Almost every United Healthcare denial rate graph in circulation traces back to one file: the Transparency in Coverage public use file that CMS publishes, which carries issuer and plan level claims and appeals data for health plans sold on HealthCare.gov. It is medical, it is individual marketplace, and it contains no dental claims. If you bill CDT codes for a living, that chart cannot tell you how your next crown will adjudicate. Here is what it does measure, why the versions you have seen disagree, and what to put on your own wall instead.
Where the chart comes from
The numbers behind the viral bar charts are not investigative journalism. They are a spreadsheet. Issuers selling qualified health plans on HealthCare.gov report claims received and claims denied to CMS, and CMS publishes the result as the Transparency in Coverage PUF alongside the other Exchange public use files. Anyone can download it. Most of the graphs are that file plotted, sorted descending, and screenshotted.
That origin explains four things at once.
It explains why the rankings shift year to year: the file is refreshed annually and issuer participation changes. It explains why the totals look large next to what you see in a dental office: the denominator counts every medical line submitted, including administrative rejections corrected and paid the following week. It explains why one chart says 31 percent and another says 33 for the same carrier: some analysts use in-network claims only, some use all claims, and some aggregate subsidiaries differently. And it explains the biggest gap, which is that the file was never built to describe dentistry.
Does UnitedHealthcare have the highest denial rate?
In several years of that file, UnitedHealthcare affiliated issuers have ranked at or near the top, and figures around 30 percent for 2023 are commonly quoted. Whether that makes the carrier the highest depends entirely on what you are willing to accept as a measurement.
Three problems sit inside the ranking.
The data is self-reported, with loose definitions. Issuers decide what counts as a denied claim. One may report every line that adjudicated to zero, including duplicates and corrected resubmissions. Another may report only claims denied after final adjudication. CMS publishes a data dictionary and flags the limitations; the charts almost never do.
The denominator is narrow. The file covers issuers on HealthCare.gov. Plans sold through state run exchanges, including California and New York, are not in it as of this writing. Employer group coverage, which is where most commercially insured Americans and nearly all dental patients sit, is not in it either.
Carrier size is confounded with carrier behavior. The largest payer processes the most claims, so it produces the largest raw count of denials. Raw counts and rates get mixed together constantly, and a headline built on counts says nothing about your odds on a single claim.
The honest answer: the carrier has topped one narrow, self-reported medical dataset. That is a real finding about that dataset, not a property of the brand and not a forecast for your operatory.
Which health insurance has the highest denial rate?
It depends on which file you open, which is why the versions you find online contradict each other. These are the three datasets people are usually arguing about without saying so.
| Dataset | What it counts | What it leaves out | Who reports it |
|---|---|---|---|
| CMS Transparency in Coverage PUF | In-network claim denials for HealthCare.gov marketplace plans, by issuer and plan | Dental, employer group plans, Medicare Advantage, Medicaid managed care, state run exchanges | Issuers, self-reported to CMS |
| Medicare Advantage prior authorization data | Prior authorization requests denied, and how many were overturned on appeal | Commercial coverage, original fee for service Medicare, dental benefits | Plans, reported to CMS |
| Your own electronic remittances | Every line your office billed, to every payer, with a reason code attached | Nothing you billed, which is the point | Your practice management system |
The first two get the charts. The third is the only one that predicts what happens to the claim you are about to send. Build a difficulty ranking from your own remittances and you get a list specific to your zip code, your employer groups and your procedure mix. It will not agree with the internet, and it will be right about your office.
The chart is medical, and dentistry does not work that way
This is the part the viral graphs cannot help with. Dental benefits are typically a separate policy with their own network, annual maximum and adjudication rules, even when the card carries the same brand as the medical plan, and the claim is adjudicated on CDT codes against a plan document an employer group selected from a menu.
That last point is the one to hold onto. Plan provisions are chosen by the purchaser, so two patients who both hand you the same carrier's card can have different frequency limits, different waiting periods, different missing tooth language and different downgrade rules. No public denial rate captures that, and no carrier level generalization survives contact with it.
The mechanics differ too.
| Medical claim on the chart | Dental claim in your office | |
|---|---|---|
| Code set | ICD-10-CM plus CPT or HCPCS | CDT, with tooth number, surface and quadrant |
| Common denial trigger | Medical necessity, prior authorization, network status | Frequency limit, annual maximum exhausted, missing attachment, waiting period |
| Typical remedy | Clinical appeal with records | Resubmit with the missing data element, often no appeal needed |
| Who sets the rule | Plan medical policy | Employer group plan design, verified per patient |
| Amount at stake per line | Often thousands | Usually 40 to 1,800 dollars, which is why it gets written off |
That last row is the quiet one. A denied medical claim is worth fighting because the number is large. A denied D0274 is worth 30 or 40 dollars, so it gets adjusted off, and the loss only becomes visible when someone adds up a year of them.
What is the decline rate of UnitedHealthcare claims?
For dental, there is no published figure, and any number you see quoted for dental specifically should be treated as an estimate someone made rather than a statistic someone measured. Commonly quoted marketplace medical figures have ranged from the high teens to the low thirties depending on the year and the issuer definition. None of those apply to a CDT claim.
Calculate your own. The arithmetic takes an afternoon and is worth more than every chart on the subject combined. Count claim lines, not claims: a claim with four lines where one denies is neither a denied claim nor a clean one.
Here is the calculation with illustrative numbers for one payer over one month:
| Metric | How to get it | Example |
|---|---|---|
| Lines submitted | Count CDT lines on claims sent to this payer | 412 |
| Lines denied or zero paid on first pass | Remittance lines with a reason code and no payment | 47 |
| Initial denial rate | 47 divided by 412 | 11.4 percent |
| Lines paid after correction or appeal | Same lines, resolved later | 31 |
| Final denial rate | 47 minus 31, divided by 412 | 3.9 percent |
| Overturn rate | 31 divided by 47 | 66 percent |
Those figures are illustrative, not a benchmark. Both rates matter because they point at different problems. A high initial rate with a high overturn rate is a submission problem: the claims were always payable, a data element was missing, and you paid for the round trip in staff time and aging. A low initial rate with a low overturn rate means the denials are correct, and the fix sits in verification and case presentation rather than the claims queue.
Run the same table per payer, per month, for twelve months. That is the graph worth having.
Why does United Health deny so many claims?
Strip out the conspiracy framing and the dental answer is mostly mechanical. Claims are adjudicated by rules engines that check frequency, age, tooth history and plan exclusions before a human is ever involved, and anything that fails a check comes back with a code. A large payer running a large book produces a large number of those, which is arithmetic rather than intent.
The codes tell you which rule fired. These are the ones that generate most of the volume in a general practice, with the fix that actually clears them.
| CARC | Reason, in brief | Typical dental cause | What clears it |
|---|---|---|---|
| 16 | Claim or service lacks information, or has a submission error | Missing tooth number, surface, quadrant or attachment | Correct and resubmit, do not appeal |
| 18 | Exact duplicate claim or service | Claim resent while the first was still in process | Check status before resending |
| 29 | Time limit for filing has expired | Claim sat in a work queue past the payer deadline | Nothing after the fact, so calendar each payer's limit |
| 96 | Non-covered charge | Plan exclusion, or benefit not in the contract | Patient responsibility if you disclosed it in advance |
| 97 | Benefit included in the payment for another service | Buildup bundled into the crown, palliative billed with an exam | Verify bundling rules before quoting |
| 119 | Benefit maximum for this period or occurrence reached | Annual maximum spent, or frequency limit hit | Track remaining benefit at verification |
| 197 | Precertification, authorization or pre-treatment absent | Predetermination required and not obtained | Obtain before scheduling, not after |
| 109 | Claim not covered by this payer or contractor | Wrong payer, terminated plan, or coordination of benefits order wrong | Re-verify and rebill to the correct payer |
Note how many of those are fixable before the claim goes out. Our breakdown of the top 10 reasons for dental insurance claim denials works through the same list from the front desk side. Two categories deserve separate attention because they are adjudicated differently: a missing tooth clause denial is a plan design question rather than a coding error, and orthodontic denials turn on lifetime maximums and age limits that no claim edit will catch.
One more category is worth naming because practices create it themselves. Code selection that does not match the documentation gets denied and, if it recurs, invites a review. Our guide to overcoming a dental claim denial for upcoding covers how to answer one.
Build the graph that pays for itself
Six steps, in order, and none of them require a chart from the internet.
- Export twelve months of remittance data from your practice management system, at the line level, with the reason codes attached. Twelve months, because frequency limits and annual maximums run on plan years and a quarter of data will lie to you.
- Define a denial once, in writing. A line that adjudicated to zero with a reason code is a denial. A clearinghouse rejection that never reached the payer is not, and mixing the two inflates the number and hides where the failure happened.
- Segment by payer and by employer group, not by carrier name. Since the employer picks the provisions, the carrier is too coarse a unit. Two groups under one carrier often behave like two different payers.
- Segment by procedure category. Denials concentrate. Perio therapy such as D4341 and D4342, crowns and buildups such as D2740 and D2950, frequency limited diagnostics such as D0274 and D1110, and implant lines such as D6010 will usually account for most of the dollars.
- Plot initial and final denial rates monthly on the same axes. The gap between the two lines is the money you are recovering; the height of the lower line is the money you are losing.
- Set a trigger, not a target. For example: any payer and category pair that crosses 15 percent initial denial in a month gets pulled for a root cause review that week, before the appeal window closes.
That last point matters more than the graph. Appeal deadlines are payer and plan specific and can be as short as 90 days from the remittance date, so a denial found in a quarterly review may already be dead. Our guide to how long you have to appeal a dental claim denial covers how those clocks run and where to find each payer's limit in writing.
Mistakes that make denial data useless
Comparing counts instead of rates. The largest payer will always have the most denials. Only the rate is comparable.
Counting front end rejections as denials. A claim rejected for a bad subscriber ID never reached adjudication. It belongs in a separate bucket with a separate fix, usually at check-in.
Treating one number as the whole carrier. A single national percentage averages across product lines that have nothing to do with each other, which is exactly the flaw in the viral charts.
Ignoring the overturn rate. A denial that you routinely win is a workflow cost, not a coverage denial. Reporting them together conceals both.
Writing off small balances rather than reworking them. This is the largest quiet loss in most offices, and it is invisible in any report that measures denials by dollar value alone.
Escalating before exhausting the payer's own process. As of this writing, appeal and external review rights depend on whether the plan is fully insured or self-funded under ERISA, and state departments of insurance generally have no jurisdiction over self-funded plans. Confirm the plan type and your options with your state insurance department before you spend the time, and use the NAIC complaint guidance when the plan is in fact state regulated.
What to do with the number once you have it
A denial rate is a diagnostic, not a scoreboard. Once you can see which payer, group and code combination produces the failures, most of the work moves upstream: verify frequency and waiting periods before scheduling, attach the radiograph the first time, and quote from the plan's actual allowable so the patient portion is right on the day. Our guide to maximizing dental insurance reimbursement rates covers the contracting side.
Curo builds that per payer and per group denial history from your remittances, flags lines matching a pattern it has seen before, and routes them for rework while the appeal window is open. The denial management overview walks through what that looks like in your own data.
The chart that started this is not fake. It is a fair reading of a small, self-reported, medical, marketplace dataset, and it became famous because it confirmed something people already felt. Just do not let it set expectations for a Tuesday morning batch of CDT claims. The only denial rate that will ever change what you collect is the one with your own tax ID on it.