MCRAFMedicaid Claims Risk Analytics FrameworkDemo — Synthetic Data
MCRAF Documentation

Framework Methodology

Reference documentation for the Medicaid Claims Integrity and Prepayment Risk Analytics Framework. Last revised May 2026.

Section 01

Framework Overview

A state-driven analytical layer that sits between Medicaid claim submission and payment, scoring every claim against a configurable rule library and producing risk classifications that enable prepayment prevention.

MCRAF is deployed by state Medicaid agencies as a complement to existing fiscal-agent and SUR (Surveillance and Utilization Review) systems. It is not a replacement for federal program integrity partners, and it does not adjudicate claims on its own — it produces risk signals that route claims into Pay, Review, or Hold dispositions for human analysts.

Section 02

Architecture

The framework is composed of three loosely coupled components that communicate through versioned, schema-validated message contracts.

  • Data Adapter — ingests claims, eligibility, and provider files from the state MMIS and normalizes them into the framework's canonical schema.
  • Detection Rule Library — a versioned catalog of domain-specific rules that produce per-claim and per-provider risk components.
  • Prepayment Review Interface — the analyst-facing workbench (this application) for triage, evidence review, and disposition.
State MMISclaims · eligibilityData Adaptercanonical schemaDetection Rule Libraryscoring enginePrepayment Reviewanalyst workbenchPAYREVIEWHOLD
Section 03

Detection Rule Library

The active rule catalog as deployed for this demonstration. In production, rules are versioned per state and tunable via the Configuration console.

No detection rules registered.
Section 04

State Deployment Guide

A typical state deployment unfolds in three sequential phases. The total elapsed time is 8–18 weeks from kickoff to pilot go-live, depending on data-system maturity.

  1. 1

    Data Adapter Configuration

    2–4 weeks

    Map the state's MMIS extracts (claims, eligibility, provider master) to the framework's canonical schema. Stand up automated ingest, schema validation, and reconciliation reports.

  2. 2

    Rule Library Calibration

    2–6 weeks

    Tune default thresholds against the state's historical claim distribution. Suppress rules that conflict with state policy. Establish baseline false-positive rate targets.

  3. 3

    Integration and Pilot

    4–8 weeks

    Deploy the prepayment review interface to a designated analyst team. Run in shadow mode against a sample of claims; transition to live prepayment holds after sign-off.

Total Timeline
8–18 weeks
Section 05

Data Governance

This demonstration runs entirely on synthetic data generated using CMS-acknowledged methods. No protected health information, personally identifiable information, or real Medicaid claims data is used, stored, or processed.

The production framework is designed for HIPAA compliance, 42 CFR Part 455 compliance, and CMS MARS-E 2.0 security standards.

Section 06

Research Foundation

This framework builds on peer-reviewed research including [CITATION TO BE INSERTED] on using predictive analytics to model policy and medication outcomes in Medicaid populations.