Medicare Fraud Case Analytics

Cutting legal case preparation from 3 weeks to 2 days through direct database access and Power BI modernization

Medicare Fraud Case Analytics

Role: Lead Healthcare Data Analyst · Organization: Tennessee Attorney General’s Office · Timeline: February 2021 – April 2023

SQL Power BI TennCare Healthcare Data Legal Analytics Data Standardization

The Problem

The Tennessee Attorney General’s Office prosecutes Medicare fraud cases that involve large, complex healthcare datasets. When I joined the team, the analytical workflow had several critical bottlenecks:

3-week turnaround for dataset requests — analysts had to submit requests through intermediaries to access TennCare (Tennessee’s Medicaid program) data, with no direct query capability
Excel-based reporting for attorney-facing deliverables, which limited the depth of analysis and made it difficult to explore data interactively during case preparation
No standardized data organization — each analyst structured their outputs differently, creating inconsistencies across cases and making it difficult for attorneys to compare analyses
Time pressure — fraud cases have legal deadlines, and a 3-week data turnaround meant analysts were often working with stale information or rushing to compensate for delays

The Approach

I tackled this from three angles simultaneously: access, tooling, and process standardization.

Securing Direct Database Access

The single highest-impact change was obtaining direct access to the TennCare database. This required:

Building trust with the TennCare data team by demonstrating responsible data handling practices and understanding of HIPAA compliance requirements
Documenting the specific query patterns the AG’s office needed, showing that direct access would reduce burden on the TennCare team (fewer ad-hoc requests for them to fulfill)
Establishing protocols for secure data extraction and storage that met both agencies’ governance requirements

This alone collapsed the data turnaround from 3 weeks to 2–3 days — the time it takes to write, validate, and document a query rather than wait in a request queue.

Modernizing the Reporting Stack

I transitioned the team’s attorney-facing deliverables from Excel to Power BI, which provided:

Interactive exploration — attorneys could filter by provider, date range, procedure code, and billing pattern without requesting new reports
Visual pattern detection — fraud patterns that were invisible in spreadsheet rows became obvious in visualizations (billing spikes, geographic clustering, procedure code anomalies)
Compliance-ready formatting — standardized layouts that met courtroom presentation requirements

Standardizing Data Organization

I designed a consistent data organization framework for case analytics:

Standard naming conventions for datasets, queries, and deliverables
Template structures for common case types so new analyses started from a proven foundation rather than a blank page
Documentation requirements for every dataset — source, extraction date, query logic, and known limitations — so any analyst could pick up where another left off

flowchart LR
    A[Case Assignment] --> B[Direct TennCare<br>Database Query]
    B --> C[Data Extraction<br>& Validation]
    C --> D[Standardized<br>Dataset Structure]
    D --> E[Power BI<br>Analysis & Visuals]
    E --> F[Attorney-Facing<br>Deliverable]
    F --> G[Case Prosecution]

    style B fill:#3B82F6,color:#fff
    style E fill:#3B82F6,color:#fff

Working with Sensitive Data

Healthcare fraud analytics involves protected health information (PHI) and legally sensitive case data. Every aspect of this project was designed with compliance in mind:

All data access followed HIPAA protocols and agency security policies
Datasets were stored in secured, access-controlled environments
Power BI reports were designed for internal use only with appropriate access restrictions
Query logic and analytical methods were documented for potential courtroom testimony and cross-examination

Recruiter Note

Due to the sensitive nature of this work (active legal cases, protected health information), I cannot share specific datasets, dashboards, or case details. I’m happy to discuss the methodology, technical approach, and analytical patterns in a conversation setting.

Results & Impact

3 wks → 2 days

Dataset turnaround reduction

Direct Access

TennCare database query capability

Excel → Power BI

Attorney-facing reporting modernized

Standardized

Data organization across all cases

Additional Outcomes

Stronger cases — faster access to fresher data meant analyses were based on current information rather than snapshots that were weeks old
Attorney empowerment — interactive Power BI reports let attorneys explore the data themselves during case preparation, reducing back-and-forth with the analytics team
Team scalability — standardized processes and templates meant new analysts could become productive faster and produce consistent outputs
Cross-case pattern recognition — standardized data structures made it easier to identify patterns that spanned multiple cases or providers

Lessons Learned

1. Access is the first bottleneck to solve. The most sophisticated analytical tools in the world don’t matter if you’re waiting 3 weeks for data. Securing direct database access was a relationship and trust-building exercise as much as a technical one — and it delivered the single largest time savings.

2. Analysts aren’t the only users. Attorneys needed to interact with the data directly, not just receive static reports. Designing for their workflow (filtering by provider, exploring billing patterns, presenting in court) produced better outcomes than designing for analytical elegance.

3. Standardization pays compound interest. The time invested in templates, naming conventions, and documentation standards felt slow at first, but it paid back exponentially as the team handled more cases in parallel with consistent quality.

Technical Stack

Component	Technology	Purpose
Data Source	TennCare Database (direct access)	Medicare/Medicaid claims, provider data
Query Layer	SQL	Data extraction, validation, case-specific analysis
Visualization	Power BI	Interactive attorney-facing dashboards
Prior Tooling	Excel (replaced)	Legacy reporting format
Governance	HIPAA-compliant protocols	PHI handling, access controls
Process	Standardized templates	Consistent data organization across cases

← Back to All Projects ← Prev: Data Reconciliation