Machine Learning for Fraud Detection - Key ML Concepts for Beginners (classification, anomaly detection, clustering, graph analysis)
• Hands-on Exercise: Building a Simple ML Model (using a simple tool)
• Transition to Healthcare Fraud: Briefly discuss how these basic concepts scale to real-world healthcare fraud data.

Diving into Generative AI, LLMs, and Agentic AI

• Generative AI Demystified – What it is?
• Understanding Large Language Models (LLMs)
• Introduction to Agentic AI/AI Agents
• Demonstrating LLM Capabilities for Healthcare Fraud (using popular LLMs such as GPT, Gemini, Claude, Grok, DeepSeek)
  • Scenario 1: Medical Record Comprehension
  • Scenario 2: Claim Verification
  • Scenario 3: Suggesting Appropriate Diagnosis/Procedure Codes
  • Discussion: Emphasize the potential for investigators to quickly get context and initial verification, saving significant manual review time.

Applications of ML/AI/GenAI to Healthcare Fraud Detection (Deep Dive)

• Visualizing Fraud Schemes without Coding (Revisited and Expanded)
• Advanced Visualization Techniques: Show how to visualize complex fraud schemes using intuitive queries (e.g., Phantom Billing/Services Not Rendered, Upcoding, Unbundling, Patient Roaming/Doctor Shopping Discussion)

Reinforce the concepts of visualization, basic ML, and initial GenAI applications.

Generating Business Rules & Data Preparation for AI Models

• Generating Simple Business Rules using Generative AI – concept creation, tool selection, creating and running the rules
• Essential Steps for Applying AI Technologies:
  • Data Ingestion
  • Data Preparation (Data Wrangling/Preprocessing):
    • Cleaning data
    • Feature Engineering
    • Transforming Data from Claim Level to Provider/Patient Level

Building ML Models with Generative AI Assistance

• Building Machine Learning Models Using Generative AI Technology – demo showing concept creation, tool selection, building the model and running the model
• Model Interpretation and Explainability

Automating Fraud Detection with Agentic AI & Practical Schemes

• Automating Fraud Detection Activities with Agentic AI - demo showing concept creation and Illustrative Workflow
  • Walk Through Several Fraud Detection Schemes (e.g., Durable Medical Equipment (DME) Fraud, Opioid Diversion/Over-Prescribing, Kickbacks & Collusion)

Emerging Trends & The Future of Healthcare Fraud Detection

• The Shift to Pre-Pay Fraud Detection
• The Future Investigator – Enhanced by AI
• Emerging Trends in Healthcare AI
• Key Considerations - Compliance and Governance

• Roadmap for AI Adoption in Fraud Detection
• Open Group Discussion: Applying Learnings to Your Context
• Your AI Journey Continues: Beyond This Course

• Final questions and discussion
• Course evaluation and feedback