Denial Management for Physician Groups

You Have 1,300 Denials. Your Team Can Work 108. The Math Isn't Working.

Over 60% of winnable denials are never resubmitted. Learn how to close the execution gap and stop the revenue leak.

ANKA — AR Execution Dashboard
1,300
Denials / Mo
108
FTE Capacity
65%
Overturn Rate
Monthly Denial Recovery

If you're a revenue cycle management (RCM) director for a specialty practice or a mid-market physician group, you probably start your Monday morning the same way: looking at a denial worklist that grew while you were sleeping.

You have a team of talented, hard-working billers. You might even have a denial management software tool that flags these claims with bright red icons. But despite the hustle, the "Days in AR" metric isn't moving and the "write-offs" are piling up.

Here is the uncomfortable truth: Your team isn't failing. The math is.


The Math of a 25-Provider Group

In the mid-market healthcare RCM, we have reached a breaking point where the volume of administrative friction from payers has officially outpaced human capacity. For a typical 25-provider physician group, the reality looks like this:

1,300
Appealable denials/month
(at a conservative 10% rate on ~13,000 claims)
35–60%
Of denied claims are never resubmitted
$25–$181
The cost to manually rework a single claim
33%
The average annual turnover rate for billing staff
50–65%
Overturn rates for appealed claims

The problem isn't just denial volume; it's execution capacity.

Let's look at the math that is quietly killing your margin.

Stat Graphic 1: The Human Capacity Wall

The Execution Gap: Why over 60% of your winnable denials never get touched.

1,300+
Appealable Denials / Month
108
Appeals / Month (1 FTE)

A side-by-side comparison showing "The Reality" vs. "Human Capacity" — typical for a 25-provider group.


The "108 Appeals" Rule: Why Your Team Can't Catch Up

Why is the "Right Bar" in the chart above so low? Because a high-quality appeal is a complex, multi-step process. A manual appeal requires:

  1. Logging into fragmented payer portals.
  2. Interpreting cryptic EOB (explanation of benefits) and reason codes.
  3. Gathering clinical documentation and medical records from the EHR.
  4. Drafting a specific, persuasive appeal letter.
  5. Submitting and—most importantly—tracking the follow-up.

Assuming approximately 140 productive hours per month per FTE, and an average of 20 to 30 minutes per appeal (though complex cases can take much longer), one dedicated FTE can realistically complete 108 high-quality appeals per month.

The Mathematical Deficit: If your group generates 1,300 denials, you would need to hire 12 additional billers just to handle rework. In a market where RCM talent is vanishing and turnover is at an all-time high, that isn't just expensive—it's nearly impossible.


The Silent Revenue Leak: The 60% Abandonment Rate

When the math doesn't work, teams are forced to "triage." They work the high-dollar claims and ignore the rest.

According to the Journal of AHIMA, as many as 60% of denied claims are never resubmitted. This "silent leakage" accounts for a 3% to 5% hit to your net revenue every year. For a practice collecting $10M annually, that represents $500,000 in earned revenue left on the table simply because nobody had the time to do the work.

The Hidden Cost of Rework

It's not just the lost revenue; it's also the cost of the chase. The cost to rework a single claim can go up to $181 for providers. When you factor in the rising cost of RCM labor, the math gets even bleaker. If your team spends two hours ($60 in labor/overhead) to recover a $150 denial, your actual margin on that service has vanished. You are "winning" the claim but losing the business. This is why "working harder" is no longer a viable RCM strategy; it's a recipe for margin compression.


Stop Buying Dashboards. Start Buying Execution.

The healthcare tech market has spent the last decade selling "insights." You likely already have a dashboard that shows you a beautiful pie chart of your denial reasons.

But dashboards don't submit appeals. In fact, most "AI-powered" RCM tools are what we call Level 2 AI: they identify the problem and then add it to your team's to-do list. This is what we call "dashboard fatigue." Giving a drowning team a more accurate report of how deep the water is doesn't help them swim.


The ANKA Difference: Level 3 Execution

ANKA was built specifically for mid-market provider groups that don't have the enterprise budget of a massive health system but faces the same enterprise complexity. ANKA represents Level 3 Execution:

ANKA doesn't just flag a denial. It reads the EOB, pulls the medical record, writes the appeal, submits it to the payer, and tracks it until the check arrives.

All appeals are documented, auditable, and configurable to align with your internal compliance policies. Exception-based cases are routed for human review when clinical judgment or payer negotiation is required.

This is execution—not recommendation.

Stat Graphic 2: The Level of AI Execution

From Dashboards to Autonomous Action

Level 1
The Dashboard
Tells you the problem.
Level 2
The Recommendation
Tells you what to do.
Level 3
ANKA — The Execution Layer
Does the work — submitting, tracking, and recovering.

The Staffing Trap: Turnover and Knowledge Wipeout

One of the greatest risks to a physician group is what we call "knowledge wipeout." In many groups, one or two senior billers hold the "tribal knowledge" of how to get a specific payer to pay a specific code. When that person leaves, your revenue cycle grinds to a halt. By shifting to an "execution layer," you institutionalize that knowledge.

ANKA's AI doesn't take vacation, it doesn't get headhunted, and it doesn't forget how to appeal a "medical necessity" denial for a radiology claim or a "global period" error in orthopedics.

Whether you are in anesthesia (dealing with concurrency errors), radiology (fighting prior-auth mismatches), or orthopedics (managing global period bundles), the problem is the same: the administrative burden is designed to make you give up. The payers are using AI to deny you at scale. If you are still using a manual spreadsheet and a highlighter to fight back, you are bringing a knife to a drone fight.


The Bottom Line: Scalability Over Labor

As your provider count grows, denial volume increases proportionally. If your model depends solely on labor, your cost-to-collect rises linearly.

Infrastructure-based execution enables denial processing to scale without proportional staffing increases. The math doesn't have to be broken. You just need a system that does the work.