The cost of claim denials and rework: what the studies say, and how to estimate yours

The short answer

The best-traced figures put the cost of working one denied claim at about USD 43.84 (Premier’s 2022 figure, all payers, as reported by HIA)¹ and the cost of an appealed claim at about USD 118 (a 2016 Change Healthcare study of 724 hospitals).² On frequency, Kodiak Solutions found an initial denial rate of 11.8% in 2024 data³, and a Premier survey found nearly 15% of claims to private payers initially denied.⁴ These are averages from different years and methods. Use them to frame the question, then measure your own.

The figures we could trace

We left out a widely repeated “rework cost per claim for medical practices” figure because we couldn’t trace it to a primary source we were comfortable citing. If a number in a vendor deck has no study behind it, leave it out too.

What each figure measures

The numbers aren’t interchangeable.

• USD 43.84 vs USD 118. The first is an average across denials; the second is for appealed claims, which take more work. They’re also six years apart. Don’t average them.
• 11.8% vs 15%. Kodiak’s figure comes from its own client data across payers; Premier’s is a survey about private payers. Different populations, different methods.
• Initial vs final. Initial denial rates include claims later paid after rework. The cost is the rework, not necessarily lost revenue.
• US only. All four figures are from the US. In India the pressure point is time: IRDAI asks insurers to decide cashless requests within one hour and give final discharge authorisation within three.⁶

A model for your own cost

Your own data beats any average. Here’s a simple model, with placeholder inputs:

With the example inputs: 20,000 × 12% = 2,400 denials a month; 70% worked = 1,680 reworked claims. Multiply by your minutes and cost per minute, then add the appeal share at its higher cost. If you have no time study yet, the published USD 43.84 and USD 118 give a rough range for the first estimate¹², but replace them as soon as you can.

Add one line people forget: denials written off without being worked. That’s lost revenue, not rework cost, and it can be larger.

A time study doesn’t need software. Pick 50 denials across your main payers and reasons, and have the staff who work them note start and end times, calls made and resubmissions. Repeat each quarter. The averages will move as payer rules change, and the spread between denial reasons usually matters more than the overall mean: a few reasons typically take most of the time.

Where denials come from

Your remittance codes tell you which denials are worth preventing. Group them into a few buckets:

1. Eligibility and registration: wrong member id, inactive coverage.
2. Prior authorization: missing or not matching the service delivered.
3. Coding and documentation: codes not supported by the record.
4. Medical necessity: payer disagrees the service was needed.
5. Timely filing and duplicates: process errors.

Buckets 1, 2 and 5 are mostly process and data problems, and the cheapest to fix. In the US, CMS-0057-F’s prior authorization deadlines and its requirement for specific denial reasons from 2026 should make bucket 2 easier to act on, because a specific reason says what to fix.⁵

Where automation helps, and where it adds rework

Automation can reduce denials in buckets 1, 2 and 5 by checking eligibility and authorisation before submission, and catching duplicates. It can also create rework if it writes wrong data at speed. Automated coding from general-purpose models has been shown to be inconsistent; our coding accuracy post summarises the studies.

Two rules keep automation on the right side:

• Measure denials by cause before and after any automation goes live, not just total volume.
• Keep a record of what the automation submitted and what the payer received, so a denial can be traced to either the automation or the payer’s rules. A claim that says one thing in your system and another at the payer is rework waiting to happen.

Questions people ask

What does it cost to rework a denied claim?

Published figures range from about USD 43.84 per denial (Premier, 2022)¹ to about USD 118 per appealed claim (Change Healthcare, 2016).² Measure your own.

What’s the average claim denial rate?

Kodiak found 11.8% initially denied in 2024 data³; Premier found nearly 15% for private payers.⁴

Is it worth appealing?

Often, since initial denials include claims later paid. Your appeal success rate by denial reason tells you which appeals are worth the cost.

For payers and their vendors, our insurer page shows how automated claim actions can be recorded, and the provenance spec describes the record format. To try approvals and read-backs on a real database in your browser, open the playground; to grade what an agent can reach on your own machine, run npx deedglass scan (the grade, explained).

Written with AI assistance and checked against our claims register.