Case Study * RAF / HCC Coding

$10M+ in risk adjustment improvements. 90%+ diagnosis accuracy. 7 states.

A value-based care organization serving Medicare Advantage patients needed to improve risk adjustment accuracy. Missed HCC codes were leaving millions on the table. We built an AI-powered platform that transformed their RAF scores while reducing audit risk.

$10M+

RAF Improvements

90%+

Diagnosis Accuracy

20K+

Patients Served

245

Providers

RAF/HCC Coding Platform - Average RAF Score Dashboard with facility filters and CMS scoring algorithms
The Challenge

Millions in missed risk adjustment from incomplete coding.

In Medicare Advantage and value-based care models, accurate risk adjustment is everything. HCC codes determine payment rates. Missed diagnoses mean missed revenue-and inaccurate patient risk profiles that affect care planning.

This organization had a problem at scale: 20,000+ patients across 7 states, 245 providers, and no systematic way to ensure diagnoses documented in clinical notes were being captured in claims. The gap between documentation and coding was costing millions.

The Risk Adjustment Challenge

X

Documentation-to-coding gap

Diagnoses documented in notes but not making it to claims

X

Chronic condition recapture

HCCs from prior years not being recaptured annually

X

Audit risk

Coding without clear documentation linkage creates exposure

X

Scale complexity

20K+ patients, 7 states, 245 providers-impossible to review manually

The Solution

AI-powered risk adjustment with explainable recommendations.

We built a platform that reviews clinical documentation-EHR data, lab results, medications, problem lists-and surfaces HCC coding opportunities with direct links to supporting evidence. Every recommendation is audit-defensible.

Missed Opportunity - ICD code recapture flow showing prior year vs current year diagnosis comparison

Holistic Patient Views

Unified view of each patient's clinical data across all sources-EHR, labs, medications, claims history. AI identifies HCC gaps by comparing documented conditions against coded diagnoses.

* Multi-source data integration (EHR, claims, labs, medications)

* Prior year HCC recapture identification

* Suspected condition flagging

* Clinical evidence linking

How We Recapture - Disease detection and code recapture workflow from EHR to accurate RAF scoring

Provider Workflow Integration

Recommendations surface at the point of care, not in a separate system. Providers see HCC opportunities during the visit, with supporting documentation pre-populated.

* EHR integration (Gehrimed, PointClickCare, Acclivity)

* Pre-visit preparation reports

* In-visit prompts for missing HCCs

* One-click documentation support

RAF Score Analytics Dashboard showing facility comparisons, provider performance, and HCC gap analysis

Population Analytics

Enterprise visibility into RAF performance across facilities, providers, and patient populations. Identify systematic gaps, track improvement, and prioritize interventions.

* RAF score trending by facility

* Provider performance comparisons

* HCC gap analysis by category

* ROI tracking and forecasting

Audit-Ready Documentation

Every HCC recommendation linked directly to clinical evidence. When auditors ask "why did you code this?"-the answer is one click away. Built for defensibility.

* Source document linking

* Clinical evidence chains

* Audit trail logging

* Peer-reviewed CMS methodology

Technical Approach

Responsible AI with explainability at the core.

Healthcare AI isn't just about accuracy-it's about explainability. Every recommendation must be defensible. We built this platform with auditability as a first-class requirement.

Peer-Reviewed Methodology

AI models built on CMS HCC scoring methodology. Not black-box predictions-transparent, validated approaches.

Evidence Linking

Every recommendation tied to specific clinical documentation. Auditors can trace from code to evidence in seconds.

Multi-Source Integration

EHR data, lab results, medications, claims history-all combined for comprehensive patient view.

Continuous Learning

Platform improves with feedback. Provider corrections and audit results feed back into the model.

The Results

$10M+ in improvements across 7 states.

Production deployment serving real patients, real providers, real revenue. These numbers come from actual RAF score improvements tracked over time.

Financial Impact

$10M+

Total RAF improvements

Higher

ICD/HCC recapture rates

Reduced

Audit exposure and claw-back risk

Clinical Impact

90%+

Diagnosis accuracy

Better

Patient risk profiles for care planning

Reduced

Hospitalizations through better targeting

Scale

20K+

Patients served

7

States in production

245

Providers using the platform

Integration

Wrapped around existing EHR systems.

No rip-and-replace. The platform integrates with existing EHR systems, pulling data for analysis and surfacing recommendations within provider workflows. Providers don't need to log into another system.

Production Integrations:

Gehrimed

Primary EHR integration

PointClickCare

SNF data integration

Acclivity

Claims and billing data

Lab Systems

Results integration

Why Integration Matters for RAF

Y

Complete patient picture

Lab results, medications, and clinical notes all factor into HCC identification.

Y

Point-of-care delivery

Recommendations surface during the visit, not in a separate report reviewed later.

Y

No duplicate data entry

Providers accept or modify recommendations-data flows back automatically.

Y

Audit trail preservation

Source documentation links maintained through the workflow.

Related Work

More healthcare results.

What's your risk adjustment opportunity?

If missed HCCs are costing you millions in RAF revenue, let's talk. 30-minute call to assess your opportunity.

Schedule Assessment Call

Related Insights

From Our Blog

Healthcare Technology

Ready to Transform Your Healthcare Technology?

We’ll scope a working AI prototype for your RAF/HCC coding challenge in one week.

19 years in business · 200+ products launched · One team, concept to scale