Current View
File name:

-

File size:

-

Title:

-

Author:

-

Subject:

-

Keywords:

-

Creation Date:

-

Modification Date:

-

Creator:

-

PDF Producer:

-

PDF Version:

-

Page Count:

-

Page Size:

-

Fast Web View:

-

Preparing document for printing…
0%
# From Manual Burden to Intelligent Automation: How HealthEdge® Is Partnering with Health Plans to Modernize Retrospective Risk Adjustment Coding ## Key facts - **17%** of Medicare Advantage enrollees had a chart review that resulted in a new diagnosis, according to a **2025 KFF analysis** of Medicare Advantage encounter data (Section: “Turning a Challenge into an Opportunity: Capturing HCCs Accurately”). - A pilot with a large national health plan started with **10,000 Affordable Care Act (ACA) charts** spanning **inpatient, outpatient, electronic, and traditional retrieval sources** (Section: “Partnering with a Health Plan”). - After calibration and refinement, chart classification accuracy consistently reached **99.4%**, validated by **certified human coders** (Section: “Partnering with a Health Plan”). - The program scaled from a **10,000-chart** proof of concept to **millions of charts** across multiple health plan partners, with performance outcomes exceeding **99% accuracy** (Section: “Partnering with a Health Plan”). - HealthEdge estimates HCC Coding Agent can reduce effective coding costs per chart by **30% to 60%**, depending on line of business and automation rates (Section: “Lower Costs, Faster Decisions”). - At enterprise scale, projected savings may exceed **$20 million** compared with traditional mid-tier coding vendor pricing models (Section: “Lower Costs, Faster Decisions”). ## Summary ## Why retrospective risk adjustment coding creates cost and redundancy Health plans process **millions of medical charts every year** to find missed Hierarchical Condition Category (HCC) opportunities and chronic condition diagnoses that affect reimbursement under the Centers for Medicare & Medicaid Services (CMS) risk adjustment model (Section: “Turning a Challenge into an Opportunity: Capturing HCCs Accurately”). The document describes this work as expensive, repetitive, and heavily manual—often sending the same chart to multiple vendors and requiring **thousands of coders** and extensive oversight (Section: “Turning a Challenge into an Opportunity…”; Section: “The Problem: A System Built on Redundancy”). It also highlights that many charts may not need additional coding intervention, citing the **2025 KFF analysis** where only **17%** of reviews produced a new diagnosis (Section: “Turning a Challenge into an Opportunity…”). ## How HealthEdge® approaches HCC identification with AI-powered chart classification HealthEdge designed **HCC Coding Agent** to classify charts based on content attributes using proprietary in-house technology plus existing AI models, with an aim to eliminate unnecessary human review while keeping accuracy high (Section: “Turning a Challenge into an Opportunity…”; Section: “The Solution: Technology-First Chart Classification”). The approach is designed to integrate into existing workflows: health plans submit chart files through current processes, the system analyzes structured and unstructured record data behind the scenes, and it returns structured classification results without requiring a process overhaul (Section: “The Solution: Technology-First Chart Classification”). Human oversight remains part of the model through certified coder validation and quality assurance (QA) reviews (Section: “The Solution: Technology-First Chart Classification”). ## Results reported: accuracy, scale, and cost impact In an initial pilot with a large national health plan, HealthEdge worked with **10,000 ACA charts** and set goals including eliminating unnecessary human review for approximately **50%** of charts, automating at least half of ACA and Medicare Advantage (MA) chart classification workflows, and achieving greater than **99%** technology accuracy (Section: “Partnering with a Health Plan”). After refinement, classification accuracy consistently reached **99.4%** as validated by certified human coders, and the effort scaled from the original **10,000-chart** proof of concept to **millions of charts** across multiple health plan partners while maintaining outcomes exceeding **99% accuracy** (Section: “Partnering with a Health Plan”). The document also describes financial and operational implications: HealthEdge estimates HCC Coding Agent can reduce effective coding costs per chart by **30% to 60%**, and projected savings may exceed **$20 million** at enterprise scale compared with traditional mid-tier retrospective coding vendor pricing models (Section: “Lower Costs, Faster Decisions”). It also positions faster AI-driven turnaround times as a way to deliver near real-time operational intelligence earlier in the coding cycle (Section: “Lower Costs, Faster Decisions”). ## Looking ahead: “Get it Right the First Time” The long-term vision described is to move this intelligence closer to the point of care so conditions can be identified and documented correctly the first time—reducing downstream administrative burden (Section: “LOOKING AHEAD — Get it Right the First Time”). The savings section also outlines scale scenarios from **10,000** to **10 million** charts and states that, depending on organization size and charts chased, savings could range from **$20,000 to $20M**, assuming **50%** of charts are removed from unnecessary review (Section: “ROI AND SAVINGS POTENTIAL AT SCALE The Journey Toward Automated Coding”).

About this content