The People Behind the Denial

Read a denial letter slowly and an institution begins to appear.

There is a category into which the request has been placed. There is a rule. There may be a code, a deadline, a reference to information that was not received or a criterion that was not met. Somewhere there is a route to appeal.

What may be absent is an ordinary explanation of what the person should do next, who can correct the decision and how long the correction can reasonably take.

The letter can be technically complete while leaving its recipient practically stranded.

This is where power becomes intimate. A decision about a service enters the body, the household and the night. It asks the sick person, or someone caring for them, to become an administrator at the moment when they may have the least capacity for another task.

No theory of efficiency is adequate if it refuses to count that work.

The strongest claim needs the right denominator

A 2026 report from the U.S. Department of Health and Human Services Office of Inspector General examined requests for admission to skilled nursing facilities made in June 2024 to nineteen large Medicare Advantage parent organizations. These organizations accounted for 86 percent of enrollment. About 12 percent of the requests were denied. Eighteen percent of denials were appealed, and 95 percent of those appealed denials were overturned. The median interval from initial request to appeal decision was six days.¹

Those figures identify an important problem worth investigating. They do not mean that 95 percent of all denials were wrong. People who appeal may differ from those who do not, and additional information can affect a decision. The administrative data did not determine how many initial denials were inappropriate. Still, a system in which nearly all challenged decisions in this bounded group were reversed raises serious questions about initial decision quality and the burden placed on beneficiaries and providers.

An earlier OIG report provides a different kind of evidence. It reviewed a stratified sample of denials from fifteen large Medicare Advantage organizations during one week in June 2019. It estimated that 13 percent of denied prior-authorization requests met Medicare coverage rules. The review involved coding experts and physician review of a subset.²

That was a historical sample of denied requests, not all care requested or every insurer. Its significance lies in identifying decisions that met the applicable standard yet were denied. A later reform or improved process should be assessed on its own evidence rather than assumed absent.

These findings are sufficient to justify scrutiny. We do not need to multiply them into an invented nationwide death toll to make the matter morally serious.

An appeal can correct and conceal

An effective appeal is valuable. It can correct a mistake, bring new evidence into view and allow an independent judgment. A system without an appeal would be worse.

But a successful appeal can also conceal a defective first decision if the institution counts only the eventual resolution. The person got the service in the end. The file is closed. The delay, fear, administrative labor and possible deterioration during the interval fall outside the success measure.

Think of a process with two gates. At the first, some valid requests are rejected. At the second, well-supported challenges are often accepted. The institution can describe the second gate as proof that its safeguards work. The person who lacked the time, language, support or clinical documentation to reach that gate experiences a different system.

The question is not only whether appeals succeed. It is who can appeal, what it costs them and whether the pattern of reversals changes the original process.

An appeal mechanism should generate learning. If a rule repeatedly produces decisions that are reversed, the institution should investigate the rule, its interpretation and its implementation. If missing information is a frequent cause, the system should examine whether it can obtain that information before denying a request. If one provider or reviewer has unusual outcomes, there should be a careful audit rather than automatic suspicion or automatic protection.

Correction should travel upstream.

Otherwise, the appeal becomes a second application process reserved for people capable of doing the work twice.

Scroll sideways to see all columns. Keyboard: focus the table and use the left and right arrow keys.

Text equivalent: A decision should be repairable

Stage or condition	What an answerable process requires
Across every step	Time limits matched to urgency; help, translation and accessible formats.
1. Initial decision	Name the responsible institution, the rule and the deadline.
2. Explanation and information	Give usable reasons and obtain missing evidence.
3. Accessible review	Provide independent judgment through a route the person can use.
4. Remedy	Correct the decision or harm when justified; confirm the remedy reaches the person.
5. Upstream correction	Investigate patterns and repair faulty rules or processes.
Return to the first decision	Assign an owner and verify changes. The dashed return is a responsibility to build, not automatic learning.

Figure 2. Correction should reach the person and the process. Help and time limits belong throughout; an appeal does not by itself guarantee a remedy or a better first decision.

Nobody owns the whole consequence

Healthcare decisions cross organizational boundaries. A clinician recommends treatment. A practice submits information. A plan applies coverage rules. A contractor may supply a review system. A hospital has its own discharge pressures. A public agency establishes requirements and supervises some parts of the process.

Each participant may have a defensible account of its role. The patient needs the roles to add up to care.

This is where fragmented responsibility becomes dangerous. A missing document can travel between organizations while each waits for the other. A provider may say the insurer has delayed; the insurer may say the provider has not supplied what is required. Both statements can contain truth. The patient is poorly placed to determine which next step will end the loop.

A humane system must assign responsibility for resolving the coordination problem. It should not turn the patient into the unpaid manager of institutions that possess far more information and resources.

That might mean a case owner with authority to contact the relevant parties, a shared status record, a deadline for identifying missing information or an independent navigator for complex situations. The exact design will vary. The obligation should be clear: somebody must be responsible for getting an answer that the person can use.

An institution that can issue a consequential denial should be able to identify who owns it.

Why review exists

We need to confront the strongest practical defense of gatekeeping. Healthcare resources are finite. Some treatments are ineffective, inappropriate or unnecessarily costly. Fraud and overbilling exist. A clinician’s request should not automatically require unlimited payment regardless of evidence.

Review can protect patients and shared resources. A process that distinguishes useful care from harmful or wasteful care performs an important function.

The problem arises when the design rewards withholding without adequately measuring the cost of an erroneous refusal. It also arises when approval becomes so administratively expensive that clinicians and patients abandon reasonable requests.

These are testable institutional questions. What are reviewers rewarded for? What evidence is available? Which errors are measured? Is the burden of a mistaken approval treated as more important than the burden of a mistaken denial? Does the organization examine whether administrative savings merely move costs into hospitals, households or clinicians’ unpaid time?

A financial target does not establish malicious intent. It does establish a reason to inspect the behavior it encourages.

Good review requires clinical standards, transparent reasoning, appropriately qualified judgment and proportionate procedures. Urgent cases need routes suited to urgency. Repeatedly justified care should not necessarily require an identical struggle each time. An automated screen should not be allowed to make responsibility disappear when a human life does not fit its categories.

There is room to debate the best mechanism. There should be less room to debate whether the affected person deserves a comprehensible reason.

The price of getting through

The Federal Reserve’s survey of household well-being found that 26 percent of U.S. adults reported skipping some medical care because of cost in 2025, down from 28 percent in 2024. The measure includes several types of care, including dental care; it does not identify insurer denials or establish that every skipped service was lifesaving.³

The improvement should be acknowledged. So should the scale of the remaining difficulty.

Money and administrative access can compound one another. A person may be able to pay privately to avoid a delay, take unpaid time to pursue an appeal or hire help with a complicated process. Another person has to decide whether the effort required to obtain the benefit will cost more immediately than they can spare.

This makes the design of a process part of the distribution of care. A nominal benefit can be less valuable when accessing it requires resources its intended beneficiaries lack.

We sometimes call this a matter of health literacy, as though the primary defect were located in the person trying to understand. People can benefit from clearer information and support. Institutions can also stop making ordinary tasks unnecessarily difficult.

A person should not need to become unusually skilled at navigating a system in order to receive what the system says they are entitled to receive.

Dignity has procedural content

Dignity is often invoked as a warm word after the operational decisions have been made. It should influence those decisions.

A dignified process explains the issue without making the person feel that asking is an offense. It provides reasonable ways to supply information. It recognizes disability, language and the practical limits of a person under strain. It avoids demanding the same evidence repeatedly when the institution can retrieve it itself.

It also permits an answer to be challenged without treating the challenge as suspicious behavior. A patient may disagree, misunderstand or be frustrated. None of those facts justifies humiliation.

This obligation extends to staff. A worker given impossible caseloads and punished for spending time on a difficult case will struggle to provide humane service. Training alone cannot compensate for an operating model that makes care a performance failure.

The institution should examine the quality of resolution, not only the speed of closure. Did the person understand the decision? Was the appropriate service obtained when approved? Was an error corrected before harm grew? Did the process reduce or increase avoidable work elsewhere?

These measures are harder than counting completed calls. They are closer to the purpose of the institution.

A public system can fail this test too

The profit motive creates particular tensions where denying payment can improve a payer’s financial position. It deserves scrutiny. Public administration can also become unresponsive, arbitrary or protective of its own routines.

A ministry can hide a shortage behind a procedure. A public hospital can tolerate poor treatment. An agency can make eligibility difficult because its budget is strained or because avoiding criticism has become more important than serving people. Removing a private insurer does not remove the need to govern decisions well.

My position is that access to necessary healthcare should be secured universally, and that the people administering it should face clear obligations to the people who need it. That commitment requires financing, clinical capacity and competent administration as well as a declaration of rights.

Different systems can be judged against those requirements. Does everyone have meaningful coverage? Can they obtain appropriate care in a medically reasonable time? Are costs shared fairly? Can errors be challenged and corrected? Is there independent oversight of both public and private power?

A public guarantee should make people more secure. If it leaves them pleading before an opaque gatekeeper, the guarantee needs further work.

The aspiration is ambitious because a person’s need for care should not become another person’s license for arbitrary control.

Put the burden where the knowledge lives

Institutions often hold the information needed to explain a decision. They know which rule was applied, which evidence was considered and which system produced the result. The person affected may receive only the outcome.

A fair process should require the institution to reveal enough of its reasoning for a meaningful challenge, while protecting legitimate privacy and security interests. A generic code or a circular explanation is insufficient when the consequence is serious.

The same logic applies to the discovery of patterns. Patients should not have to assemble thousands of individual cases to reveal a defect an insurer or agency can detect in its own data. Regulators should be able to examine denial patterns, reversals, delays and disparities using appropriate safeguards and sound methods.

Unusual rates require investigation, not instant conviction. Case mix, documentation and differing services can explain some variation. But a system should be designed to notice, ask and follow through.

There should also be consequences when an institution repeatedly fails its obligations. A mild administrative penalty that costs less than the behavior saves may do little to alter incentives. Remedies should be proportionate, lawful and strong enough to matter, with due process for those subject to them.

The person making the mistake may need training. The organization benefiting from the pattern may need a different rule.

Trust must have an object

We are often told that public trust is declining and must be restored. Restored to what?

A request for trust is meaningful when an institution can show that its decisions deserve it. That includes admitting errors, publishing useful information, responding to challenges and changing a process that causes avoidable harm.

Public-relations language cannot carry this burden indefinitely. If people repeatedly encounter a system that says one thing and requires another, their skepticism may be a rational response to experience.

At the same time, distrust can be exploited. A person who has been badly treated may become vulnerable to someone promising certainty, a secret cure or a simple enemy. Institutional failure creates an opening for other forms of power, some of them even less answerable.

That is another reason to repair the real thing. People need competent care, reliable information and institutions that can earn confidence through what they do.

Criticism should point toward that repair. It should identify a responsibility and a remedy rather than teaching people that every professional is an adversary and every process a trap.

The answer should arrive before the damage

An appeal that eventually proves a person right can still arrive too late. This is true in healthcare with particular force, but it extends to housing, wages, benefits and other essential decisions.

Time is part of justice. A remedy must fit the timescale of the harm.

That may require expedited review, interim protection where justified, assistance for people unable to navigate a process and obligations that begin before a formal complaint. It requires resources. A system cannot promise prompt scrutiny while starving the staff responsible for providing it.

The expense should be weighed against the cost of leaving errors in place. Delay can consume clinical capacity, household income, professional time and public confidence. Cheap administration can be very expensive elsewhere.

The person holding the denial letter should not need an extraordinary act of persistence to be treated as an ordinary human being. They should encounter an institution whose normal operation includes understanding, explanation and correction.

A society reveals its character in the amount of organized effort it requires from people who are already in pain.

We can require less.