3 min read
Sep 17, 2025
Why RADV is Different - and How HealthMC Approaches It
Medicare Advantage risk scores are built on retrospective coding. Plans identify chronic conditions, coders confirm supporting documentation, and the results flow into risk scores that set revenue for future years. It’s a process that has grown familiar to payors, providers, and vendors alike.
But Risk Adjustment Data Validation (RADV) audits play by a different rulebook. They aren’t about maximizing revenue through coding. They’re about defending past submissions against CMS review. And that distinction changes everything.
The RADV Challenge
The stakes are High
CMS is increasing the scale of RADV audits, including extrapolation. A single chart with insufficient support can have the weight of thousands of members.
Financial risk is asymmetric: you can’t “win” extra revenue, but you can lose revenue you’ve already booked.
The Rules are Different
In retrospective coding, the standard is: Would a qualified coder at the plan code this diagnosis?
In RADV, the standard is: Would a CMS coder accept this diagnosis as supported? Only CMS’s determination matters.
The timelines are compressed
Retrospective projects often follow annual or semi-annual cycles.
RADV audits don’t wait. When CMS selects an audit sample, plans have a fixed, tight window to prepare and submit.
Why Traditional Coding Falls Short
Most coding teams are trained to think in binary terms: either a diagnosis is supported and codable, or it isn’t. That works for revenue capture. It does not work for defending against RADV.
In reality, diagnoses fall along a spectrum of support:
Clear support: Any coder would agree (diagnosis plus treatment plan documented).
Ambiguous support: Coders may disagree (complication listed, but not clearly linked).
Minimal support: Most coders would not code, but there’s a plausible argument.
No support: Nothing present.
For RADV, this variation matters. CMS allows multiple charts per condition. That means it’s not about finding the codable diagnosis. It’s about assembling the best possible portfolio of charts to maximize defensibility.
HealthMC’s Approach
We designed HealthMC’s platform with RADV in mind. Our strategy combines AI-driven breadth with actuarial and coding expertise to support nuanced decision-making:
Comprehensive chart review
Our AI reviews every available chart for every audited member, capturing all possible manifestations of conditions.
This goes beyond “yes/no” coding to identify weaker signals that might still justify submission.
Nuanced condition grading
We categorize potential diagnoses along the spectrum of support.
This allows plans to prioritize their strongest cases while keeping backup options for appeal and litigation.
Strategic chart selection
Because submission is constrained, we help plans decide which charts to send.
Sometimes the best move isn’t the clearest code, but the chart that opens the door to additional unaudited conditions that help offset risk.
Why This Matters Now
RADV is moving from an occasional concern to a defining challenge for Medicare Advantage. The combination of increased audits, extrapolation risk, and CMS scrutiny means that plans can’t afford to rely on legacy coding workflows.
HealthMC is uniquely positioned because:
We blend actuarial judgment with clinical coding. We understand not just what’s codable, but how decisions impact risk scores, revenue, and audit outcomes.
We use AI to scale nuance. Human coders struggle to consistently apply non-binary judgments. Our models can do this systematically across millions of charts.
We see audits as opportunity. By reframing audit prep as a strategic data-selection problem, we can help plans protect revenue while positioning themselves for stronger coding processes going forward.
Closing Thoughts
RADV isn’t just another coding project. It’s a different game with different rules, higher stakes, and shorter timelines. Traditional vendors aren’t built for it.
At HealthMC, we believe that AI, coupled with deep domain expertise, is the only way to meet the challenge head-on. We’re already supporting plans through active audits, and we’re building partnerships to scale this work.
If your organization is facing an audit — or expects to — we’d welcome a conversation.
