Daniel Cham MD

Thursday, September 3, 2026

Mark Andrew Collins Asked for Help. His Story Exposes a Healthcare Handoff Problem Physicians Can’t Afford to Ignore.

The problem isn't always a lack of care. Sometimes it's what happens between the people who provide it.

“Redesigning workflows attacks the source of distress instead of giving people skills to cope with it.” — Dr. Liz Harry, Chief Well-Being Officer, Michigan Medicine, quoted by the American Medical Association

Mark Andrew Collins was 26 years old when he finally asked for help.

That sentence should bother every person who works in healthcare.

Not because asking for help guarantees recovery.

It doesn't.

Not because every emergency department can solve addiction in one visit.

It can't.

And not because Mark's story, by itself, proves that a particular clinician or hospital caused his death.

It doesn't.

But because Mark did something healthcare constantly tells patients to do.

He raised his hand.

He said, in effect:

I need help.

And then the system had to decide what happened next.

That is where this story becomes much bigger than addiction.

It becomes a story about handoffs.

And healthcare has a handoff problem.

Nine Days

Diane Santos had only nine days between learning about her son's opioid addiction and finding him dead.

Nine days.

There were tears.

Confessions.

Phone calls.

And, importantly, hope.

Mark wanted treatment.

According to his mother, he wanted to stop using drugs. He wanted to understand why he was using them. He had researched treatment and was looking forward to getting help, including therapy.

He was not simply refusing care.

He was looking for it.

He had even secured an appointment for early January.

The problem was that January was still weeks away.

Then withdrawal became severe.

Diane asked whether he wanted to go to the emergency department.

Eventually, he said yes.

They went to Backus Hospital in Norwich, Connecticut.

According to Santos' account, Mark told hospital staff that he had used fentanyl that morning.

He was evaluated.

He was not admitted.

A recovery coach spoke with him and provided a number to call the next day for treatment access.

Mark went home.

The next day, he called.

There was no immediate treatment available.

The plan was to keep trying.

The following morning, Diane went to wake her son.

His bedroom door was locked.

There was no answer.

She opened it.

Mark was on the floor.

He was cold.

He was gone.

He died on December 30, 2023, at age 26. His obituary identifies him as Mark Andrew Collins and says he had asked for help overcoming addiction.

The recent Connecticut reporting says his death occurred roughly 36 hours after his emergency-department visit.

That timeline is devastating.

But there is something else about the story that deserves attention.

The Most Dangerous Word in Healthcare

It may not be “denied.”

It may not be “delayed.”

It may not even be “error.”

It may be:

“Next.”

The patient goes to the next person.

The referral goes to the next department.

The authorization goes to the next queue.

The claim goes to the next system.

The denial goes to the next worklist.

The phone call goes to the next representative.

The patient waits.

The physician assumes someone else is handling it.

The staff member assumes the physician handled it.

The payer assumes the documentation was sufficient.

The practice assumes the payer received everything.

Everybody is busy.

Everybody is doing something.

And somehow, nobody owns the outcome.

That is the paradox of modern healthcare.

We have more specialists, more software, more portals, more protocols, more dashboards, more data and more artificial intelligence than ever.

Yet we can still lose a human being between two perfectly documented steps.

A Referral Is Not Treatment

This is where Mark's story becomes uncomfortable.

A healthcare organization can truthfully say:

“We connected the patient with a recovery resource.”

That sounds like action.

But did the patient actually receive treatment?

Those are not the same thing.

A phone number is not treatment.

A referral is not treatment.

An appointment request is not treatment.

A portal message is not treatment.

A discharge instruction is not treatment.

A task placed into someone's queue is not treatment.

The outcome is what matters.

And this distinction is not unique to addiction.

Consider a patient with a suspicious imaging result.

The scan is completed.

The radiologist reports an abnormality.

The report enters the EHR.

The ordering physician is notified.

A message is generated.

The patient is told to follow up.

Everyone can point to a completed action.

But six months later, nobody can answer a simple question:

Did the patient actually receive the follow-up care?

That is a handoff failure.

Or consider a prior authorization.

The physician orders a medically necessary procedure.

The staff submits the authorization.

The payer requests additional documentation.

Someone assumes someone else responded.

The authorization expires.

The procedure is delayed.

Technically, the prior authorization workflow existed.

Operationally, it failed.

Healthcare loves confusing activity with completion.

We count tasks.

We should count outcomes.

And Then There Is Medical Billing

This is where the conversation becomes particularly interesting for physicians.

Because the exact same structural problem exists in revenue cycle management.

A patient receives care.

But the money doesn't automatically follow the care.

Why?

Because healthcare revenue is also a chain of handoffs.

Patient → registration → eligibility → encounter → documentation → coding → claim → payer → adjudication → payment → posting → reconciliation.

One broken connection can contaminate everything downstream.

And here is the uncomfortable part:

Most healthcare technology is designed to optimize individual steps.

Better scheduling.

Better documentation.

Better coding.

Better clearinghouses.

Better denial management.

Better dashboards.

Better AI.

Better analytics.

Better portals.

But we keep asking:

“How do we make this step faster?”

The more important question is:

“Why did this information have to travel through six disconnected steps in the first place?”

That's a different question.

And it leads to a different kind of healthcare technology.

Healthcare Has a Favorite Illusion

It is called workflow.

Healthcare organizations love workflows.

There is a workflow for registration.

A workflow for referrals.

A workflow for prior authorization.

A workflow for documentation.

A workflow for coding.

A workflow for claims.

A workflow for denials.

A workflow for appeals.

Soon we will probably have a workflow for managing the workflows.

At some point, someone will build an AI agent to remind another AI agent to check whether the first AI agent completed its task.

We will call it innovation.

And someone will put it on a conference stage.

The problem isn't that workflows are bad.

The problem is that we often automate fragmentation instead of eliminating it.

Automation can make a broken process faster.

It can also make a broken process fail at scale.

The Contrarian View: AI Is Not the Solution to Bad Data

This is where the current AI conversation gets backward.

Healthcare is rushing to put AI everywhere.

AI coding.

AI clinical documentation.

AI prior authorization.

AI denial prediction.

AI scheduling.

AI patient messaging.

AI revenue-cycle management.

And yes, these tools can be useful.

But there is a fundamental problem.

AI cannot manufacture trustworthy context from missing information.

If the underlying data is incomplete, ambiguous, inconsistent or disconnected, the AI may simply produce a faster interpretation of bad inputs.

Garbage in, garbage out is old.

The modern version is:

Garbage in, beautifully summarized.

That's progress of a sort.

But it isn't transformation.

The Real Problem Is Upstream

Suppose a physician documents:

“Patient presents with worsening symptoms. Continue management.”

That sentence may be clinically understandable to the physician.

But downstream, it may be remarkably unhelpful.

The coder needs specificity.

The billing system needs specificity.

The payer may need specificity.

The quality system may need specificity.

The next clinician may need specificity.

The patient may need specificity.

The problem didn't begin with the coder.

It began when the information was captured.

This is why I believe healthcare billing is fundamentally a data-quality problem before it is a tooling problem.

The claim is simply the final expression of everything that happened upstream.

If the information is incomplete at the point of capture, somebody downstream will eventually pay for it.

Sometimes the cost is a denial.

Sometimes it is staff time.

Sometimes it is a delayed payment.

Sometimes it is an appeal.

Sometimes it is a frustrated patient.

And in clinical care, the consequences can be much more serious.

The Five Handoffs Every Practice Should Audit

Forget the software demo for a moment.

Walk through the patient journey.

1. Patient → Front Desk

Did the practice capture the information correctly?

Or did someone type something quickly because the waiting room was full?

Measure: registration accuracy.

2. Front Desk → Clinical Team

Did the clinical team receive the information that actually matters?

Or did important context disappear between intake and the exam room?

Measure: missing or corrected intake data.

3. Clinical Encounter → Documentation

Did the medical record accurately capture the service delivered?

Not merely enough to close the note.

Enough to support the clinical, operational and financial consequences of the encounter.

Measure: documentation completeness and charge lag.

4. Documentation → Claim

Did the structured information actually support the claim?

Or is the billing team reconstructing the encounter after the fact?

Measure: first-pass yield, edits and preventable denials.

5. Claim → Payment

Did the claim become money?

Or did it become another work queue?

Measure: clean-claim rate, denial rate, days in A/R and avoidable rework.

The Billing Department Is Often Being Asked to Solve a Problem It Didn't Create

This may be the most underappreciated issue in medical billing.

A claim gets denied.

So what happens?

Someone blames billing.

The biller opens the chart.

Reads the note.

Looks for missing information.

Sends a message.

Waits.

Contacts the physician.

The physician modifies or clarifies documentation.

The claim gets corrected.

Then everyone celebrates.

The claim was fixed.

But was the system fixed?

No.

The system simply learned how to repair the damage.

That is reactive healthcare.

And reactive healthcare is expensive.

The Cheapest Denial Is the One That Never Exists

This sounds obvious.

It isn't how most organizations operate.

Many revenue-cycle teams are optimized around working denials.

How many were received?

How quickly were they touched?

How many were appealed?

How many dollars were recovered?

Those are useful metrics.

But they measure the cost of failure.

A more interesting question is:

Why did the denial happen in the first place?

Was the patient's insurance information wrong?

Was eligibility checked too late?

Was the authorization missing?

Was the diagnosis insufficiently supported?

Was the procedure documentation incomplete?

Was the payer rule misunderstood?

Was the wrong payer pathway used?

Was information available in one system but invisible to another?

The denial is the symptom.

The upstream data failure is the disease.

This Is Where Healthcare Gets Weird

Imagine an airline.

A passenger buys a ticket.

The airline checks the passenger.

The passenger boards.

The aircraft departs.

Then, after landing, someone discovers that the airline never actually recorded where the passenger was supposed to go.

Nobody would say:

“Don't worry. We have a really sophisticated airport team that specializes in finding passengers after they land.”

Yet healthcare routinely operates this way.

We build enormous departments to repair problems that should have been prevented earlier.

Then we congratulate ourselves for recovering 80% of the lost revenue.

That is not efficiency.

That's organized recovery from preventable confusion.

Mark's Story Forces a Bigger Question

Diane Santos has spent the years since her son's death advocating for better treatment and sharing his story.

The recent reporting says she later returned to Backus Hospital, where she found an emergency-department protocol for initiating buprenorphine that she believed could have helped Mark. She brought the issue to administrators and says changes were subsequently made.

That part of the story matters enormously.

Because it introduces a concept healthcare often struggles with:

The protocol existed.

The question was whether the system reliably translated the protocol into action.

That is a very different problem from not having a protocol.

And it is a problem that exists everywhere.

You can have:

a policy nobody follows,
a dashboard nobody checks,
an alert everyone ignores,
a referral nobody closes,
a claim nobody reconciles,
a denial nobody learns from,
and an AI tool nobody actually integrates into the workflow.

Healthcare doesn't merely need knowledge.

It needs execution.

Dr. Gail D'Onofrio's Point Is Bigger Than Addiction

Emergency physician and addiction-medicine expert Dr. Gail D'Onofrio has argued for a more routine approach to initiating treatment for opioid use disorder in emergency departments.

In the recent reporting, she described the need for a standardized process involving medication, harm reduction and continued treatment. She also acknowledged the complexity of emergency-department operations and the reluctance to legislate clinical medicine.

That tension is important.

Because healthcare has two competing instincts.

Standardize everything.

And:

Don't turn medicine into a checklist.

Both instincts are reasonable.

The answer isn't to replace physicians with protocols.

The answer is to make the right clinical action easier to execute reliably while preserving clinical judgment.

That distinction matters.

A protocol should support judgment.

It shouldn't replace it.

The Same Principle Applies to Billing

Imagine a physician sees a patient.

The practice already knows:

who the patient is,
what insurance they have,
what appointment they booked,
why they came,
what services are scheduled,
what authorizations are required,
what documentation is necessary,
what financial responsibility is expected.

Yet these facts often live in different systems.

So humans become the integration layer.

That's the hidden infrastructure of American healthcare.

People copying information from one box into another box.

And then we wonder why administrative costs are so high.

The Human API

Here is my favorite description of modern healthcare administration:

The human API.

When two software systems don't communicate, a person does.

When the EHR doesn't talk to the billing system, someone exports data.

When the payer portal doesn't integrate with the practice workflow, someone logs in.

When documentation doesn't support coding, someone sends a message.

When the claim is rejected, someone investigates.

When the referral isn't completed, someone calls.

When the patient doesn't understand the instructions, someone explains them.

Humans are constantly acting as APIs between disconnected systems.

And humans are expensive APIs.

More importantly:

Humans get tired.

Burnout Isn't Always a Physician Problem

We often talk about physician burnout as though the primary cause is seeing too many patients.

Sometimes it is.

But there is another form of burnout that receives less attention:

workflow absurdity.

The physician enters information.

The nurse re-enters information.

The front desk re-enters information.

The biller interprets information.

The payer requests the same information again.

The patient receives another form.

Then everyone asks why clinicians are exhausted.

Maybe the problem isn't that healthcare workers don't know how to work hard.

Maybe they are working hard on things that should not require human effort in the first place.

The Measurement Trap

Healthcare loves measuring productivity.

Patients per hour.

Claims per day.

Denials worked.

Calls answered.

Prior authorizations completed.

Notes closed.

Revenue collected.

But productivity can become dangerous when the measurement ignores the system.

A billing employee can process 100 claims.

Wonderful.

But what if 30 of them require rework?

A staff member can complete 50 prior authorizations.

Wonderful.

But what if the process requires information that should have been captured automatically?

A physician can close every note before leaving.

Wonderful.

But what if the documentation still fails downstream requirements?

Speed without quality is just faster rework.

And rework is one of healthcare's least celebrated industries.

What Should Physicians Measure Instead?

Start with five questions.

1. Where does information get re-entered?

Every manual re-entry is a potential failure point.

2. Where does responsibility become ambiguous?

If two people think the other person owns the task, nobody owns the task.

3. Where does work leave the system?

Every external portal, fax, spreadsheet or phone call deserves scrutiny.

4. Where are errors discovered?

The later the error appears, the more expensive it usually becomes.

5. What percentage of work is preventive versus corrective?

This may be the most revealing metric of all.

A mature organization doesn't simply become better at fixing mistakes.

It becomes better at preventing the mistakes from being created.

Three Myths Worth Killing

Myth #1: “We need more staff.”

Sometimes you do.

But adding people to a broken workflow can simply create a larger broken workflow.

Before hiring, ask:

What work are we hiring humans to compensate for?

Myth #2: “We need better AI.”

Maybe.

But first ask:

Is the underlying information structured well enough for AI to help?

If not, you may simply automate uncertainty.

Myth #3: “Our denial team is excellent.”

That's great.

But here's the contrarian question:

Why are you measuring how good your organization is at cleaning up preventable mistakes instead of measuring how few mistakes it creates?

The best denial is not the one recovered.

It's the one that never happened.

What a Better Practice Looks Like

A better practice does not necessarily have more software.

It has fewer gaps between systems.

The goal is simple:

Capture better information earlier.

Connect it.

Validate it.

Use it.

And prevent the same information from being repeatedly recreated by different people.

For a physician practice, that means building a chain of continuity:

Patient information → clinical context → documentation → coding → claim → payer response → payment.

Each stage should inherit trustworthy information from the stage before it.

Not a photocopy.

Not a reinterpretation.

Not a manual reconstruction.

Continuity.

The OnnX Thesis

This is the thinking behind what I'm building with OnnX.

I don't believe the future of medical billing is simply:

“Put AI on top of billing.”

That's too shallow.

The bigger opportunity is to move upstream.

Instead of waiting for the claim to fail, identify the missing information before the claim exists.

Instead of asking a biller to interpret a fragmented chart, structure the information closer to the point where it is created.

Instead of making humans the permanent bridge between disconnected systems, make the systems communicate.

Instead of treating billing as a separate department, treat it as the financial expression of the clinical and operational data generated throughout the practice.

That changes the question.

From:

“How do we work more denials?”

to:

“Why did we create the denial?”

From:

“How do we automate billing?”

to:

“How do we make the revenue cycle more deterministic?”

From:

“How can AI replace administrative work?”

to:

“How can better data eliminate unnecessary administrative work?”

That's a much more ambitious goal.

And, in my opinion, a much more useful one.

The Future Isn't More Automation

This may be the most contrarian prediction in this article.

The future of healthcare may not be about more automation.

It may be about less work.

Those are not the same thing.

Automation asks:

Can we make this task happen automatically?

Better system design asks:

Why does this task exist?

That distinction could save healthcare billions.

If a task is unnecessary, automating it is still waste.

If a handoff is unnecessary, automating the handoff is still waste.

If a form exists only because two systems cannot communicate, building an AI that fills out the form is still a workaround.

The ultimate automation is deleting the work.

And That Brings Us Back to Mark

Mark Andrew Collins is not a billing story.

He is not a technology case study.

He should not be reduced to a metaphor for software.

He was a 26-year-old man who wanted help.

His mother, Diane Santos, lost her son and then chose to turn grief into advocacy.

That deserves to remain the center of the story.

But his story also forces healthcare leaders to ask a difficult question:

What happens between “the patient needs help” and “the patient actually receives help”?

Because that space between those two sentences is where handoffs live.

And handoffs are where systems reveal themselves.

Dr. D'Onofrio's argument for making appropriate treatment more routine in emergency care points toward the same principle: don't rely entirely on heroic individuals to make a fragmented system work. Build processes that reliably connect patients to the care they need.

The lesson extends far beyond addiction.

It applies to referrals.

Prior authorizations.

Test results.

Medication reconciliation.

Discharge planning.

Documentation.

Claims.

Payments.

And everything in between.

Healthcare's Most Expensive Employee May Be the Person Who Keeps Fixing the Same Problem

Think about the person in your practice who everyone considers indispensable.

They know which payer portal to use.

They know which physician forgets to document something.

They know which claims are likely to deny.

They know which fax number works.

They know which insurance representative actually answers the phone.

They know the workaround.

They know the workaround for the workaround.

They are brilliant.

They are also a warning sign.

Because when one employee becomes the only person who knows how to make the system work, the organization hasn't created resilience.

It has created institutional dependency.

The goal should not be to eliminate that person's value.

It should be to capture their knowledge and build it into the system.

Otherwise, one vacation can become an operational event.

One resignation can become a crisis.

One missed handoff can become a patient-safety problem.

One bad data field can become a denial avalanche.

The Question Every Physician Should Ask

Not:

“Do we have enough staff?”

Not:

“Do we have an AI tool?”

Not:

“How many claims did we collect?”

Ask this instead:

“Where does responsibility disappear in our practice?”

Then follow the patient.

Follow the information.

Follow the money.

You may discover that the problem isn't one bad employee.

It isn't one bad payer.

It isn't one bad software platform.

It isn't even one bad process.

It is the space between processes.

That is where healthcare loses time.

That is where healthcare loses money.

And sometimes, as Mark Andrew Collins' story painfully reminds us, that is where healthcare can lose something much more important.

A Final Provocation

We often say healthcare needs more compassion.

I agree.

But compassion without execution can become another form of bureaucracy.

The system can care deeply.

The clinicians can care deeply.

The staff can care deeply.

The family can care deeply.

And the patient can still fall through the gap.

Good intentions don't close handoffs. Systems do.

That is why the future of healthcare shouldn't simply be about adding intelligence.

It should be about creating continuity.

Continuity of information.

Continuity of responsibility.

Continuity of care.

Continuity of revenue.

And ultimately:

continuity between what healthcare promises and what healthcare actually delivers.

Diane Santos has said that she continues to tell Mark's story because she wants his life and death to make a difference.

That is perhaps the most important lesson of all.

A tragedy should not become content.

It should become accountability, learning and change.

Mark asked for help.

His story should make us ask whether our systems are designed merely to respond to requests—

or actually to close the loop.

Because in healthcare, the difference between those two things can be enormous.

And sometimes, it can be everything.

What This Means for Your Practice Monday Morning

Don't start with another software purchase.

Start with a whiteboard.

Write:

Patient → Front Desk → Clinical → Documentation → Coding → Claim → Payer → Payment

Then ask your team:

Where does information get lost?

Where does somebody re-enter it?

Where does someone wait for someone else?

Where do we discover errors too late?

Where are we paying people to repair problems that better system design could prevent?

Then pick one handoff.

Fix it.

Measure the result.

Repeat.

That's how you move from reactive healthcare to deterministic healthcare.

Not by buying another dashboard.

By removing another failure point.

FAQ

Is Mark Andrew Collins' story evidence that the hospital caused his death?

No. Public reporting describes Santos' account of what happened and the timeline surrounding Mark's death. That information alone does not establish legal or medical causation. Any such conclusion would require review of the complete clinical record and other evidence.

Why use an addiction story to discuss medical billing?

Because the underlying lesson is not addiction-specific.

It is about handoffs, continuity, ownership and information integrity.

Those same structural problems appear in clinical care, prior authorization, documentation, coding, claims and payment.

Is AI the answer?

AI can help, but only when the underlying workflow and data are reliable.

The goal should not be to automate every broken task.

It should be to eliminate unnecessary work and make the remaining work more reliable.

What should practices measure?

Start with:

First-pass yield
Preventable denial rate
Charge lag
Days in A/R
Documentation corrections
Authorization turnaround
Manual re-entry
Rework volume
Exception volume
Percentage of problems discovered upstream versus downstream

The most valuable metric may be:

How much work did we prevent from happening?

The Bigger Lesson

Healthcare doesn't have a shortage of intelligence.

It has a shortage of continuity.

We have brilliant physicians.

Dedicated nurses.

Exceptional billers.

Talented administrators.

Sophisticated software.

Powerful AI.

And enormous amounts of data.

Yet we still create systems where each component can work perfectly while the overall patient journey fails.

That's the paradox.

Local optimization can create global dysfunction.

The front desk can be efficient.

The physician can be efficient.

The coder can be efficient.

The billing team can be efficient.

The payer can be efficient.

And the entire system can still be terrible.

Why?

Because healthcare isn't a collection of departments.

It's a chain.

And a chain doesn't become stronger because one link gets smarter.

It becomes stronger when the links actually connect.

About the Author

Dr. Daniel Cham is a physician and medical consultant with experience across medical technology, healthcare management and medical billing. His work focuses on practical solutions to the operational problems that sit at the intersection of clinical care, technology and healthcare economics.

Through OnnX, he is exploring a different approach to medical billing: improving the quality and continuity of information upstream so that revenue-cycle problems can be prevented rather than endlessly repaired downstream.

Connect with Dr. Cham on LinkedIn and follow his work across healthcare technology, medical practice and operational strategy.

Dr. Daniel Cham on LinkedIn

Continue the Conversation

Healthcare is changing quickly.

The people who understand the systems behind the change will be better positioned to shape what comes next.

Knowledge drives progress. Start your journey here.

Explore more insights, conversations and healthcare perspectives:

Visit Dr. Cham's website
Listen on Spotify
Watch on YouTube
Follow on X
Follow on Facebook

Free resource: Check the Featured section of my LinkedIn profile for a free download. No signup required.

Your Turn

Here's the question I want to leave you with:

Where does responsibility disappear inside your healthcare organization?

Is it between the physician and the coder?

The patient and the referral?

The practice and the payer?

The clinical note and the claim?

Or somewhere nobody has thought to look yet?

Tell me in the comments.

If this perspective challenges how you think about healthcare operations, share or repost this article and start the conversation with another physician, practice owner or healthcare leader.

Because maybe the biggest healthcare problem isn't that we don't have enough technology.

Maybe we've simply become very good at automating the gaps between things that should have been connected in the first place.

Mark Andrew Collins asked for help.

His story reminds us that healthcare's job isn't finished when we hand someone a phone number, close a task, submit a claim or move something to the next queue.

The job is finished when the loop is closed.

References

Cris Villalonga-Vivoni, Connecticut Insider / CT Post — September 3, 2026
“CT man sought help for opioid addiction. His mom found him dead 36 hours later: ‘Now I’m his voice’”
This is the primary human-interest story behind the article. It documents Mark Andrew Collins’ attempt to seek treatment, his mother Diane Santos’ advocacy, and the broader questions surrounding emergency-department access to opioid-use-disorder treatment.
Read the Connecticut Insider story
American Medical Association STEPS Forward — September 1, 2026
“Small Workflow Changes Can Make a Big Impact”
The AMA's current workflow-improvement discussion is highly relevant to the article's central argument: healthcare organizations can address friction by redesigning workflows rather than simply asking clinicians and staff to absorb more administrative burden.
Read the AMA STEPS Forward article
Medical Group Management Association (MGMA) — September 2, 2026
“Fewer than 1 in 10 practices see faster prior auth turnarounds in 2026”
MGMA reports that 44% of medical-group leaders said prior-authorization turnaround was slower in 2026, while only 7% said it was faster. It also highlights the hidden work surrounding authorizations—status checks, documentation, peer-to-peer reviews, denials, appeals and multiple payer portals.
Read the MGMA report

#Healthcare #HealthcareLeadership #HealthcareInnovation #PhysicianLeadership #HealthcareOperations #MedicalBilling #RevenueCycleManagement #HealthcareTechnology #HealthTech #PatientSafety #OpioidUseDisorder #AddictionMedicine #WorkflowDesign #HealthcareAI #ClinicalOperations #PhysicianBurnout #MedicalPractice #HealthcareTransformation #HealthEquity #OnnX

Wednesday, September 2, 2026

Dr. Paul Kalmansson, Leon Allan Loyd, and the Cello: The Five Minutes Healthcare Keeps Getting Wrong

What a physician, a veteran facing amputation, and a cello reveal about physician time, administrative burden, and the healthcare system we are building.

“The administrative burdens physicians shoulder each day directly affect the patient-physician relationship and unnecessarily interrupt the delivery of care.” — American Medical Association (AMA)

Then came the cello.

On September 2, 2026, a remarkable human-interest story emerged from Loma Linda, California.

Leon Allan Loyd, a U.S. Army veteran, was facing another amputation.

His medical journey had already taken something enormous from him. Now another surgery was approaching.

The day before the operation, his physician, Dr. Paul Kalmansson, did something unusual.

He played the cello.

Not because music could reverse the disease.

Not because Bach could change the surgical plan.

Not because the hospital needed another intervention on the chart.

Kalmansson played because he wanted his patient to feel that someone was there.

Kalmansson had earned a master's degree in music before attending medical school. He chose Bach's Prelude from the G Major Cello Suite for Loyd.

Loyd later described what the music gave him:

more strength and more hope.

His wife, Sheri Loyd, said Leon called her afterward and became emotional.

He was grateful.

Very grateful.

And the story became public after the couple's granddaughter shared video of the performance on social media.

Think about that for a moment.

A physician.

A veteran.

An impending amputation.

A cello.

Bach.

And a few minutes in which nobody was talking about productivity.

Nobody was talking about throughput.

Nobody was talking about claims.

Nobody was talking about reimbursement.

Nobody was talking about artificial intelligence.

Someone was simply caring for another human being.

And that is where this story becomes much bigger than music.

Because I think healthcare has a time problem.

Not merely a staffing problem.

Not merely a burnout problem.

Not merely an administrative problem.

A time problem.

And we have been trying to solve it with more technology without first asking the uncomfortable question:

What if the technology is sometimes consuming the very time it was supposed to save?

The Contrarian Take

Here is my unpopular opinion:

Healthcare does not have a technology shortage.

We have a friction shortage.

Actually, let me say that differently.

We have too much friction.

Physicians have EHRs.

Patients have portals.

Staff have payer websites.

Billing teams have clearinghouses.

Administrators have dashboards.

Executives have analytics.

Everyone has another login.

Everyone has another password.

Everyone has another notification.

Everyone has another queue.

And somehow, after all this technology, someone still ends up calling the insurance company to ask:

“Where is the claim?”

That should make us laugh.

Except it doesn't.

Because somebody is paying for those minutes.

Usually with time.

And in healthcare, time is not a trivial resource.

Time is clinical capacity.

Time is attention.

Time is listening.

Time is explaining.

Time is thinking.

Time is teaching.

Time is reassuring.

Time is human connection.

Sometimes, apparently, time is also five minutes with a cello.

The Cello Is the Point

It would be easy to read the story of Dr. Kalmansson and Leon Allan Loyd and conclude:

“Isn't that nice?”

It is nice.

But I think that misses the point.

The more interesting question is:

Why does this moment feel so extraordinary?

Why does a physician sitting with a frightened patient and playing music feel almost surprising?

Because modern healthcare has become remarkably good at processing people.

We are less good at being with people.

We have become very sophisticated at moving information.

We are less sophisticated at protecting attention.

We can transmit a claim electronically.

We can route an authorization.

We can generate a note.

We can calculate a risk score.

We can predict a denial.

We can automate a message.

But none of those things can sit beside a human being who is about to lose a limb and say:

I am here.

That still requires a human.

And that is precisely why administrative efficiency matters.

Not because we need physicians to see more patients.

Not because every minute must produce another billable unit.

Not because the healthcare system needs to squeeze another percentage point of productivity from exhausted professionals.

But because human attention is finite.

If we waste it on work that machines and better-designed systems could handle, we are making a choice.

We are choosing administration over attention.

Even when nobody intended to.

The Healthcare Industry Has a Strange Definition of Efficiency

Healthcare loves the word efficiency.

But what does efficiency actually mean?

For some organizations, efficiency means:

More patients.

More encounters.

More claims.

More revenue.

Fewer employees.

Shorter cycle times.

Higher collections.

Lower cost per encounter.

All of those metrics can matter.

But there is a dangerous assumption hiding underneath them.

The assumption is that the purpose of efficiency is to create more throughput.

I disagree.

The highest form of healthcare efficiency should create more capacity for care.

Those are not the same thing.

Imagine a clinic saves two hours per physician every week.

There are two ways to use those two hours.

Option A:

Book more patients.

Option B:

Give physicians more time to think, communicate, review complex cases, call patients, mentor staff, coordinate care, or simply finish the day without taking another pile of work home.

We tend to assume Option A is automatically better.

Why?

Because Option A is easier to measure.

Revenue is measurable.

Claims are measurable.

Visits are measurable.

Hours saved are measurable.

But human connection is harder to put on a dashboard.

That doesn't make it less valuable.

It makes it easier to ignore.

The Administrative Tax Nobody Sees

The American Academy of Family Physicians has described administrative burden as a major problem for physicians, noting that administrative tasks can consume approximately half of physicians' time in some settings.

The AMA likewise describes administrative burden as something that consumes physician time and focus, interrupts patient care, and contributes to burnout.

That should change how we talk about medical billing.

Billing is often treated as something that happens after medicine.

I think that is outdated.

The revenue cycle starts much earlier.

It starts with the information captured during the encounter.

It starts with documentation.

It starts with coding.

It starts with payer requirements.

It starts with whether the right information is available at the right time.

It starts upstream.

By the time a claim is denied, the problem may have already existed hours, days, or weeks earlier.

And then we call someone in the billing department and ask them to fix it.

That is like discovering a leak in the basement and congratulating ourselves because we bought a better mop.

The mop matters.

But perhaps we should also fix the pipe.

Three Experts, One Bigger Message

Several physician leaders and professional organizations have been making versions of this argument for years.

Their approaches differ.

Their specialties differ.

Their technologies differ.

But the direction is remarkably similar.

1. Dr. Paul Kalmansson: Medicine Is Still Human

Kalmansson's contribution to this conversation is not a policy paper.

It is a cello.

His story reminds us that medicine is not simply the delivery of clinical interventions.

There is also the experience of being a patient.

Fear matters.

Uncertainty matters.

Dignity matters.

Hope matters.

Presence matters.

Kalmansson explained that he wanted his patient to know he cared and that music allowed him to communicate that more effectively than words.

That is a powerful lesson for technology leaders.

Technology should protect the conditions in which physicians can be human.

It should not make physicians more efficient at becoming machines.

2. Makrina Shanbour, MD: Fix the System, Not the Physician

Physician well-being discussions increasingly recognize that burnout cannot simply be solved by telling physicians to become more resilient.

That is important.

Because if a workflow is dysfunctional, giving the physician a meditation app does not repair the workflow.

If the EHR is unnecessarily complicated, telling doctors to “practice self-care” does not simplify the EHR.

If prior authorization is consuming hours, yoga does not complete the authorization.

If claims repeatedly fail because information is missing upstream, motivational speeches do not fix the claim.

The system needs work.

The AMA has repeatedly emphasized organizational approaches to physician well-being and the need to remove obstacles that interfere with patient care.

That leads to a provocative principle:

Do not train physicians to tolerate broken systems. Fix the systems.

3. Steven Waldren, MD, MS: Technology Can Help — If We Use It Correctly

Dr. Steven Waldren, Chief Medical Informatics Officer at the American Academy of Family Physicians, has written extensively about administrative burden, documentation and the promise and limitations of AI in primary care.

His work reflects an important reality:

Technology can reduce burden.

But technology can also create burden.

AI is not automatically the answer simply because it has “AI” in the name.

A poorly designed AI workflow can become one more system clinicians have to monitor.

One more dashboard.

One more alert.

One more verification step.

One more thing to learn.

The objective should therefore not be:

“Where can we put AI?”

The better question is:

“Where is human time being wasted, and what is the safest way to remove that waste?”

That distinction matters.

The AI Industry Has a Problem

Here is another contrarian thought:

The best AI in healthcare may be the AI nobody notices.

We have become fascinated with spectacular demonstrations.

AI writes a note.

AI summarizes a chart.

AI generates an answer.

AI produces an image.

AI writes an email.

Fine.

But what if the most valuable healthcare AI is much less glamorous?

What if it quietly prevents a claim error before submission?

What if it identifies missing documentation before the billing team sees the claim?

What if it understands payer-specific rules?

What if it checks eligibility?

What if it recognizes that a seemingly minor documentation issue is likely to create a denial?

What if it helps structure information before the downstream mess happens?

Nobody will make a viral video about that.

There will be no robot dancing.

No cinematic soundtrack.

No dramatic product launch.

Just fewer problems.

And fewer problems are often more valuable than impressive demonstrations.

We Have Been Automating the Wrong End

Much of healthcare technology has historically focused on what happens after information is created.

The claim is generated.

Then someone checks it.

The claim is denied.

Then someone works the denial.

The authorization is rejected.

Then someone appeals it.

The chart is incomplete.

Then someone chases the physician.

The payer changes a rule.

Then someone discovers it.

The patient receives a bill.

Then someone calls.

This is downstream automation.

It can help.

But it is still reactive.

The bigger opportunity is upstream.

Prevent the problem before it becomes a workflow.

That means asking:

Was the right information captured?
Was it structured correctly?
Does it support the intended code?
Does it satisfy payer requirements?
Is something missing?
Is something contradictory?
Is this likely to trigger a denial?
Can the problem be corrected while the encounter is still fresh?

This is where clinical-to-claims intelligence becomes interesting.

The objective is not to turn physicians into coders.

Quite the opposite.

The objective is to keep physicians focused on medicine while the infrastructure quietly makes the administrative consequences of that medicine more reliable.

The Five-Minute Test

Here is a simple test I would give every healthcare technology company.

Ask this:

“What will the physician do with the five minutes your product saves?”

Not:

“How many clicks did you remove?”

Not:

“How many claims did you process?”

Not:

“What is the accuracy rate?”

Those metrics matter.

But ask the next question.

What happens to the five minutes?

If the answer is:

“Now the physician can see another patient,”

that's useful.

But it is not the only possible answer.

Maybe the physician calls a patient who received a frightening diagnosis.

Maybe they review a complicated case.

Maybe they explain a treatment plan properly.

Maybe they mentor a younger clinician.

Maybe they eat lunch.

Yes.

Lunch.

Healthcare has somehow reached a point where eating lunch can sound like a technology use case.

Maybe they go home fifteen minutes earlier.

Maybe they play with their child.

Maybe they sleep.

Maybe they simply stop working.

Those are not failures of productivity.

They are evidence that time was returned to a human being.

The Great Healthcare Productivity Trap

There is a dangerous cycle:

Technology saves time.
Organizations discover the time.
The organization fills the time with more work.
Productivity rises.
The time savings disappear.
Everyone asks for another technology solution.

Repeat.

This is the productivity treadmill.

We keep making people more efficient at doing more work.

At some point we should ask:

What if efficiency is supposed to make work more humane rather than merely more abundant?

That is a radical idea in modern healthcare.

But perhaps it should not be.

Medical Billing Is Not the Enemy

I want to be careful here.

Medical billing is not evil.

Revenue matters.

Claims matter.

Accurate coding matters.

Compliance matters.

Physician practices cannot survive if they do not get paid.

A clinic with beautiful patient relationships and terrible cash flow eventually becomes a clinic that closes.

That helps nobody.

So the answer is not:

“Forget billing and focus only on patients.”

That's romantic nonsense.

The answer is:

Build billing infrastructure that respects the clinical mission.

A physician should not have to choose between caring for patients and running a financially viable practice.

A clinic should be able to do both.

That requires better infrastructure.

Why Small and Independent Practices Matter

This conversation becomes particularly important for small and medium-sized physician practices.

Large health systems can have entire departments dedicated to revenue cycle management.

They can employ analysts.

They can negotiate contracts.

They can build internal teams.

They can absorb implementation costs.

A small physician practice has a different reality.

The office manager may be handling five different functions.

The physician may still be involved in billing decisions.

The staff may be switching between the EHR, payer portals, clearinghouse systems, spreadsheets, phone calls and email.

The technology stack becomes a patchwork.

One vendor handles one problem.

Another vendor handles another.

A third vendor promises to integrate them.

Eventually the practice needs a consultant to explain why the integrations don't integrate.

We laugh because it is absurd.

Then somebody gets another denial.

This Is Why I Believe the Future Is Upstream

My own interest in this problem led me to build OnnX around a simple idea:

Revenue-cycle intelligence should begin closer to the point where clinical information is created.

Not after the claim fails.

Not after the payer rejects it.

Not after the billing team spends an hour figuring out why.

Earlier.

Much earlier.

The vision is not “AI replaces the billing department.”

It is not “AI replaces physicians.”

It is not “AI replaces staff.”

That would miss the point.

The goal is to create infrastructure that helps humans make fewer preventable administrative mistakes.

The system should do the repetitive work.

Humans should handle judgment, exceptions, relationships and accountability.

That is a much more interesting future than simply replacing people.

Human-in-the-Loop Is Not a Weakness

There is a strange obsession in technology with eliminating humans from workflows.

I think that is backwards in healthcare.

Healthcare is full of ambiguity.

Clinical judgment is contextual.

Patient preferences matter.

Exceptions happen.

Payers behave inconsistently.

Documentation can be nuanced.

Rules change.

Data can be incomplete.

So the best system is often not:

Human → machine disappears

It is:

Human → machine assists → human verifies when needed

That is not failure.

That is good system design.

The machine handles scale.

The human handles judgment.

The machine remembers rules.

The human understands context.

The machine watches the queue.

The human decides what deserves attention.

That division of labor is much more realistic.

What We Should Stop Automating

Here is where I will be deliberately contrarian.

We should not automate something simply because it is technically possible.

Some things should remain deeply human.

1. Empathy

Don't automate the moment when someone is scared.

2. Difficult conversations

A chatbot should not become the default interface for every emotionally significant conversation.

3. Clinical judgment

Decision support can be valuable.

Decision abandonment is not.

4. Accountability

If an automated process causes a serious problem, someone must own the outcome.

5. Trust

Patients should understand when technology is involved in their care.

6. Exceptions

The strangest cases are often the ones that require human attention.

7. The physician-patient relationship

If technology gives physicians more time with patients, excellent.

If technology becomes another barrier between them, we have solved the wrong problem.

What We Should Automate Aggressively

Now the other side.

There are tasks where healthcare should be much more aggressive.

Automate repetitive work.

Automate data checking.

Automate routine status checks.

Automate predictable workflows.

Automate duplicate entry.

Automate claim-quality checks.

Automate administrative reminders.

Automate repetitive payer-rule matching where appropriate.

Automate the hunting.

Automate the copying.

Automate the reconciliation.

Automate the things humans hate doing and machines are good at doing.

Then give the human back the time.

That's the deal.

A Practical Five-Step Test for Clinic Owners

If I were running a physician practice today, I would not begin by asking:

“What AI should we buy?”

I would ask five questions.

Step 1: Find the Time Leaks

Track where physicians and staff spend time for one week.

Not estimates.

Actual time.

Phone calls.

Portal checks.

Claim follow-up.

Prior authorization.

Documentation.

Coding questions.

Denial work.

Eligibility.

Faxing.

Data entry.

Find the leaks.

Step 2: Separate Clinical Work From Administrative Work

Create two buckets.

Only a human should do this.

And:

A machine should probably do this.

You may be surprised how much work sits in the second bucket.

Step 3: Attack the Upstream Cause

Don't only count denials.

Ask why the denial happened.

Was documentation incomplete?

Was coding wrong?

Was eligibility incorrect?

Was the payer requirement misunderstood?

Was information entered inconsistently?

Was the claim scrubbed properly?

Was the problem detectable earlier?

Move upstream.

Step 4: Measure Time Returned

Do not measure only revenue.

Measure:

Physician minutes returned
Staff minutes returned
Denial rate
Clean-claim rate
Days in A/R
First-pass resolution
Administrative touches per claim
Manual payer interactions
Rework
After-hours administrative work

And then ask the most important metric:

What did people do with the time we returned?

Step 5: Protect the Savings

This is the step most organizations skip.

If automation saves five hours, don't automatically schedule five more hours of work.

Decide what the recovered capacity is for.

Patient communication.

Complex cases.

Staff development.

Quality improvement.

Professional development.

Or simply sustainability.

Otherwise, automation becomes a machine for creating more work.

The ROI We Don't Put on the Spreadsheet

Suppose an administrative improvement saves a physician one hour a week.

The obvious calculation is financial.

But there are other returns.

One hour may mean:

One difficult patient conversation.

One family phone call.

One chart reviewed more carefully.

One trainee taught.

One staff member supported.

One less late night.

One less weekend spent catching up.

One additional moment of presence.

Not everything valuable in healthcare appears as revenue.

Some things appear as capacity.

Some appear as trust.

Some appear as better decisions.

Some appear as less exhaustion.

And some appear as a physician sitting beside a frightened patient with a cello.

The Most Dangerous Word in Healthcare Technology

The word is:

More.

More patients.

More claims.

More automation.

More dashboards.

More data.

More alerts.

More productivity.

More throughput.

More revenue.

More features.

More integrations.

More AI.

More.

Perhaps the healthcare technology industry should become obsessed with another word:

Enough.

Enough alerts.

Enough clicks.

Enough duplication.

Enough portals.

Enough manual reconciliation.

Enough administrative noise.

Enough work that nobody should have been doing manually in the first place.

Because sometimes the greatest technology achievement is not producing more.

It is eliminating what never needed to happen.

A Myth We Need to Kill

Myth:

“If AI makes physicians more productive, healthcare automatically improves.”

No.

Not automatically.

Productivity without purpose can make a bad system worse.

If AI helps a physician complete documentation faster and the organization immediately fills the recovered time with more appointments, the patient may not experience the benefit.

If AI helps billing process twice as many claims but creates another verification queue, staff may not experience the benefit.

If an AI tool saves ten minutes but requires twenty minutes of setup, training and monitoring, congratulations:

You invented a new job.

Technology should be judged by net burden, not theoretical efficiency.

Another Uncomfortable Question

Why do we celebrate technology that saves a physician ten minutes...

and then use those ten minutes to schedule another patient?

Maybe that's appropriate.

Sometimes it is.

But sometimes the better question is:

What would happen if we used those ten minutes to make the existing patient experience better?

Healthcare has spent decades optimizing volume.

Perhaps the next era should optimize attention.

The Cello and the Claim

This brings us back to Leon Allan Loyd.

Imagine two versions of healthcare.

In Version One, Dr. Kalmansson finishes his administrative work late.

He has charts.

Messages.

Documentation.

Billing questions.

Administrative interruptions.

He is exhausted.

There is no cello.

There is no extra time.

There is no special moment.

In Version Two, the infrastructure around him works better.

The repetitive administrative work is reduced.

The physician has more capacity.

And one day, instead of clicking through another queue, he sits down with his patient.

He plays Bach.

The disease is still there.

The operation is still coming.

The medicine has not changed.

But the experience of medicine has changed.

That difference matters.

Recent Healthcare News Is Telling Us Something

The broader healthcare conversation is moving in the same direction.

The AMA continues to focus on reducing administrative burden and identifying workflow and technology barriers that prevent physicians from spending more time with patients.

The AAFP has similarly emphasized administrative burden as a major threat to the physician experience and has highlighted technologies that can reduce burdens such as documentation and prior authorization.

And current physician discussions around AI increasingly recognize that its value is not simply generating impressive outputs. The real opportunity is reducing work that consumes professional time without adding equivalent clinical value.

The trend is clear.

The question is whether we will use technology to give time back or simply to fill every empty minute with more work.

Three Things I Would Change Tomorrow

If I had the authority to redesign a physician practice tomorrow, I would start with three rules.

Rule #1: Protect Physician Attention

Do not allow low-value administrative interruptions to compete constantly with clinical work.

Attention is a scarce resource.

Treat it like one.

Rule #2: Move Problems Upstream

Don't celebrate fixing a denial.

Celebrate preventing the denial.

Don't celebrate correcting missing information.

Celebrate capturing it correctly the first time.

Don't celebrate finding a billing error.

Celebrate never creating it.

Rule #3: Measure Human Capacity

Revenue matters.

Productivity matters.

But measure physician time and staff time too.

Because a healthcare system that collects more money while consuming every remaining minute of its people may be financially efficient and humanly bankrupt.

The Legal and Ethical Question

There is also an important compliance issue here.

Automation does not eliminate responsibility.

Healthcare organizations still need appropriate controls around:

HIPAA and privacy.

Data security.

Access controls.

Auditability.

Coding compliance.

Documentation integrity.

Payer requirements.

Human oversight.

Transparency.

Vendor accountability.

A system that automatically makes a decision should not become a black box that nobody understands.

The more consequential the decision, the more important explainability and oversight become.

The goal is not:

“The AI did it.”

The goal is:

“The system assisted, the workflow was auditable, and the appropriate human remained accountable.”

That distinction will become increasingly important as AI moves deeper into healthcare operations.

The Five Questions Every Healthcare Founder Should Ask

If you are building healthcare technology, ask yourself:

1. What human problem am I actually solving?

Not what feature are you building.

What problem?

2. Am I removing work or moving work?

This is a killer question.

If you eliminate one task but create three verification tasks, you haven't eliminated burden.

You've redistributed it.

3. Does my technology protect attention?

If not, why not?

4. Can the user understand why the system made a recommendation?

If not, what happens when it is wrong?

5. What does the human do with the time I give back?

If you cannot answer that question, you may be optimizing the wrong thing.

The Future of Healthcare May Be Less About AI Than We Think

The future will certainly contain more AI.

But I don't think the defining question will be:

“How intelligent is the AI?”

It may be:

“How intelligently did we redesign the work?”

AI inside a broken workflow can simply create a faster broken workflow.

AI inside a well-designed workflow can remove friction.

That is a much bigger opportunity.

The winners may not be the companies with the most impressive demonstrations.

They may be the companies that quietly remove thousands of tiny frustrations every day.

No applause.

No headlines.

No futuristic robot.

Just fewer interruptions.

Fewer errors.

Fewer denials.

Fewer clicks.

Fewer calls.

Fewer unnecessary handoffs.

More attention.

More capacity.

More human work.

And That Is Why I Keep Thinking About the Cello

Dr. Paul Kalmansson could not save Leon Allan Loyd from the reality he was facing.

But he could change how Leon experienced that reality.

That distinction is enormous.

Medicine cannot always change the outcome.

But physicians can influence the experience.

And technology should help them do that.

Not by pretending machines can replace compassion.

Not by turning every human interaction into a workflow.

Not by measuring every minute solely by revenue.

But by removing unnecessary work around the people doing necessary work.

That is the promise I see in healthcare technology.

Not a hospital full of robots.

Not physicians replaced by algorithms.

Not endless automation.

Something much more practical.

A physician with enough time to be a physician.

My Hot Take

We have spent years asking:

“How can we make doctors more productive?”

I think we should ask:

“How can we make doctors less interrupted?”

Those are not the same thing.

A more productive physician can see more patients.

A less interrupted physician may listen better.

Think more clearly.

Notice something subtle.

Explain something better.

Call someone back.

Teach.

Rest.

Or sit down and play music for a patient who is afraid.

Maybe that is the future we actually want.

What I Got Wrong About Efficiency

For years, healthcare has trained us to think about efficiency as doing something faster.

But speed is only one dimension of efficiency.

The deeper question is whether the work should exist at all.

A five-minute task that should never have existed is not efficient just because we completed it in four minutes.

That is not efficiency.

That is optimized waste.

And healthcare has plenty of it.

The future belongs to organizations willing to ask:

What should disappear?

The OnnX Thesis

This is ultimately what drives my work with OnnX.

I don't believe physicians need another piece of software demanding their attention.

They need infrastructure that quietly removes friction.

I don't believe AI should replace the people who understand patients.

I believe AI should help remove repetitive administrative work that keeps those people away from patients.

I don't believe the objective of medical billing technology should be to make healthcare feel more transactional.

I believe it should help make the administrative side of medicine less visible to the people who are trying to practice medicine.

That means moving from:

reactive → predictive

downstream → upstream

manual → intelligent

fragmented → connected

transactional → contextual

more work → more capacity

That is a different vision of healthcare technology.

And I think it is worth pursuing.

A 30-Day Experiment for Any Practice

You don't need to buy a new AI system tomorrow.

Start with observation.

For 30 days, track five things:

1. Administrative interruptions per physician

2. Manual billing touches per claim

3. Preventable denials

4. Time spent on payer-related work

5. After-hours administrative work

Then ask:

What are the top three sources of wasted human attention?

Don't start with technology.

Start with the waste.

Then decide whether technology, process redesign, staffing, training or policy change is the best solution.

Sometimes the answer will be AI.

Sometimes it won't.

That's okay.

Good healthcare innovation is not about selling AI.

It is about solving problems.

Metrics That Actually Matter

Healthcare leaders should continue tracking traditional financial metrics.

But add human-capacity metrics.

Consider measuring:

Physician administrative minutes
Staff administrative minutes
Denial rate
Clean-claim rate
Days in accounts receivable
Claim rework
Number of manual touches
Payer portal interactions
Prior-authorization turnaround
After-hours work
Physician satisfaction
Staff satisfaction
Patient communication time

The goal is not to make every metric rise.

The goal is to understand where the system is consuming human capacity.

Questions Worth Asking Your Team

At your next practice meeting, don't ask only:

“Are we collecting enough?”

Ask:

What administrative task annoys everyone?

What task do we repeat every day?

What information do we enter twice?

What do we manually check that should be automatically checked?

What causes the most rework?

What creates the most interruptions?

What keeps physicians after hours?

What could we prevent rather than repair?

And finally:

If we gave everyone five hours back next month, what would we want them to do with it?

That answer tells you what your technology strategy should actually be.

The Bigger Lesson

The story of Leon Allan Loyd and Dr. Paul Kalmansson is emotionally powerful because it contains something technology cannot manufacture on command.

Presence.

A physician saw a human being.

Not a case.

Not a diagnosis.

Not a claim.

Not an encounter number.

A person.

And for a few minutes, he gave that person something medicine sometimes struggles to find:

time.

That is why the story matters.

Because healthcare technology should ultimately be judged by what happens to the human beings on the other side of the screen.

If technology gives physicians more time to care, it is doing something valuable.

If it gives staff more capacity to solve problems, it is doing something valuable.

If it reduces errors before they become expensive downstream problems, it is doing something valuable.

If it allows a frightened patient to receive more attention from the person caring for them, it is doing something valuable.

But if technology merely creates more screens, more alerts, more workflows, more verification and more work...

we shouldn't call that innovation.

We should call it administrative inflation.

Three Things I Want Healthcare Leaders to Remember

Protect physician attention.

Move administrative problems upstream.

Measure success by the human capacity your technology creates—not simply the transactions it processes.

Get Involved

Here is the uncomfortable question I want to leave with physicians, clinic owners, healthcare executives and healthcare technology builders:

If we could give every physician five additional minutes with every patient, what would they do with those five minutes?

Would they listen?

Explain?

Think?

Teach?

Reassure?

Call a family member?

Or perhaps, once in a while, do something completely unexpected.

Like play Bach.

I would genuinely like to hear your answer.

Comment below: What is the single administrative task that steals the most valuable time from you or your clinical team?

And if this perspective resonates with you, share or repost it.

The conversation about healthcare technology should not be limited to what AI can do.

It should include what humans can finally do when unnecessary work gets out of their way.

Final Thoughts

Technology should not make physicians better machines. It should give physicians more room to be human.

The best administrative system is not the one that processes the most work. It is the one that prevents unnecessary work from reaching the people who should be caring for patients.

The ultimate ROI of healthcare technology may not be another claim processed. It may be another human moment made possible.

Frequently Asked Questions

Is this article arguing against medical billing technology?

No.

Quite the opposite.

Accurate, efficient billing is essential to practice sustainability. The argument is that billing technology should be designed around the clinical mission rather than operate as an isolated financial function.

Is AI going to replace medical billing staff?

Some repetitive tasks will increasingly be automated.

That does not necessarily mean humans disappear.

The more likely model is that technology handles repetitive work while people handle exceptions, judgment, communication, compliance and accountability.

Why connect a cello performance to medical billing?

Because both involve the same scarce resource:

human time.

The cello story demonstrates what can happen when a physician has the time and inclination to provide meaningful human presence.

Administrative inefficiency can reduce that available capacity.

Isn't physician productivity important?

Absolutely.

But productivity should not automatically mean seeing more patients.

It can also mean spending more attention on complex patients, improving communication, reducing errors, mentoring staff or sustaining a healthier clinical practice.

What is the upstream revenue-cycle problem?

Many downstream billing problems originate earlier in the workflow.

Documentation, coding, eligibility, payer requirements and data quality can influence what happens when a claim is eventually submitted.

The earlier a problem can be identified, the less expensive and disruptive it may be to correct.

Should every medical practice adopt AI?

No.

A practice should adopt technology when it solves a clearly identified problem better than the alternatives.

Sometimes that means AI.

Sometimes it means workflow redesign.

Sometimes it means better training.

Sometimes it means hiring the right person.

Technology should serve the problem, not the other way around.

What should physicians look for in healthcare AI?

Look beyond the demo.

Ask:

Does it reduce net workload?

Does it integrate into the existing workflow?

Can users understand its recommendations?

Does it preserve human oversight?

Is patient data protected?

Can the organization audit what happened?

Does it actually save time?

And perhaps the most important question:

What will the physician do with the time saved?

Three References From This Week

1. Dr. Paul Kalmansson and Leon Allan Loyd: The Cello Before Amputation

ABC7 Los Angeles reported on September 2, 2026, how Dr. Paul Kalmansson played Bach on the cello for U.S. Army veteran Leon Allan Loyd before Loyd underwent amputation surgery, giving him strength and hope during an extraordinarily difficult moment.

ABC7 Los Angeles — Loma Linda doctor's cello performance gives veteran patient hope

2. AMA: Reducing Administrative Burden

The American Medical Association continues to emphasize that administrative burden consumes physician time and focus and can interfere with patient care, reinforcing the need to redesign workflows rather than simply ask physicians to work harder.

AMA — Reducing Administrative Burden

3. AAFP: Relieving Administrative Burden

The American Academy of Family Physicians identifies administrative work as a major burden on physicians and highlights technology and workflow innovations designed to reduce documentation, prior authorization and other administrative demands.

AAFP — A Guide to Relieving Administrative Burden

Tools and Resources

For practice leaders evaluating administrative technology, start with the fundamentals:

Workflow mapping — identify where information enters, moves and breaks.

Denial analysis — determine the actual causes of denials rather than simply counting them.

Time tracking — measure administrative minutes rather than relying on assumptions.

Payer-rule monitoring — identify changes before they create downstream problems.

Documentation review — find recurring information gaps.

Human-in-the-loop AI — automate repetitive work while preserving appropriate human oversight.

Audit trails — make important automated decisions traceable.

Security controls — protect patient information throughout the workflow.

The Future Outlook

Healthcare is not heading toward a world where machines simply replace people.

It is heading toward a much messier and more interesting world.

Humans and machines will work together.

The challenge will be deciding which work belongs to which.

Machines are exceptionally good at repetition.

Humans are exceptionally good at context.

Machines can monitor thousands of transactions.

Humans can recognize when something doesn't feel right.

Machines can remember rules.

Humans understand relationships.

Machines can identify patterns.

Humans decide what those patterns mean in context.

The future belongs to systems that understand this division of labor.

And the best healthcare technology may ultimately be almost invisible.

Patients may never know it exists.

Physicians may barely notice it.

Staff may simply wonder why the day feels less chaotic.

Claims may move more cleanly.

Errors may be caught earlier.

Denials may decline.

Workflows may become quieter.

And somewhere in that quieter system, a physician may have enough time to sit beside a patient.

Maybe talk.

Maybe listen.

Maybe hold a hand.

Maybe play a cello.

That is not a productivity failure.

That is the point.

About the Author

Dr. Daniel Cham is a physician and medical consultant whose work spans medical technology consulting, healthcare management and medical billing. His focus is practical: helping healthcare professionals and organizations navigate complex challenges at the intersection of medicine, technology, operations and the business of practice.

Connect with Dr. Cham on LinkedIn:

Dr. Daniel Cham on LinkedIn

This article represents the author's professional perspective and is intended for educational and informational purposes. It does not constitute legal, medical, coding, billing or compliance advice. Healthcare organizations should obtain appropriate professional guidance for decisions involving their specific circumstances.

Continue the Conversation

More healthcare insights, commentary and practical ideas:

Website: Dr. Daniel Cham

Spotify: Dr. Daniel Cham on Spotify

YouTube: Dr. Daniel Cham on YouTube

X: Dr. Daniel Cham on X

Facebook: Dr. Daniel Cham on Facebook

Knowledge drives progress. Start your journey here.

Free Resource

Check the Featured section of my LinkedIn profile for a free download.

No signup required.

One Last Thought

If this perspective resonates with you, consider reposting it.

Not to promote a product.

Not to promote a technology.

But to help other physicians, clinic owners and healthcare leaders rethink a deceptively simple question:

What would healthcare look like if we stopped asking humans to do work machines could safely do—and started giving humans back the time to do what only humans can do?

Because Leon Allan Loyd did not need another transaction.

He needed a physician.

And for a few minutes, Dr. Paul Kalmansson gave him exactly that.

#Healthcare #PhysicianLeadership #MedicalBilling #HealthcareAI #RevenueCycleManagement #PhysicianBurnout #AdministrativeBurden #HealthcareTechnology #DigitalHealth #PatientCare #MedicalPractice #PhysicianWellbeing #AIinHealthcare #HealthcareInnovation #IndependentPractice #OnnX