What actually gets a claim rejected
Six years reviewing claims from small practices. The rejections were rarely clinical — they were demographics, dates and modifiers.
I spent six years on the payer side, working through claims from practices much like yours. The thing that surprised me most was how few rejections had anything to do with clinical judgement.
Practitioners tend to worry about whether a treatment was justified. In the queue I worked, that was a small minority of denials. The overwhelming majority were failures of data entry, and almost all of them were preventable at the point of booking.
Here is the actual distribution, roughly, from what crossed my desk.
1. Demographic mismatch — the largest single bucket
The name on the claim does not match the name on the policy. A middle initial that is present in one system and absent in the other. A married name updated with you but not with the payer. A date of birth transposed.
None of this is interesting and all of it stops a claim dead. The fix is unglamorous: verify eligibility electronically before the first session, not after the third, and store the client’s name exactly as the payer holds it — not as they introduced themselves.
2. Stale eligibility
Coverage that was valid in January and lapsed in March. Employer plans change at renewal; individual plans lapse quietly. A practice that verifies once at intake and never again will accumulate a slow trickle of denials for clients it believes are covered.
Re-verify at a fixed interval. Monthly is more than enough for most caseloads, and it turns a denial six weeks after the fact into a conversation before the session.
3. Missing or wrong modifiers
This is where genuine clinical work gets rejected for clerical reasons. Two services in one session without the modifier that says they were distinct. A telehealth session submitted without the place-of-service code that makes it a telehealth session.
Modifiers are the part of coding that changes most often and gets communicated worst. If your software does not warn you at submission, you will find out six weeks later.
4. Timely filing
Every payer has a window — commonly 90 or 180 days, sometimes tighter. A claim that sits in a “to submit” pile past that window is not a denial you can appeal. It is simply gone.
The practices that lose money here are almost never disorganised in general. They are practices where claim submission depends on one person having a quiet afternoon, and quiet afternoons stopped happening in March.
5. Authorisation
Some services need prior authorisation and the requirement is inconsistent across payers for the same code. This one genuinely is hard, and it is the one place where a phone call still beats any software.
What this adds up to
Look at the list again. Four of the five are data problems that exist before the session happens. That is the useful conclusion: most of your claim outcomes are determined at booking, not at submission.
Which means the highest-leverage change is usually not better coding knowledge. It is:
- Electronic eligibility verification at intake, and on a schedule after that.
- Storing the payer’s version of the client’s identity, not the friendly one.
- Scrubbing against payer rules before submission, so the error surfaces in seconds rather than weeks.
- Submitting on a fixed cadence that does not depend on anyone’s afternoon.
On first-pass rate
If you track one number, track first-pass acceptance — the share of claims paid without any intervention. It is more honest than a collection rate, which can look healthy while hiding an enormous amount of unpaid administrative work behind it.
A practice at 70% first-pass and a practice at 92% may eventually collect similar amounts. One of them spends several hours a week doing it, and that time comes out of clinical hours or out of evenings.