A nurse knew healthcare from the inside. Then her daughter Zoe became a 1-pound, 7-ounce NICU patient—and Jennie Willis discovered how quickly a medical crisis can become a family crisis.
“Patients should not have to fight the healthcare system
to get the care they need and deserve.” — Willie
Underwood III, MD, MSc, MPH, President, American Medical Association
The nurse became the patient’s mother.
Jennie Willis had spent her career caring for patients.
As a registered nurse in the PACU at Sunrise Hospital
& Medical Center in Las Vegas, she understood healthcare from the
inside.
She knew the terminology.
The procedures.
The waiting.
The uncertainty.
The quiet conversations in hallways.
Then, after multiple miscarriages, Jennie became pregnant
with her third child.
At 24 weeks, she learned that her baby was not receiving
enough blood and nutrients through the placenta.
She was admitted to MountainView Hospital.
The objective was simple:
Give the baby more time.
At 26 weeks and three days, Jennie delivered her
daughter, Zoe, by C-section.
Zoe weighed just 1 pound, 7 ounces.
She would spend more than four months in the NICU.
Suddenly, Jennie was experiencing healthcare from the other
side of the bed.
Not as the nurse.
As the mother.
And the story became much bigger than medicine.
Jennie and her husband were caring for Zoe while also caring
for their two older children.
Eventually, Jennie returned to work while Zoe was still
hospitalized.
She visited Zoe during her breaks.
Then returned to the NICU after her shifts.
Think about that.
She was working in a hospital while her own daughter
remained hospitalized.
The clinical problem was enormous.
Then another problem appeared.
The bill.
The family was relying on one income.
Medical bills were arriving.
Repeated trips to the hospital added expenses.
The weeks became months.
A nurse at MountainView told Jennie about the HCA
Healthcare Hope Fund.
The assistance helped with household and medical expenses
and gave the family some financial stability during an extraordinarily
uncertain period.
Today, Zoe is home and thriving.
That is the good ending.
But Jennie's story leaves us with a bigger question.
What if the problem we can see is not the problem that
started first?
Healthcare has a strange habit.
We tend to notice problems at the end of the process.
The patient misses an appointment.
We notice.
The authorization expires.
We notice.
The claim denies.
We notice.
The physician gets a query.
We notice.
The patient receives a bill.
We notice.
Then we ask:
“Who is going to fix this?”
That question is understandable.
It is also frequently too late.
Because by the time a problem becomes visible, someone may
already have spent hours creating it.
Sometimes days.
Sometimes weeks.
And occasionally an entire department.
The denial may be the smoke alarm.
Imagine your kitchen has a water leak.
You see water on the floor.
So you mop it up.
The next morning, more water.
You mop again.
More water.
Eventually, someone proposes a brilliant operational
solution:
Hire a faster mop.
That is roughly how bad process design can look.
The denial team becomes the mop.
The denial becomes the puddle.
The real leak remains underneath the sink.
And everyone congratulates themselves because the floor is
dry by 4 p.m.
Until tomorrow.
This is the uncomfortable question:
Are we fixing the problem—or becoming extremely efficient
at repairing the consequences of the problem?
The revenue cycle has become very good at consequences.
We have:
Denial teams.
Appeal teams.
Authorization teams.
Coding teams.
Eligibility teams.
Revenue-integrity teams.
Patient-financial-services teams.
Collections teams.
Analysts.
Dashboards.
Work queues.
Escalation queues.
And enough passwords to qualify as a cardiovascular stress
test.
Yet practices still experience recurring problems.
That should tell us something.
Maybe the industry does not have a shortage of people fixing
problems.
Maybe it has a shortage of systems that prevent the same
problem from being created repeatedly.
The contrarian question
Here is the question I would put on every practice manager's
desk:
When did we first have enough information to know this
could become a problem?
Not:
“Why did the payer deny it?”
Not:
“Who forgot to submit it?”
Not:
“Who is going to work the denial?”
Those questions matter.
But they are downstream.
The more interesting question is:
When could we have known?
That changes everything.
Denial management is not denial prevention.
The words sound similar.
They are not.
Denial management asks:
How do we recover money after something went wrong?
Denial prevention asks:
How do we reduce the probability that it goes wrong?
And upstream intelligence asks an even earlier
question:
What information, decision, or workflow condition created
the risk in the first place?
Those are three different problems.
Treating them as one problem is how organizations end up
buying another dashboard when what they really needed was better information
flow.
The data is telling us something.
A September 2026 MGMA Stat poll found that 44% of medical
group leaders said payer prior-authorization turnaround had become slower in
2026, while 40% said it was about the same.
Only 7% reported faster turnaround.
The poll included 178 applicable responses.
But here is the interesting part.
The delay does not necessarily begin when the payer receives
the request.
It can begin before submission.
Someone has to determine whether authorization is required.
Someone has to find the correct submission channel.
Someone has to collect documentation.
Someone has to enter information.
Someone has to monitor status.
Someone may have to respond to requests for additional
information.
Then comes the peer-to-peer review.
Then perhaps the denial.
Then the appeal.
Then another round.
So when we say:
“The payer took seven days.”
We may be measuring only seven days of a much longer
administrative journey.
That distinction matters.
Faster does not always mean better.
This may be one of the strangest truths in healthcare
administration.
Suppose a practice reduces authorization turnaround from
five days to two.
Fantastic.
But what if it accomplished that by hiring two additional
people to spend their entire day checking payer portals?
The patient may get an answer faster.
The practice may still be carrying the same structural
burden.
We have improved speed.
We have not necessarily improved efficiency.
That is why the right metric is not always:
How fast did the process finish?
Sometimes it is:
How much unnecessary work did the process require?
The seven-portal problem.
MGMA reporting has highlighted another revealing reality:
many practices have staff accessing seven or more payer portals each week.
Seven.
Not seven patients.
Seven portals.
And each portal can have different:
Rules.
Passwords.
Submission requirements.
Documentation requirements.
Interfaces.
Status screens.
Workflows.
It is hard not to wonder whether we have accidentally
created a new medical specialty:
Portal Navigation.
No medical school teaches it.
Yet healthcare workers perform it every day.
And then there is Susan.
Every practice has a Susan.
Susan knows:
Which payer hates modifier 25.
Which plan requires authorization.
Which portal actually works.
Which fax number nobody answers.
Which payer changed a rule three months ago.
Which physician's notes tend to trigger questions.
Which claim needs a phone call.
Which payer representative actually calls back.
Susan is incredibly valuable.
But there is a problem.
Susan is not a database.
She can retire.
Take vacation.
Change jobs.
Get promoted.
Or simply have a bad Tuesday.
When operational intelligence exists only in people's
memories, the organization is borrowing knowledge rather than owning it.
The goal of technology should not be to eliminate Susan.
It should be to make sure Susan's useful knowledge does
not disappear when Susan walks out the door.
This is where the Jennie Willis story becomes relevant.
Jennie's story is not a billing-denial story.
The published account does not say that a denial caused her
family's hardship.
It says something more human.
A nurse who understood healthcare suddenly experienced it as
a mother.
The medical problem was only one part of the family's
reality.
There was also:
Work.
Transportation.
Two older children.
Medical expenses.
Household expenses.
Uncertainty.
Stress.
And the need to keep functioning while a tiny baby remained
in intensive care.
That is what context looks like.
One event. Multiple systems. One human being living
through all of them.
Healthcare systems often separate those experiences.
Patients do not.
The patient is not a claim.
A claim is a representation of an encounter.
It is not the encounter itself.
A diagnosis code is not the patient.
A procedure code is not the patient.
An authorization number is not the patient.
A denial code is certainly not the patient.
Yet revenue-cycle systems often encounter patients primarily
through these artifacts.
That creates a dangerous mental shortcut:
We start optimizing the representation instead of
understanding the event that created it.
That is backwards.
Billing is often a data problem wearing a billing
costume.
Consider what happens before a claim exists.
Registration.
Eligibility.
Coverage.
Scheduling.
Authorization.
Clinical encounter.
Documentation.
Diagnosis.
Medical decision-making.
Procedure.
Time.
Charge capture.
Coding.
Claim creation.
Every step creates information.
Every step can lose information.
Every handoff can introduce ambiguity.
Every delay can make information stale.
Every disconnected system can create another version of the
truth.
Then the claim arrives downstream carrying the accumulated
consequences.
And someone says:
“The claim is wrong.”
Maybe.
But perhaps the claim is simply reporting a problem that
began five steps earlier.
The claim is where the problem becomes visible.
It may not be where the problem began.
That distinction is the heart of the OnnX thesis.
Not:
“Let's build a better denial dashboard.”
But:
“Let's understand why the claim became vulnerable before
it was submitted.”
That means looking upstream.
Not because downstream work is unimportant.
Because downstream work is expensive.
AI does not magically fix bad workflows.
This needs to be said loudly.
Healthcare does not need another AI system that produces 300
alerts.
Nobody wakes up thinking:
“I wish my EHR gave me more things to click.”
The objective should not be maximum automation.
It should be:
maximum useful information at the right moment.
AI can identify patterns.
Compare information.
Summarize documentation.
Detect inconsistencies.
Prioritize potential problems.
Surface payer-specific patterns.
Learn from historical outcomes.
But AI should not pretend that uncertainty has disappeared.
Sometimes the correct answer is:
“I don't know.”
That is not failure.
That is responsible intelligence.
Prediction is not prevention.
Suppose an AI system predicts that a claim has a high
probability of denial.
Great.
Then what?
If the practice does nothing, the prediction accomplished
very little.
The real chain is:
Prediction → Decision → Intervention → Outcome → Learning
Not:
Prediction → Dashboard → Meeting → Another dashboard
A prediction without an actionable response is simply an
expensive notification.
Healthcare already has plenty of those.
Human judgment still matters.
A physician understands clinical context.
A coder understands coding rules.
A biller understands payer behavior.
A practice manager understands operational reality.
A good system should connect those perspectives.
It should not pretend one algorithm can replace all of them.
The future is not:
AI versus humans.
It is:
AI for scale.
Humans for judgment.
Data for learning.
That is a much more useful model.
Three myths worth retiring
Myth 1: “Our denial team will fix it.”
Maybe.
But if the same denial keeps returning, you have not
necessarily fixed the system.
You may simply have built an excellent repair shop.
There is a difference.
Myth 2: “More automation means less work.”
Not necessarily.
Bad automation can create:
More alerts.
More exceptions.
More reconciliation.
More cleanup.
More distrust.
Automation magnifies process quality.
If the process is bad, congratulations.
You just made bad work faster.
Myth 3: “The most sophisticated AI wins.”
No.
The system that people actually use wins.
A tool that identifies three meaningful problems may be more
valuable than a spectacular platform producing 300 theoretical risks.
Useful beats impressive.
Every time.
The hidden cost is not always the denial.
Suppose one recurring problem requires 20 minutes of staff
time.
One claim?
Fine.
A hundred claims?
More than 33 hours.
Every month?
Now you have created a job.
Every year?
You have created a small department.
And that is before accounting for:
Delayed cash.
Patient calls.
Physician queries.
Appeals.
Management time.
Staff frustration.
Opportunity cost.
The financial loss is not always sitting inside the denial
amount.
Sometimes it is sitting inside everyone's calendar.
What should a practice measure?
Days in A/R matters.
Denial rate matters.
Collection rate matters.
But consider adding:
Preventable denial rate
First-pass acceptance rate
Rework minutes per 100 claims
Documentation-related denial rate
Authorization-related delay
Eligibility error rate
Repeat-denial rate
Time from detection to resolution
Staff hours spent on recurring problems
And perhaps the most interesting metric:
How many problems did we prevent instead of repair?
That question changes the conversation.
A 30-day experiment
You do not need a massive transformation project.
Try one problem.
Week 1: Find the leak
Choose your most common recurring denial.
Pull 25 examples.
Look for patterns.
Week 2: Find the earliest signal
Ask:
What was the first point at which someone could
reasonably have known this was going to become a problem?
Week 3: Change one thing
One workflow.
One checkpoint.
One documentation prompt.
One payer-specific rule.
One authorization check.
Do not redesign the entire universe.
Week 4: Measure
Compare:
Before.
After.
Rework.
Staff time.
Denials.
Resolution.
Then decide whether the intervention deserves to survive.
The legal and ethical line
Revenue optimization has a boundary.
It should improve the accuracy of representing care.
It should never encourage manipulation of care documentation
or coding simply to increase reimbursement.
The objective is not:
Find the highest-paying code.
It is:
Accurately represent the care that occurred.
That distinction is fundamental.
Technology also has to account for:
HIPAA.
Security.
Coding compliance.
Medical necessity.
Payer contracts.
Auditability.
AI governance.
Human oversight.
Automation does not transfer accountability.
If software recommends something, someone still needs to
determine whether that recommendation makes sense.
Small practices cannot afford enterprise-sized
complexity.
A large health system can throw people at a problem.
An independent practice often cannot.
The physician may be the owner.
The practice manager may handle HR.
The nurse may handle authorization.
The front desk may handle eligibility.
The biller may handle denials.
Everyone already has another job.
That means technology for small and medium-sized practices
needs to respect something very scarce:
attention.
Not just money.
Attention.
A system that requires seven dashboards, three consultants,
two months of training, and a 94-page implementation manual may be
technologically impressive.
It may also be completely wrong for a five-physician
practice.
Simple is not unsophisticated.
Sometimes simple means someone finally understood the
workflow.
What OnnX is exploring
This is the problem space behind OnnX.
OnnX is an AI-powered healthcare revenue-cycle and medical
billing SaaS platform being developed for small and medium-sized medical
practices and clinics.
But the interesting part is not “AI billing.”
The interesting part is where intelligence enters the
workflow.
Instead of waiting for the claim to fail:
Understand.
Detect.
Decide.
Act.
Learn.
The goal is to move intelligence closer to the point where
information and decisions are created.
That means thinking beyond:
“Can we automate this claim?”
And asking:
“Can we make the decision that creates this claim
better?”
That is a different category of problem.
The future may be less about faster billing.
It may be about fewer reasons to bill reactively.
Imagine a system that understands:
The practice.
The payer.
The provider.
The encounter.
The documentation.
The workflow.
The historical outcome.
And the uncertainty.
Not to replace people.
To give them context.
Before the problem becomes expensive.
That is the direction worth exploring.
One uncomfortable possibility
Maybe healthcare has spent decades optimizing the wrong end
of the pipe.
We became very good at:
Finding denials.
Counting denials.
Categorizing denials.
Reporting denials.
Appealing denials.
Forecasting denials.
Hiring people to work denials.
Buying software to manage denials.
And yet we may still be asking the wrong question.
Why did the denial have to exist in the first place?
That question moves the conversation upstream.
Jennie Willis gives us the human version of the lesson.
Jennie knew healthcare.
Then Zoe became her patient.
Jennie understood the clinical system.
But suddenly she was navigating the system as a mother.
The story reminds us that healthcare is never just the
clinical event.
There is always a surrounding human system.
Work.
Money.
Family.
Transportation.
Time.
Stress.
Support.
Information.
The same principle applies operationally.
A claim is never just a claim.
It is the final expression of dozens of earlier decisions
and information exchanges.
When something goes wrong downstream, we should not only
ask:
“How do we fix it?”
We should ask:
“Where did we first have a chance to prevent it?”
Maybe that is the real healthcare AI opportunity.
Not replacing the biller.
Not replacing the physician.
Not replacing the practice manager.
Not predicting everything.
Not producing another dashboard.
Maybe the opportunity is much simpler.
Give the right person the right context early enough to
make a better decision.
That is intelligence.
Everything else is infrastructure.
Get Involved
Here is the question I want to put to physicians, practice
owners, administrators, coders, and billers:
What problem does your practice keep fixing that you
suspect actually starts somewhere else?
Eligibility?
Authorization?
Documentation?
Coding?
Payer behavior?
Registration?
Workflow?
Something nobody has named yet?
Tell me what you see in the comments.
The recurring problems are often the most interesting ones
because someone has usually become so good at working around them that nobody
remembers to question why the workaround exists.
And if this perspective resonates, repost it for another
physician or practice owner who spends too much time repairing problems that
should have been prevented upstream.
Don't just fix the denial. Find the decision that created
it.
Don't just automate the queue. Question why the queue
exists.
Don't just measure what went wrong. Measure what could
have been prevented.
About the Author
Dr. Daniel Cham is a physician,
physician-entrepreneur, and medical consultant focused on healthcare
technology, healthcare management, medical billing, and revenue-cycle
intelligence.
As founder of OnnX, Dr. Cham explores how AI,
clinical context, workflow intelligence, and better data can help
physician-owned practices move from reactive revenue-cycle management toward
more proactive decision-making.
His work focuses on practical questions at the intersection
of medicine, healthcare operations, technology, revenue integrity, and
innovation.
Connect with Dr.
Daniel Cham on LinkedIn
Disclaimer
This article is intended for general educational and
informational purposes only. It does not constitute medical, legal, coding,
compliance, financial, or professional advice.
Healthcare regulations, payer requirements, contracts,
coding rules, and technology standards can change and vary by circumstance.
For advice concerning a particular patient, practice,
claim, contract, compliance issue, legal matter, or technology implementation,
consult an appropriately qualified professional.
Continue the Conversation
Healthcare innovation rarely begins with a new piece of
software.
It often begins with a better question.
For conversations about healthcare operations, medical
technology, physician entrepreneurship, medical billing, revenue-cycle
intelligence, and innovation, connect with Dr. Cham through his professional
channels.
Knowledge creates options. Better questions create better
systems. Start upstream.
Featured Resource
Visit the Featured section of my LinkedIn profile for
a free resource.
No signup required.
If this perspective resonates, consider reposting it so
another physician, administrator, or practice owner can join the conversation.
References
HCA
Healthcare Hope Fund — Jennie Willis and Zoe
The September 22, 2026 story identifies Jennie Willis as a
registered nurse in the PACU at Sunrise Hospital & Medical Center in Las
Vegas. It describes Zoe's birth at 26 weeks and three days, her 1-pound,
7-ounce birth weight, more than four months in the NICU, Jennie's return to
work while Zoe remained hospitalized, and the family's financial strain and
assistance through the HCA Healthcare Hope Fund.
American
Medical Association — Willie Underwood III, MD, MSc, MPH
The AMA President's September 4, 2026 article states:
“Patients should not have to fight the healthcare system to get the care they
need and deserve.” The same article reports findings from the AMA's latest
survey of 1,000 practicing physicians, including an average of 40 prior
authorizations per week and approximately 13 hours of physician/staff time
devoted to them.
MGMA
— Prior Authorization Turnaround, 2026
MGMA reported that in a September 1, 2026 poll of 178
applicable responses, 44% of medical group leaders said payer
prior-authorization turnaround was slower in 2026, 40% said it was about the
same, 7% said it was faster, and 9% were unsure. MGMA also described workload
occurring before and after the formal payer decision, including submission,
documentation, status checks, additional-information requests, peer-to-peer
reviews, denials, and appeals.
Hashtags
#Healthcare #MedicalBilling #RevenueCycleManagement
#HealthcareAI #MedicalCoding #DenialManagement #HealthcareInnovation
#HealthTech #PhysicianEntrepreneur #RevenueIntegrity #HealthcareOperations
#MedicalPractice #ClinicalDocumentation #PriorAuthorization
#HealthcareTechnology #PhysicianOwnedPractice #PracticeManagement
#AIinHealthcare #OnnX
