Where dental claim denials actually come from — and 5 you can fix today

The average dental practice loses six to nine percent of submitted claim revenue to denials, downcoding, and re-submission delays. For a practice grossing $1.5M a year, that's $90K to $135K of work that got done, billed, and then ignored, partially paid, or rejected.

Most owners treat that number as a cost of doing business with insurance. They are wrong.

Most of what gets called "insurance trouble" is really five very specific problems. Each of them is well-understood. Four of them are fixable before the claim leaves your office. The fifth is mostly fixable if you know what you're looking at.

This post is the map. We'll go through the five buckets, the math for each, and what you can change today on whatever PMS you're running.

1. Eligibility and coverage gaps

This is the largest bucket. The patient's plan changed, you didn't catch it, the claim came back denied. Or: the patient has dual coverage, the coordination of benefits is set up wrong, the primary payer denies, the secondary defers to the primary, and your AR rep is now in a three-way phone call.

The underlying issue is timing. Most practices verify eligibility when the appointment is made — which might be three weeks before the visit — and don't re-check at the chair. In that window, plans change. People get laid off. Open enrollment shifts coverage. The Medicaid carrier reassigns provider networks. None of it shows up until the claim bounces.

The fix: Verify eligibility twice. Once at booking, once at check-in. The second check is the one that catches everything. If your PMS doesn't do real-time eligibility — or only does it on a clearinghouse poll once a day — your front desk has to do it manually, and they won't, because they have eight other things going on. Pick one or fix the other.

There is a related sub-problem: knowing your frequency limits against this specific patient's specific plan. We'll get to that one separately, but it shares a root cause: eligibility data that's a day old is worse than no eligibility data at all, because it lets the team think the work is done.

2. Coding errors

CDT code wrong. Tooth number wrong. Surface code missing or incorrect. Wrong provider attached. Wrong place of service. Missing modifier.

A coding error doesn't always trigger a hard denial. Sometimes it triggers a downcoding event — the payer reimburses at a lower rate for what they think you did, and you find out about it on the EOB three weeks later. That's the worst kind of denial because nobody on your team flags it. The claim looks paid; the practice just lost money on it.

The truth here is that coding is hard, dental coding is harder than medical coding because the procedure-to-code mapping changes by tooth and surface, and even experienced coders make mistakes when they're moving fast.

The fix: Two queues. Every claim should be in coded state before it hits submission state, and a human should look at the coded queue. The human is usually your office manager or a billing specialist. The review takes thirty seconds per claim once you're used to it. The catch rate is high — most teams flag two to four codeable errors a day. Multiply by $50-$200 per corrected claim and the math is obvious.

A modern PMS should help here by validating codes against tooth/surface combinations before the claim is even ready to submit. Many do not.

3. Missing attachments

The claim is technically valid but the payer requires supporting documentation that didn't go out with it. A bitewing. A perio chart. A narrative explaining medical necessity. A pre-op x-ray for a crown.

Each payer has its own list of what triggers an attachment requirement, and the list changes. When you submit without the required attachment, the claim either denies outright or sits in a "pending — provider information" state for two to six weeks while the payer waits for you to send the missing thing.

The fix: Build the attachment policy into the claim creation step, not the submission step. The moment a procedure code that requires an x-ray is entered for a patient, the system should refuse to mark the claim ready-to-submit unless an x-ray is attached. If your PMS doesn't enforce this, your billing team needs a manual checklist taped above their monitor with the top thirty procedures and their attachment requirements. A monitor checklist is not glamorous but it works.

This is the bucket where the gap between "good" PMS and "bad" PMS is widest. Good ones turn this from a process problem into a software problem. Bad ones make it a memory problem.

4. Frequency limits

Patient X had a prophy four months ago. The plan covers two per year, every six months. You schedule another prophy at five months because the patient was due, the front desk didn't catch the timing, the hygienist did the work, and the claim came back denied with the helpful explanation: "frequency exceeded."

This is a denial that costs you the entire procedure. The work is done. The chair was occupied. You can't bill the patient because they assumed insurance was covering it. You either eat it or have an awkward conversation.

The frustrating thing about frequency denials is that they are entirely predictable. Every plan publishes its frequency rules. Every patient's history is in your PMS. The question of "can I bill this prophy" has a deterministic answer at the moment of scheduling.

The fix: Frequency check at scheduling, frequency check at check-in. The check at scheduling prevents the wasted appointment. The check at check-in catches the cases where the plan changed in between. A PMS that does this automatically is the bare minimum here; if yours doesn't, your scheduling protocol needs an explicit "check last service date against plan frequency" step before any appointment is confirmed.

There is one nuance: when the patient knows they're past frequency and is paying out of pocket anyway, the system should let you proceed. It just shouldn't surprise anyone with a denial after the fact.

5. Payer-specific rules

This is the bucket that earns insurance its reputation. Every payer has its own little playbook of bundling rules, downcoding patterns, narrative requirements, and "we'll cover this only if you also did that" gotchas. Aetna handles one situation differently than Cigna. Delta Dental varies by state. Some payers downcode posterior composites to amalgam regardless of what the chart says. Some require a narrative for every endodontic procedure on a non-vital tooth. Some pay differently if the same procedure was on the same patient by a different provider within ninety days.

You cannot memorize the full playbook. Even billing specialists who have done this for fifteen years still get surprised by new payer-specific rules quarterly, because the payers add new ones.

The partial fix: Build a cheat sheet for the top five payers you submit to most. It will be specific to your geography and your provider mix. Anything beyond that is either institutional memory (your billing person knows) or software (your PMS has the rules baked in for the top payers and surfaces warnings at coding time).

This is the bucket where software earns its keep, because no human can hold this many edge cases in their head. If your PMS is from the era of paper EOBs, it doesn't have this. If your PMS is from this decade, ask whether it does.

What this adds up to

Of the five buckets, four can be fixed at submission time. Eligibility, coding, attachments, frequency: all of them are checkable before the claim leaves your office, with software you can run today and process changes you can implement next week.

The fifth bucket — payer-specific rules — is harder. It's where you'll always lose some claims to surprises you couldn't predict. The goal there isn't zero denials; it's predictable denials, so you know what to expect from each payer and can build the appeals workflow around it.

The math, if you do this work:

A practice grossing $1.5M annually with current denial rates of seven percent is losing roughly $105K per year to the five buckets above. A reasonable target after a six-month process tightening is to bring the denial rate to three to four percent. That's $45K to $60K recovered annually. The work is real but the math is good.

The deeper point, the one we keep coming back to in conversations with practice owners:

Most denials happen at the chair, not at the clearinghouse.

By the time the claim is at the clearinghouse, your software either caught the problem and is asking your billing person to fix it, or it didn't, and a week from now you'll find out about it on an EOB. The version of this where catching denials becomes a job is the version where you're already losing.

The version where your software prevents denials from being created is the version where the job mostly goes away.

If you'd like to know what we've built to prevent denials at the chair, the product page has the details. If you'd rather just see it run on real data, we do 20-minute demos on actual practice data.

The denial math doesn't have to be the denial math forever. The buckets are knowable. The fixes are doable. Most of the loss is preventable.

---

PracticeCore AI is a modern practice management platform built for dental practices. One price (1% of what you collect), no per-seat fees, self-host option available. We talk to dentists every day. We write here when there is something worth saying.