The Patient in the Bed and the Patient in the Chart

There are really two versions of every hospitalized patient.

There is the patient in the bed.

And then there is the patient in the chart.

At the bedside, the patient is complicated.

You watch them get out of bed.

You see how long it takes.

You watch them grab the furniture.

You see whether they understand their oxygen tubing.

You notice that they can technically walk down the hallway, but only because a physical therapist is following behind them with a chair in case they suddenly have to sit down.

You hear them tell you they are nervous about going home.

You know what happens when they walk to the bathroom instead of 50 feet down a perfectly flat hospital hallway.

Then there is the chart.

Ambulated 50 feet.

Shortness of breath with exertion.

Oxygen saturation 89%.

Recommend home health.

Those statements may all be accurate.

But they are not the same thing as standing next to the patient.

And once another organization has to make a decision about that patient, the distinction matters.

Because the insurer does not spend 12 hours in the room.

The skilled nursing facility does not.

The durable-medical-equipment company does not.

For many of the people deciding what happens next, the patient they meet is the patient we documented.

This became much more obvious to me once I started thinking seriously about discharge planning.

Inside the hospital, a lot of information exists almost informally.

The nurse knows one part.

Physical therapy knows another.

Occupational therapy sees something different.

Respiratory therapy may understand the oxygen needs better than anyone else.

The physician knows the medical plan.

Case management understands what the insurance company and receiving facility need before the next step can happen.

Together, those people may have a pretty good understanding of the patient.

But another organization cannot approve something based on:

“Trust us. We know this patient.”

There has to be evidence.

Medicare Advantage regulations make that explicit. When an MA organization makes a medical-necessity determination, federal rules require it to consider the enrollee’s medical history—including diagnoses, conditions and functional status—along with physician recommendations and clinical notes. The plan also has to follow applicable Medicare coverage rules. [1]

That sounds completely reasonable.

If an insurer is being asked to pay for skilled nursing facility (SNF) care, inpatient rehabilitation, oxygen or another covered service, somebody should be able to demonstrate why the patient needs it.

The problem is that the quality of that decision can only be as good as the information that survives the trip from the bedside to the reviewer.

One example that has always stood out to me involves patients with significant lung disease.

Imagine two patients who both walk 50 feet.

On paper, that sounds identical.

But maybe the first patient walks independently, manages their oxygen tubing, turns around safely and sits back down without much difficulty.

The second patient also technically walks 50 feet.

But they become extremely short of breath.

They struggle not to get tangled in the nasal-cannula tubing.

A therapist follows them with a chair because they may need to sit suddenly.

They require constant cueing.

They finish the walk exhausted.

Those are not functionally identical performances.

But if the cleanest number traveling through the chart is:

50 feet

then I worry about how much of the real performance survives.

I am not arguing that insurance reviewers literally look at nothing except distance. Medicare Advantage rules specifically require consideration of functional status, physician recommendations and clinical notes. [1]

The larger point is that a number is easy to transmit.

Context is harder.

Researchers have found that mobility documentation really can become fragmented across professions. A 2025 observational study across four hospital sites identified 22 different documentation sources used to communicate patient mobility. Physical therapists and nurses looked for and recorded different kinds of mobility information, and an audit of 104 records found information that was variably missing or inconsistent between sources. [2]

That study was conducted in Australia, not the United States, so I would not use it to claim every American hospital has the same problem.

But the mechanism feels very familiar:

Everybody is documenting the patient. Nobody is necessarily documenting the patient in exactly the same way.

I think part of the reason is that every discipline is looking at the patient through a different lens.

Physical and occupational therapists are deeply involved in function.

Can the patient transfer?

Walk?

Use equipment safely?

Perform activities of daily living?

What assistance do they require?

What rehabilitation needs remain?

Physicians are answering a different set of questions.

What medical condition is being treated?

Is the treatment still medically necessary?

What medication or equipment is being prescribed?

What clinical problem justifies that prescription?

Nurses contribute another view.

We spend long stretches of time watching what the patient actually does between therapy sessions and physician rounds.

Respiratory therapists may know details about oxygen delivery and respiratory support that are not obvious anywhere else.

Case managers then have to take pieces of all of that information and translate them into an actual transition:

What setting is being requested?

What does the payer require?

Which facility can take this person?

Which documents need to be sent?

What equipment company can provide what has been ordered?

It is tempting to ask:

Who owns the chart?

During our conversation, my first instinct was the physician.

After looking at the evidence, I think that is too simple.

Physicians have enormous responsibility. Physician recommendations and clinical notes explicitly matter in Medicare Advantage coverage determinations, and orders from physicians or other authorized treating practitioners are required for many services, including home oxygen. [1][5]

But the functional patient is an interprofessional product.

CMS’s skilled-nursing guidance makes that visible. When skilled need depends on the patient’s condition and complications, the record can require support from physician orders and notes as well as nursing or therapy documentation. CMS also tells providers to verify that information in the medical record is correct and to avoid vague or subjective descriptions that do not sufficiently show the need for skilled care. [4]

So maybe nobody gets to say:

“That part wasn’t mine.”

This is where documentation stops being a clerical issue and starts becoming a coverage issue.

CMS says SNF records should include a detailed rationale for skilled services, the complexity of those services, the patient’s response and other patient characteristics. It specifically warns against vague or subjective descriptions that do not sufficiently demonstrate the need for skilled care. [4]

That means:

“Patient weak”

is not the same as explaining what the patient cannot safely do.

“Needs assistance”

is less useful than documenting what type of assistance, how much, and why.

“Short of breath with activity”

does not communicate the same thing as documenting what happened during activity, how much support was required and how the patient’s function changed.

This is not about discovering some secret vocabulary that unlocks insurance approval.

It is about turning an observation into information another clinician can understand.

I have also seen what happens when different parts of the record point in different directions.

A physical therapist may recommend short-term rehabilitation.

Occupational therapy may document that home with services could be appropriate.

Now the chart contains two different answers to what sounds like the same question.

In my experience, that can stall the process.

The team may want another therapy assessment before pursuing the post-acute plan. If the disagreement happens before a weekend or before the next scheduled therapy visit, that reassessment can add time before the authorization process even begins.

I want to make an important correction to how I first described this.

There is not a universal Medicare rule saying PT and OT must both vote “rehab” before someone can receive SNF care.

Medicare coverage depends on the applicable benefit and medical-necessity requirements, not a two-therapist voting system. A recommendation for rehabilitation does not by itself establish covered SNF care; the record must support the need for daily skilled services and the other applicable requirements. [4]

But disagreement still matters.

A 23-hospital study of 427 stroke patients found that discharge planners and physical therapists agreed on the recommended destination in 83.1% of cases. When they agreed, the actual discharge destination matched their consensus recommendation 92.5% of the time. Insurance barriers were among the reasons cited when the final destination differed. This was a study of recommendations and destinations, not proof that disagreement caused a delay. [3]

So the point is not:

Two therapy notes are legally required.

The point is:

When the record itself cannot tell a coherent story about the patient’s function, everyone downstream has a harder decision to make.

Home oxygen is probably the cleanest example I know of the difference between:

“This seems helpful.”

and

“The record establishes that Medicare covers it.”

At the bedside, you may watch a patient breathe easier with supplemental oxygen.

They may walk farther.

They may tell you they feel substantially better.

That can be clinically meaningful.

But Medicare’s home-oxygen benefit has actual qualifying criteria.

Under the common Group I criteria, an oxygen saturation of 88% or lower at rest on room air can establish qualifying hypoxemia; specified sleep or exercise findings can also qualify. Other coverage and testing requirements still apply. For exercise qualification, a patient whose saturation is 89% or higher at rest can qualify if the saturation drops to 88% or lower during exercise and supplemental oxygen improves that demonstrated hypoxemia. [5]

So consider a patient whose oxygen saturation reaches 89%, not 88%.

From the bedside, I may still think:

This person looks dramatically better with oxygen.

But 89% does not satisfy that particular Group I threshold.

Group II permits 89% with dependent edema suggesting heart failure, documented pulmonary hypertension or cor pulmonale, or erythrocythemia with hematocrit above 56%. Arterial blood-gas and other specified coverage pathways also exist. So 89% alone cannot settle eligibility. [5][6]

And here is where research offers an important counterweight to my bedside instinct.

In the Long-Term Oxygen Treatment Trial, 738 patients with stable COPD and moderate resting or exercise-related desaturation were randomized to long-term supplemental oxygen or no long-term oxygen. The trial found no significant improvement in time to death or first hospitalization, and no sustained benefit in measured quality of life, lung function or six-minute walking distance. [7]

That does not mean oxygen never helps a patient feel better in the moment. This trial studied stable COPD, not acute illness, severe resting hypoxemia or every diagnosis. Its exercise group could drop below 89% during walking; it was not simply a trial of people who never crossed Medicare’s oxygen threshold. [7]

It means that seeing an immediate bedside benefit does not automatically prove that long-term oxygen improves the outcomes we care about in the population studied.

That makes the coverage threshold easier for me to understand, even when it is frustrating in an individual case.

Coverage criteria are not always asking:

“Would this help?”

They may be asking:

“Has the documented clinical situation met the defined coverage standard?”

Those questions overlap.

They are not identical.

The oxygen example gets more complicated when a patient’s needs change during hospitalization.

Patients sometimes describe themselves as being on “three liters at baseline.”

But what does that mean?

Three liters sitting on the couch?

Three liters while sleeping?

Three liters continuously?

Three liters at rest but five while walking?

Those are very different things.

If the admission note simply carries forward:

3 L baseline

that statement can become the patient’s accepted story surprisingly quickly.

Then the disease progresses.

Maybe the patient’s actual needs after hospitalization are higher.

If nobody clearly documents the new requirement, the old description can keep traveling.

Medicare’s oxygen policy is explicit that qualifying testing has to occur at the time of need. For an inpatient being discharged, that means within two days of discharge. [5]

The current oxygen LCD also has separate payment rules when stationary oxygen flow exceeds four liters per minute. For beneficiaries qualifying in Group I or II, a higher allowance requires a qualifying blood-gas study performed while the patient is on four or more liters per minute; without that supporting qualification, payment can be limited to the standard allowance. [6]

That is more precise than saying:

“If seven liters isn’t written down, insurance denies the concentrator.”

Here, “blood-gas study” can mean qualifying oximetry or an arterial blood-gas test. It does not always mean an arterial blood draw. [6]

Different equipment, circumstances and payers matter.

But the underlying bedside observation survives fact-checking:

A changing oxygen need has to become a documented oxygen need before another organization can reliably act on it.

The oxygen requirement that exists only in the nurse’s head does not travel home with the patient.

I do not think this problem exists because clinicians do not care.

Hospitals produce enormous amounts of documentation.

And each profession tends to record the information most relevant to its own work.

That can create blind spots.

An older study comparing structured assessments with hospital records in patients age 75 and older found that impairments in basic activities of daily living were missing from 40% to 60% of nursing documentation and 80% to 97% of physician documentation. Instrumental activities of daily living were absent even more often, and cognitive function appeared in only about 30% to 40% of records. [8]

That study involved two Nordic hospitals and was published in 2008, so it should not be treated as a current estimate of documentation quality in American hospitals.

But it demonstrates something important:

Important functional information does not automatically become visible simply because a medical record exists.

Someone still has to notice it.

Someone still has to document it.

And someone else still has to find it.

This may be the most important counterargument in the entire article.

It would be dangerous to tell patients:

If your clinicians just document everything correctly, insurance will approve the care.

That is not true.

Documentation is evidence.

It is not a magic password.

In a report published in 2022, the HHS Office of Inspector General reviewed a sample of prior-authorization denials issued June 1–7, 2019, by 15 large Medicare Advantage organizations. OIG estimated that 13% of denied prior-authorization requests actually met Medicare coverage rules. In some cases, insurers said there was not enough documentation, while OIG reviewers concluded that the existing medical record was already sufficient to support medical necessity. [9]

The newer SNF data are even harder to ignore.

OIG’s 2026 report examined June 2024 SNF admission decisions from 19 large Medicare Advantage organizations: 12% were denied, with organization-level rates ranging from 0.4% to 23%. Eighteen percent of denials were appealed; 95% of those appealed were overturned. [10]

naviHealth reviewed about half the requests and denied 14%, versus 11% for internal reviews and 9% for other contractors. MA organizations overturned 97% of appealed naviHealth denials. [10]

Those numbers need restraint.

Appealed cases are a selected subset. An overturn can involve new documentation. OIG did not determine how many initial denials were inappropriate, and these comparisons do not hold patient mix constant. [10]

But the variation still matters.

The same Medicare benefit does not produce identical decision patterns across organizations.

Interpretation, workflow and contractor practice are possibilities worth investigating. These data do not tell us what each reviewer would do with the same chart.

Sometimes the chart really is incomplete.

Sometimes the chart is adequate and the reviewer still gets the decision wrong.

Those are different problems.

There is another danger here.

Once people realize how consequential documentation can be, it becomes tempting to talk about:

“What words get rehab approved?”

I hate that framing.

Clinical documentation should not be an exercise in manipulating coverage.

If a patient walked independently, the chart should not imply they required assistance because that makes rehabilitation easier to authorize.

If someone does not meet oxygen criteria, we should not manufacture a saturation number that gets the equipment covered.

The goal is not to make the record look worse.

The goal is to make the record more accurate.

If someone walked 50 feet but required a chair follow, say that.

If they needed repeated cues to manage oxygen tubing, say that.

If their resting oxygen requirement differs from their exertional requirement, make the distinction clear.

If the home situation materially changes what safe function looks like, document the relevant information in the appropriate place.

If PT and OT disagree, the answer may not be to pick whichever recommendation is more convenient.

Maybe the patient genuinely needs another assessment.

Good documentation should make the patient easier to understand—not easier to sell.

During the conversation that became this article, I said something pretty strong.

I said the physician probably carries the largest responsibility because physicians write progress notes, place orders and ultimately prescribe many of the services or equipment a patient leaves with.

There is still truth in that.

If the physician never incorporates a major change into the medical assessment and plan, important information can absolutely get lost.

But I think I was assigning responsibility too narrowly.

A doctor cannot recreate a physical-therapy assessment from memory.

A therapist cannot independently prescribe every piece of equipment.

A nurse’s twelve hours of observation do not automatically appear in a physician progress note.

A case manager cannot invent clinical evidence that nobody documented.

And a Medicare Advantage reviewer is explicitly expected to consider more than the physician note: current federal regulations mention medical history, functional status, physician recommendations and clinical notes. [1]

So the responsibility is shared.

Which is uncomfortable, because shared responsibility can easily become nobody’s responsibility.

I do not think the solution is simply:

Everybody needs to document more.

Clinicians already spend an extraordinary amount of time documenting.

More words can make the problem worse if the important information becomes harder to find.

What I think matters is better transmission.

If function is important to the next decision, the function should be easy to find.

If a patient’s oxygen needs changed, the current requirement should not be buried underneath a copied-forward “baseline” from three days earlier.

If therapy recommendations conflict, the team should know they conflict.

If a coverage decision depends on a particular clinical fact, somebody should be able to identify whether that fact has actually been established.

The mobility-documentation study is interesting here because the problem was not simply absence of charting.

Researchers found 22 different places where mobility information could live. [2]

Twenty-two places to document mobility may actually be worse than one good place.

The future probably should not be clinicians typing the same fact into more boxes.

It should be healthcare systems getting better at turning multidisciplinary observations into a coherent, current picture of the patient.

This is what I ultimately want families to understand when an insurance decision seems disconnected from what they are seeing.

The person making that decision may not have watched your mother struggle to stand.

They may not have seen your father trying to keep his oxygen tubing untangled.

They may not have been there when the patient made it 50 feet only because somebody followed with a chair.

They have the evidence that made it into the record.

That does not automatically make the decision right.

OIG’s work makes very clear that even records containing sufficient evidence can still result in inappropriate denials. [9]

But it explains why documentation matters so much.

Healthcare constantly asks one institution to make decisions about work performed inside another institution.

The chart is how the patient crosses that border.

And every time we reduce a complicated human being to a checkbox, a distance, an oxygen saturation or a copied-forward phrase, there is a risk that something important gets lost in translation.

The goal should not be to document a patient who looks sick enough to qualify.

The goal should be to document the patient who is actually there.

Because eventually the person in the bed leaves the room.

The patient in the chart keeps traveling.