A toddler’s kidney transplant reveals a bigger healthcare problem: we have become remarkably good at moving patients, data and organs—but surprisingly bad at removing unnecessary work from the people caring for them.
“Sometimes we put handcuffs on ourselves.” — Donald M. Berwick, MD, MPP, healthcare quality expert, and health policy leader.
John Miller Got a Kidney. Physicians Got Another Portal. Something Is Wrong.
John Miller was two years old when his family learned that another family’s tragedy might give him a second chance at life.
John had been seriously ill since infancy.
His mother, Greer Miller, had watched her son struggle with severe kidney disease. His father, Kris Miller, watched his little boy spend far too much of his early childhood in hospitals. His older sister, Charlotte Miller, experienced the disruption that illness can bring to an entire family.
Then John needed a kidney transplant.
In May 2026, his cousin, Lance Miller, died in a motorcycle accident.
Lance’s family was suddenly dealing with an unbearable loss.
Yet in the middle of that grief, they chose organ donation.
Then came the extraordinary connection.
Lance’s kidney was an exceptionally strong match for John.
John’s father, Kris Miller, and Lance’s father were twin brothers.
On June 3, 2026, John received the transplant at Children’s Healthcare of Atlanta.
The operation succeeded.
John had more energy.
He could play.
He could act like a two-year-old.
The hospital also reached an institutional milestone: John’s transplant was its 2,000th pediatric solid-organ transplant.
That number is impressive.
But it is not the story.
The story is a little boy who got more childhood.
And that is where I want to take this conversation somewhere slightly uncomfortable.
Because healthcare has become remarkably good at moving complicated things.
Organs.
Patients.
Images.
Lab results.
Medications.
Millions of pieces of clinical data.
And yet somehow we still struggle to move a simple piece of information from one administrative workflow to another without asking someone to log into a different portal.
We can transplant a kidney into a toddler.
But we still make physicians chase paperwork.
Maybe we should talk about that.
The Healthcare Machine Has a Strange Sense of Humor
I am a physician.
I am also building a healthcare technology company.
So I have a front-row seat to one of healthcare’s strangest contradictions.
Medicine has become extraordinarily sophisticated.
We can perform robotic surgery.
Sequence genomes.
Replace organs.
Treat diseases that were once considered hopeless.
Monitor patients remotely.
Use advanced imaging to see inside the human body.
And then we ask a physician to spend 20 minutes looking for a fax.
There is something almost poetic about it.
Except nobody is laughing.
Healthcare’s technological progress and administrative progress have not moved at the same speed.
In some places, the clinical side looks like the future.
The administrative side looks like it missed the train.
And then another train.
And possibly lunch.
The Problem Is Not That Physicians Hate Technology
Physicians are often described as resistant to technology.
I don’t buy it.
Physicians use incredibly sophisticated technology every day.
The problem is bad workflow disguised as technology.
Nobody wakes up and says:
“I wish I had another login today.”
Nobody says:
“I hope this afternoon contains three payer portals.”
Nobody went to medical school dreaming about reconciling spreadsheets.
And yet these tasks have become part of the modern physician experience.
We keep telling clinicians that the answer is another tool.
Another dashboard.
Another application.
Another AI assistant.
Another notification.
Another inbox.
At some point, we have to ask:
What if the problem isn’t that physicians need more technology?
What if they need less work?
The Contrarian Idea
Here is my contrarian view:
Healthcare may not have an automation problem. It may have an elimination problem.
We keep asking:
What can AI automate?
I think the first question should be:
Why are we doing this at all?
That distinction changes everything.
Suppose a billing employee spends 15 minutes copying information between systems.
The obvious technology question is:
“How can AI copy it faster?”
The better operational question is:
“Why does the information have to be copied?”
Those are completely different questions.
One produces automation.
The other produces redesign.
And redesign usually wins.
John Miller’s Story Gives Us a Useful Test
Think about John’s transplant.
Nobody would have looked at the transplant process and said:
“Let’s add six unnecessary steps and see how that goes.”
Nobody would have said:
“Before the surgeon begins, let’s make the family call three different departments.”
Nobody would say:
“The kidney information exists in the system, but the surgeon needs to manually re-enter it.”
Why?
Because the stakes are obvious.
The system is designed around the human outcome.
Save the child.
But administrative healthcare often gets designed around something else.
The department.
The software.
The payer.
The form.
The workflow.
The policy.
The process.
The organization chart.
And eventually the patient becomes the thing we work around.
That is backwards.
The Patient Never Asked for Our Organizational Chart
Patients do not care which department owns the problem.
They do not care whether a claim is sitting in:
Billing.
Coding.
Authorization.
Credentialing.
Registration.
The clearinghouse.
The payer.
The EHR.
They simply know something isn’t working.
And physicians experience the same problem from the other side.
A physician may know exactly what happened clinically.
But the administrative system may ask for the information again.
And again.
And again.
Healthcare has developed an unusual relationship with repetition.
We call it documentation.
We call it compliance.
We call it verification.
We call it reconciliation.
Sometimes those things are necessary.
Sometimes they are.
But sometimes repetition is simply repetition.
Not every administrative task is sacred.
Some are just old.
Recent News Is Telling Us Something
The current healthcare environment makes this problem increasingly difficult to ignore.
Recent MGMA reporting found that 44% of medical-group leaders said prior-authorization turnaround became slower in 2026 compared with the previous year.
That is remarkable.
Especially because healthcare has spent years investing in electronic processes designed to make these workflows faster.
Technology does not automatically eliminate friction.
Sometimes it just gives friction a login page.
The AMA continues to report substantial administrative pressure from prior authorization, including roughly 40 requests per physician per week.
Approximately 94% of physicians report that prior authorization contributes to burnout.
And approximately 95% say it can delay necessary care.
Those numbers should make us uncomfortable.
Not because prior authorization is inherently bad.
Clinical review can have legitimate purposes.
The problem is when the administrative process becomes so cumbersome that it consumes the very people who are supposed to deliver care.
Three Expert Lessons
1. The AMA’s Practice-Sustainability Lesson
AMA guidance for private practices emphasizes a problem many small practices know intimately:
A practice can become dangerously dependent on one or two people who understand the billing and insurance process.
That is fragile.
If the person leaves, gets sick, retires or simply takes a vacation, institutional knowledge can disappear.
The lesson:
Knowledge trapped inside one employee is not a system.
It is a vulnerability.
Physician owners should document workflows.
Measure performance.
Standardize recurring processes.
And build systems that make institutional knowledge visible.
2. The Denial Lesson
Experts in revenue-cycle management repeatedly emphasize that practices need to understand why claims are denied.
This sounds obvious.
Yet many practices still spend enormous amounts of time reacting to denials instead of preventing them.
Think about that.
We have built an entire industry around fixing problems after they happen.
There is something strangely comfortable about that.
A denial creates work.
The work creates reports.
The reports create meetings.
The meetings create action items.
And eventually everyone feels productive.
Except the denial is still happening.
Activity is not the same thing as improvement.
3. The Patient-Access Lesson
The AMA has repeatedly highlighted how administrative barriers can interfere with patient access to care.
That matters because administrative work does not stay administrative forever.
A delayed authorization can delay treatment.
A billing problem can confuse a patient.
A repeated documentation request can consume clinical time.
A staffing shortage can slow operations.
A financially stressed practice can reduce capacity.
The administrative system is connected to the patient experience.
We should stop pretending otherwise.
The Medical Billing Paradox
Here is the paradox.
Physicians are expected to provide excellent clinical care.
Practice owners are expected to operate financially sustainable businesses.
Staff are expected to handle growing administrative complexity.
Payers are expected to control costs.
Patients are expected to navigate the system.
And technology companies are expected to somehow make everyone happy.
Good luck.
The result is a system where everyone is busy.
But being busy does not mean the system is working.
A practice can have:
More staff.
More software.
More reports.
More dashboards.
More automation.
And still have a broken revenue cycle.
Why?
Because the system may be optimizing activity instead of outcomes.
The First Question Should Be: Where Does the Money Leak?
Physician owners should know three things.
Where are we losing legitimate revenue?
Why are we losing it?
Can we prevent it from happening again?
That third question is the most important.
If the same denial happens 200 times, you don’t have 200 billing problems.
You have one workflow problem occurring 200 times.
That distinction can save enormous amounts of time.
Stop Treating Every Denial as an Individual Event
Imagine your practice has 1,000 claims.
One hundred are denied.
Your billing team works those 100 claims.
That sounds reasonable.
But suppose 60 of those denials have the same underlying cause.
Now you have an opportunity.
Instead of asking:
“How do we work 60 claims faster?”
Ask:
“How do we stop creating those 60 claims?”
That is where AI becomes interesting.
Not because AI can type faster.
Because AI can potentially recognize patterns humans may not notice across thousands of transactions.
Where AI Can Actually Help
I believe AI has a meaningful role in medical billing.
But not the role that marketing often promises.
AI should not be the new employee that magically does everything.
It should be an intelligent layer that helps people see, prioritize and act.
For example:
Before submission: identify potential inconsistencies.
After submission: monitor payer responses.
When denied: classify the denial.
When information is needed: locate relevant documentation.
When an appeal is appropriate: prepare a draft.
When a human decision is required: escalate.
After resolution: learn from the outcome.
That is a workflow.
Not a chatbot.
The AI Rule I Would Put on Every Billing System
Here is my rule:
If the system is uncertain, it should say so.
That sounds basic.
It isn’t.
AI systems can produce very confident answers.
Healthcare does not need confident nonsense.
A responsible system should distinguish between:
Known.
Likely.
Uncertain.
Needs human review.
That is especially important when claims, coding, documentation and reimbursement are involved.
The objective is not to make AI look intelligent.
The objective is to make the workflow safer.
Human-in-the-Loop Is Not a Failure
There is a strange obsession in technology with eliminating humans.
I think that is the wrong goal.
Suppose AI identifies a likely denial.
It finds the relevant documentation.
It explains why the claim may have failed.
It drafts a response.
Then a trained person reviews it.
That is not an inferior system.
That is a responsible system.
The machine handles the repetitive work.
The human handles judgment.
Human oversight is not a weakness.
It is a safety feature.
What AI Should Never Do
AI should not:
- Invent documentation.
- Invent diagnoses.
- Change clinical facts.
- Create unsupported medical necessity.
- Manipulate coding simply to increase reimbursement.
- Hide uncertainty.
- Submit questionable claims without review.
- Replace clinical judgment.
- Override compliance controls.
- Make accountability disappear.
If the technology does those things, you do not have innovation.
You have a liability generator with a user interface.
That is not the future healthcare needs.
Myth: “We Just Need More Billing Staff”
Maybe.
But first diagnose the workflow.
If five employees spend their days correcting the same mistake, hiring a sixth employee does not solve the mistake.
It increases payroll.
If the problem is information flow, hire less slowly and redesign faster.
If the problem is payer complexity, understand the payer pattern.
If the problem is documentation, fix documentation.
If the problem is eligibility, move verification upstream.
More people can increase capacity.
They cannot automatically fix bad architecture.
Myth: “Electronic Means Efficient”
No.
Electronic means electronic.
That’s it.
A fax converted into a PDF is still a fax problem wearing nicer clothes.
A manual workflow inside a web portal is still manual.
A spreadsheet uploaded into another spreadsheet is still a spreadsheet.
Digital transformation becomes meaningful only when the workflow itself improves.
Myth: “AI Means Fully Automated”
Also no.
The goal is not 100% automation.
The goal is 100% appropriate attention.
Routine work should move quickly.
Exceptions should receive attention.
High-risk decisions should receive human review.
Important cases should not disappear into an algorithmic black box.
That is a much better definition of intelligent automation.
The Five-Step OnnX Philosophy
This is the thinking behind the direction I am pursuing with OnnX.
1. Eliminate
Ask whether the task needs to exist.
2. Simplify
If it must exist, remove unnecessary steps.
3. Standardize
Create predictable workflows.
4. Automate
Use technology for repetitive, rules-based or pattern-heavy work.
5. Escalate
Move exceptions and high-risk decisions to humans.
That order matters.
Most technology projects start at number four.
I think they should start at number one.
A Practical Exercise for Your Practice
Take one claim.
Not a dashboard.
Not a monthly report.
One claim.
Follow it from:
Patient → registration → eligibility → documentation → coding → claim → payer → adjudication → payment
Write down every handoff.
Every login.
Every manual entry.
Every phone call.
Every delay.
Every time someone says:
“Let me check.”
Then ask:
Why?
You will probably find something interesting.
Your Seven-Day Billing Reality Check
Day 1: Find your biggest denial category.
Do not guess.
Measure it.
Day 2: Find your slowest payer.
Measure staff time, not just payment time.
Day 3: Find your oldest unresolved claims.
Ask why they are still unresolved.
Day 4: Find the most repetitive task.
Then ask why a person is still doing it.
Day 5: Find the most fragile workflow.
Which process collapses if one employee is absent?
Day 6: Find one task suitable for automation.
Choose something low-risk and repetitive.
Day 7: Establish a baseline.
Measure before changing.
Then measure again.
The Metrics I Would Watch
Forget vanity metrics.
Start with useful ones.
Clean-claim rate
How many claims move through without correction?
Denial rate
How many fail?
Repeat-denial rate
How often does the same problem recur?
Days in A/R
How long does legitimate revenue remain outstanding?
A/R over 90 days
How much money is becoming increasingly difficult to recover?
Appeal success rate
Are appeals working?
Time to resolution
How quickly does a problem become someone’s responsibility?
Staff time per claim
How much human labor does the workflow consume?
Preventable denial rate
How much failure could have been prevented upstream?
That last metric deserves far more attention.
The Hidden Metric: Physician Attention
There is another metric we rarely put on the dashboard.
Physician attention.
How many minutes per week does a physician spend on work that does not require a physician?
That is a real cost.
A physician’s attention is not free.
It represents years of education, training and experience.
If technology can save a physician 30 minutes of low-value administrative work every day, that is not merely a productivity improvement.
It may be the difference between finishing the day exhausted and finishing it with enough energy to think.
Sometimes the best healthcare technology is the technology that gives someone back 20 minutes.
The Patient Does Not See the Denial Queue
John Miller’s family probably did not care about the internal workflow architecture surrounding his transplant.
They cared about whether John would live.
Whether he would have energy.
Whether he could grow up.
That is how patients experience healthcare.
They do not see the system.
They experience the consequences of the system.
That is why physicians should care about administrative design.
Not because they want to become billing experts.
Because bad administrative design eventually becomes someone’s healthcare experience.
Legal and Compliance Considerations
Healthcare automation comes with responsibility.
Medical billing involves patient information, coding, reimbursement, documentation and payer requirements.
That creates potential exposure involving:
- Privacy.
- Security.
- False claims.
- Improper coding.
- Documentation.
- Medical necessity.
- Auditability.
- Access controls.
- Data retention.
- Human oversight.
Technology does not eliminate legal responsibility.
If anything, automation makes governance more important.
Every organization adopting AI should be able to answer:
What does the system do?
What information does it use?
Who reviews its output?
What happens when it is wrong?
Can we reconstruct what happened?
Who is accountable?
If nobody can answer those questions, the system should not be operating independently.
Ethical Considerations
There is also a moral question.
Healthcare technology should not make it easier to do the wrong thing at scale.
AI can make a good workflow faster.
It can also make a bad workflow faster.
That means ethics has to be part of product design.
The rule should be:
Technology can organize facts. It should never manufacture facts.
It can identify a missing document.
It should not create one.
It can flag a coding concern.
It should not manipulate a code to increase reimbursement.
It can prepare an appeal.
It should not invent clinical justification.
The difference between assistance and deception matters.
What Healthcare Founders Should Stop Saying
I would like healthcare technology marketing to retire a few phrases.
“Fully autonomous.”
“Zero-touch.”
“Replaces your billing team.”
“Never make another denial.”
“AI handles everything.”
They sound impressive.
They also create unrealistic expectations.
Healthcare is too complicated for magic.
The better pitch is often less exciting:
We help your team identify problems earlier.
We reduce repetitive work.
We prioritize the cases that need attention.
We keep humans accountable for important decisions.
That may not sound futuristic.
It sounds useful.
And useful wins.
What Physicians Should Stop Accepting
Physicians should also challenge a few assumptions.
“We’ve always done it this way.”
“That is the payer’s process.”
“Billing handles it.”
“Someone checks that manually.”
“We don’t have enough staff.”
“We just need another person.”
“We need another system.”
Maybe.
But maybe not.
The better question is:
Can we redesign this?
That question creates options.
The Bigger Opportunity for OnnX
The opportunity I see is not simply building another medical billing platform.
It is building a different relationship between physicians and the administrative machinery surrounding their practice.
Small and medium-sized clinics should not need enormous departments to manage increasingly complex workflows.
They should have access to intelligent infrastructure.
That infrastructure should help them understand what is happening.
Why it is happening.
What needs attention.
What can be prevented.
What can be automated.
And what should remain human.
That is the direction behind OnnX.
Not:
“AI replaces people.”
But:
“AI removes work people should not have to do.”
There is a meaningful difference.
The Future of Medical Billing May Be Boring
And I mean that as a compliment.
Imagine a future where:
A claim is checked before submission.
A likely problem is flagged.
Relevant information is automatically gathered.
Routine claims move through.
Exceptions rise to the top.
Denials are categorized automatically.
Recurring patterns are detected.
Appeals are prepared.
Humans review what actually requires judgment.
Payments are reconciled.
And the physician barely notices any of it.
That is not flashy.
There is no futuristic robot standing behind the front desk.
Nobody needs a headset.
Nobody needs to ask ChatGPT to explain a claim.
The workflow simply works.
That would be progress.
The Best AI May Be the AI You Barely Notice
Healthcare technology has an unusual problem.
We often celebrate visible technology.
The dashboard.
The chatbot.
The robot.
The interface.
The feature.
But some of the most valuable technology will be invisible.
A system catches an error before it becomes a denial.
Nobody notices.
A system routes a claim to the right person.
Nobody notices.
A system finds the missing documentation.
Nobody notices.
A system prevents a repetitive task.
Nobody notices.
That is okay.
The goal is not applause.
The goal is fewer problems.
What John Miller Teaches Us
John Miller’s story began with illness.
It moved through tragedy.
Then generosity.
Then medicine.
Then recovery.
At every point, information and people had to connect.
The system had to work.
That is the lesson I want to carry into medical billing.
Healthcare is ultimately a coordination business.
Patients coordinate with physicians.
Physicians coordinate with nurses.
Nurses coordinate with specialists.
Hospitals coordinate with families.
Billing teams coordinate with payers.
Everyone coordinates with everyone.
Every unnecessary handoff creates friction.
Every unnecessary handoff creates another opportunity for failure.
So perhaps the next great healthcare innovation is not another layer.
Perhaps it is removing layers.
Final Thoughts: Maybe Healthcare Needs Fewer Things
John Miller does not need another portal.
He needs to grow up.
He needs to play.
He needs to run around.
He needs to annoy his parents like a perfectly healthy toddler.
He needs a future.
That is what his transplant made possible.
And that should remind all of us what healthcare is ultimately for.
Physicians do not enter medicine to manage administrative machinery.
They enter medicine to care for people.
Clinic owners do not build practices because they dream about denial queues.
They build them to serve communities.
Medical billers should not spend their entire day fixing problems that could have been prevented upstream.
And healthcare technology should not create more work simply because it can.
The best workflow is the one that removes itself from the way.
The best automation is the automation that gives people their time back.
The best AI is not necessarily the one that does the most.
It may be the one that knows when not to act.
And the best healthcare technology is ultimately measured by something very simple:
Did it make it easier for a human being to care for another human being?
That is the standard I want to build toward.
Get Involved
Now I want to hear from the people actually living this problem.
If you could eliminate one administrative task from your practice tomorrow, what would disappear first?
Is it prior authorization?
Denial follow-up?
Insurance verification?
Coding questions?
Payer portals?
Documentation requests?
Something else?
Tell me in the comments.
And if this perspective challenges the way you think about medical billing, share or repost it so another physician or clinic owner can join the conversation.
Healthcare does not need another slogan.
It needs people willing to question the workflow.
Question one unnecessary task.
Challenge one broken handoff.
Eliminate one piece of friction.
That is where meaningful change begins.
Continue the Conversation
Healthcare is changing faster than many of its administrative systems.
I share practical ideas about medical billing, healthcare operations, AI, physician entrepreneurship, medical technology and the future of medicine.
You can continue exploring these conversations through my website, podcast, YouTube channel and social platforms.
The goal is simple:
Learn something useful.
Question something we take for granted.
Try something better.
Because knowledge matters most when it changes what we do.
Explore more perspectives on healthcare operations, physician entrepreneurship, medical technology, revenue-cycle management and healthcare innovation:
Listen to the podcast on Spotify
Knowledge creates leverage.
Better questions create better systems.
Better systems give healthcare professionals more time to do the work only they can do.
Start there.
About the Author
Dr. Daniel Cham is a physician, medical consultant and entrepreneur working at the intersection of healthcare management, medical technology, medical billing and AI.
As founder of OnnX, an AI-powered medical billing SaaS focused on small and medium-sized physician practices, Dr. Cham explores practical ways technology can reduce unnecessary administrative work while preserving human judgment and accountability.
His work focuses on a simple question:
How can healthcare technology make healthcare feel more human—not more complicated?
Connect with Dr. Cham on LinkedIn for perspectives on healthcare operations, medical billing, AI, entrepreneurship and the future of medicine.
Disclaimer
This article is intended for general educational and informational purposes only. It does not constitute medical, legal, coding, reimbursement, financial or compliance advice.
Healthcare rules, payer policies and individual circumstances vary. Practices should consult appropriately qualified professionals for guidance regarding their specific clinical, legal, compliance and revenue-cycle situations.
Free Resource
Check the Featured section of my LinkedIn profile for a free resource you can download without signing up.
Use it as a starting point for your next conversation about medical billing, workflow and healthcare operations.
Final Three Sentences
Healthcare does not need more technology for technology’s sake.
It needs better systems that give physicians more time to do the work only humans can do.
And sometimes the smartest thing an AI system can do is help us realize which work should disappear entirely.
References
- People — John Miller’s Kidney Transplant Story
The human-interest source covering John Miller, Greer Miller, Kris Miller, Charlotte Miller, and Lance Miller, including the family connection, organ donation, and June 3, 2026 transplant at Children’s Healthcare of Atlanta. - American Medical Association — Prior Authorization and Physician Burden
Current AMA survey data documenting the impact of prior authorization on physicians, including delays in care, administrative workload, burnout, and clinical outcomes. - Medical Group Management Association — Prior Authorization Turnaround Times
Current MGMA data showing that 44% of medical-group leaders reported slower prior-authorization turnaround times in 2026, highlighting the continuing administrative friction facing medical practices.
#MedicalBilling #RevenueCycleManagement #HealthcareAI #HealthcareInnovation #PhysicianEntrepreneur #PhysicianPractice #MedicalPractice #HealthcareTechnology #AIinHealthcare #HealthcareAdministration #MedicalCoding #DenialManagement #PracticeManagement #PhysicianBurnout #DigitalHealth #HealthTech #ClinicManagement #OnnX
Leave a Reply