Assurance
Pennsylvania Medicaid · Home Care Fraud Prevention
A pre-payment prevention and post-payment forensic audit system for home care billing, purpose-built for the Special Investigation Units of Pennsylvania Medicaid MCOs.
The Problem
In August 2026, DOJ's Eastern District of Pennsylvania and a national Medicaid Fraud Strike Force announced coordinated charges for over $5.76 million in fraudulent Pennsylvania Medicaid home care billings.
Named patterns included a caregiver who billed 64,000+ physically impossible hours, another with 8,700+ overlapping billed hours, billing during a client's incarceration or hospitalization, falsified EVV records, and kickbacks paid to clients to sign off on hours never worked.
Four of the five billing patterns named in that case are live in Assurance today — the fifth, client kickbacks, is a future consideration.
Source: DOJ Eastern District of PA, "Dozens Charged With Health Care Fraud… Involving $5.76 Million in Billings to Pennsylvania's Medicaid Program," Aug. 4, 2026 (justice.gov/usao-edpa).
Market-sizing estimate: PA DHS FY2024-25 Blue Book; CareRing Health PAS Rate Study for the PA Homecare Association. Conservative, directional estimate — see full sourcing notes available on request.
The Platform
Built and tested on a fully synthetic, fictitious 90-day dataset simulating a 9-agency provider network (123 caregivers, 88 clients, 6,941 shifts) — fraud spread realistically across several agencies.
Every module traces to one of three sources: DOJ's Eastern District of PA enforcement action (Aug. 4, 2026), CMS's "Vulnerabilities and Mitigation Strategies in Medicaid Personal Care Services," or a decade operating a PA home care agency from the inside.
Built to grow beyond Pennsylvania. The fraud patterns these modules catch aren't unique to Pennsylvania's billing system — they trace to federal DOJ enforcement patterns and CMS's own national Medicaid personal-care vulnerability research. The underlying detection logic adapts to any state Medicaid MCO's claims, EVV, and billing data.
Module-by-module detail — exact triggers, thresholds, and detection logic — is reserved for direct conversations with MCO SIU teams.
Request a walkthroughAgency-Level Oversight
Every module above catches something at the level of one caregiver or one claim. Assurance also rolls those same flags up to the agency level — surfacing when flagged activity clusters at a single provider rather than spreading evenly across a network. A pattern like that points to a different problem: not one bad caregiver, but an agency whose own oversight isn't catching it, or isn't trying to.
The goal isn't to police the industry — it's to remove the agencies whose conduct puts funding and reputation at risk for everyone else, so the honest majority of agencies aren't competing against, or paying the price for, the few that aren't.
In our own 9-agency simulated network, this rollup surfaced one agency running at over 2.5x the network's average flag concentration — while the quietest agencies sat at under a fifth of that average, well within normal range. That's the bar: find the real outlier, without flagging agencies that are actually clean.
Assurance flags. It doesn't decide. Every agency-level flag comes with the underlying incidents attached — never a verdict, a score with no explanation, or a recommendation to remove an agency from a network. That determination requires real investigation, due process, and judgment only an SIU can exercise. What Assurance provides is a documented, defensible reason to open that investigation in the first place.
Request a walkthroughData Handling & Compliance
Because we're pre-launch, our discovery and pilot phases run on a strict zero-PHI architecture — we don't accept or process live patient data. Pilots run on fully de-identified, historical extracts, so there's no HIPAA exposure to your organization at this stage.
We're designing our production environment for deployment onto HIPAA-compliant, BAA-backed cloud infrastructure, and will map our internal controls to SOC 2 Type I requirements as we move toward a signed pilot.
Specific data handling terms — retention, deletion, and processing commitments — are detailed in our Data Specification & Handling Sheet, available on request.
Let's Talk
If you're with a Pennsylvania Medicaid MCO's Special Investigations Unit, I'd like to set up a live demo — a walkthrough of how Assurance flags fraud before it's paid, with the option to run it against your own claims data.
Fill out our contact form → or email [email protected] directlyAssurance is currently in early conversations with Pennsylvania Medicaid MCOs and is not yet in production with any payer.
