- About 9% of hospital claims are denied the first time, and reworking each one costs roughly $118 (national denials index, reported by Healthcare Dive).
- Most denials are preventable. They start before the claim exists: at scheduling, check-in and documentation.
- Prevention beats appeals. Fix the process that produced a denial, not just the claim.
- Measure your initial denial rate by payer and by reason every month, and work the top three causes.
A denied claim is money your practice has already earned and has not been paid. Every denial has to be found, understood, corrected and resubmitted, and some are never recovered. The good news is that most denials follow a small number of patterns, and each pattern has a fix.
Why denials matter more than they look
A national denials index looked at more than 3.3 billion transactions from 724 hospitals and found that about 9% of claims were initially denied, worth roughly $262 billion in a single year (2016 data). It also estimated that about 63% of denied claims were recoverable, but that reworking each denial cost around $118.
Two things follow from those numbers. First, a denial is never free even when you win the appeal: someone on your team spends time on it, and payment arrives weeks later. Second, more than a third of denied dollars are never recovered at all. For a practice billing $2 million a year, every percentage point of unrecovered denials is $20,000.
Denied vs rejected: know which one you have
Before fixing anything, separate the two:
- Rejected claims failed the clearinghouse or payer’s front-end checks: a missing field, an invalid member ID, a provider not on file. The payer never processed them. Fix the data and resubmit, usually the same day.
- Denied claims were processed and refused. The remittance (ERA) carries a reason code and remark that explains why. These need correction, a corrected claim, or an appeal.
Treating a denial like a rejection (resubmitting the same claim) usually produces a duplicate-claim denial. Treating a rejection like a denial wastes days.
The most common causes, and how to prevent each
That index found that registration and eligibility problems were the largest single cause (about 24% of denials) and missing or invalid claim data the next (about 15%). The rest spread across a familiar list.
1. Eligibility and registration errors
The patient’s coverage ended, the plan changed, the member ID or date of birth was mistyped, or the wrong payer is primary.
Prevent it: verify eligibility electronically before every visit, not just the first one. Coverage changes at the start of every year and whenever someone changes jobs. Capture the insurance card at check-in, compare it to what is on file, and update the master record once so every new and unsubmitted claim picks it up.
2. Missing or invalid claim data
Missing referring provider, wrong place of service, missing modifiers, a diagnosis that does not support the procedure, a missing NPI or taxonomy.
Prevent it: scrub every claim before submission. A good scrubber checks required fields for the claim type, diagnosis-to-procedure support, modifier logic and payer-specific rules, and it tells the biller exactly what to fix.
3. No prior authorization
The service needed approval and none was on file, or the authorization expired, or the visit count ran out.
Prevent it: flag services that commonly need authorization at the point of ordering, track authorization numbers, dates and remaining visits against the patient, and warn before a visit is scheduled outside an active authorization.
4. Medical necessity
The documentation does not support why the service was needed, or the diagnosis is not one the payer covers for that service.
Prevent it: document the reason for every test and procedure in the note, link diagnoses to the lines they support, and for Medicare, collect an Advance Beneficiary Notice when coverage is uncertain.
5. Coding and bundling
Procedures billed together that the payer considers part of one service, an unsupported modifier, or an outdated code.
Prevent it: keep your code library current every year, check pairs of codes against bundling edits before submission, and add modifiers only when the documentation supports a distinct service.
6. Duplicate claims
A claim resubmitted while the original was still processing, or two charges entered for the same service.
Prevent it: check claim status before resubmitting, use corrected-claim (replacement) submissions instead of new claims, and block duplicate charges for the same patient, date and service.
7. Timely filing
The claim arrived after the payer’s filing limit. For Original Medicare, the limit is 12 months from the date of service; commercial plans are often much shorter and are set in your contract.
Prevent it: report on charges not yet billed every week, and keep each payer’s filing limit visible in the work queue so the oldest claims are worked first.
Reactive denial work
- Find the denial weeks later on the ERA
- Research the reason
- Correct and resubmit or appeal
- Wait another payment cycle
Prevention at the source
- Verify coverage before the visit
- Track authorizations as care is ordered
- Scrub every claim before it leaves
- Fix the process behind each denial type
Working the denials you still get
Even with strong prevention, some denials arrive. A disciplined denial workflow recovers most of them.
- Post the ERA the day it arrives so denials surface immediately, not at month end.
- Route each denial to a work queue by reason with the payer’s reason and remark attached. Eligibility denials go to the front desk lead, coding denials to the coder, authorization denials to whoever owns prior auths.
- Decide: correct, appeal or write off. Fixable data errors become corrected claims. Disagreements about coverage or medical necessity become appeals with supporting documentation.
- Watch the appeal deadline. For Original Medicare, a redetermination must be requested within 120 days of receiving the initial determination. After that come reconsideration (180 days), an Administrative Law Judge hearing, the Medicare Appeals Council and federal court, each with its own deadline.
- Close the loop. Every denial type that repeats should change something upstream: a check-in step, a template, a scrubber rule.
The numbers to watch every month
You cannot manage denials you do not measure. Four numbers tell you most of what you need:
| Metric | How to calculate it | What it tells you |
|---|---|---|
| Initial denial rate | Claims denied on first submission ÷ claims submitted | How well your front end and scrubber work |
| Clean claim rate | Claims accepted with no edits or rejections ÷ claims submitted | Quality of claims leaving your office |
| Denials by reason and payer | Count and dollars per reason, per payer | Where to focus prevention |
| Days in A/R | Total A/R ÷ average daily charges | How fast you turn work into cash |
How software changes the math
Most of the prevention steps above are checks a person can forget on a busy day: verifying coverage, catching a missing modifier, noticing an expired authorization. Software does them every time.
In XEHR, eligibility is checked before the visit, insurance is captured once and carried to every claim, the claim is built from the signed note, and every claim is scrubbed before your biller approves it. Denials that still arrive land in a work queue with the payer’s reason and a suggested next step, and your team decides what happens. The goal is simple: fewer denials created, and the ones that do arrive worked while they are still easy to win.
