Why Independent Medical Practices Are Spending More Money Automating Problems They Should Have Prevented
“The secret of getting ahead is getting started.” — Mark Twain
The most expensive mistake in medical billing may not be
a denied claim. It may be the information error that nobody noticed three weeks
earlier.
The Patient Who Had Two Medical Histories
Imagine a patient arriving at a hospital for a routine
procedure.
The procedure goes perfectly. The physician is satisfied.
The patient is recovering.
Then something goes terribly wrong.
A medication is administered despite a documented,
life-threatening allergy.
Why?
The allergy was recorded under the patient's first name,
Jonathan. The hospital registered him under his legal first name, John.
Two names. Two profiles. One patient.
The physician reviewed one record. The allergy existed in
another.
The information was in the database. It simply wasn't
connected to the person who needed it.
This is an illustrative scenario, not a documented
individual case. But the underlying problem is real: fragmented patient
records and identity-matching failures can create serious patient-safety risks.
Now, let's bring that problem closer to home.
Imagine running an independent medical practice.
Your front desk verifies insurance. Your medical assistant
prepares the chart. You examine the patient. Your documentation is completed.
Your biller submits the claim.
Three weeks later, the claim comes back.
Rejected.
The reason?
An insurance discrepancy that could have been identified
before the appointment.
Now somebody has to investigate.
The front desk checks the registration information. The
biller opens the payer portal. The office manager searches old notes. Someone
calls the patient.
Everyone is looking for information that should have been
available in the first place.
Congratulations. Your practice has just funded another
episode of CSI: Medical Billing.
Except nobody is solving a crime. They're trying to collect
money for work already completed.
Here is the uncomfortable question:
Why are we investing so much effort in correcting
information after it becomes a problem instead of preventing the problem before
it enters the system?
Healthcare has purchased electronic health records, practice
management systems, clearinghouses, billing platforms, automation tools, and
now artificial intelligence.
Yet physicians and medical billers continue to spend
enormous amounts of time chasing missing information.
We have digitized the paperwork.
We have not necessarily fixed the process.
And now we want AI to clean up the mess.
That may be the biggest opportunity in healthcare
technology. It may also be one of its biggest mistakes.
Bad information does not become good information simply
because an expensive AI model reads it.
The Great Healthcare Technology Paradox
Healthcare has more technology than ever.
So why does so much administrative work still depend on
someone remembering to make a phone call?
Why does a medical biller need five browser tabs open to
resolve one claim?
Why does an office manager know more about certain payer
exceptions than the software supposedly designed to manage them?
Why does the physician sometimes have to explain an
encounter that happened weeks ago?
We have sophisticated software.
But we still have remarkably unsophisticated information
handoffs.
Consider the journey of one patient:
Patient → Insurance → Eligibility → Referral →
Authorization → Encounter → Documentation → Coding → Claim → Payer → Payment →
Denial → Resolution.
Every stage depends on the information generated by the
previous stage.
Yet many practices manage these stages using separate
applications, different databases, manual entry, and disconnected
communication.
The result?
A mistake made at registration may not become visible until
billing.
A missing authorization may not become apparent until after
the service.
An incomplete clinical note may not become a financial
problem until weeks later.
And when the problem finally appears, the billing department
gets to investigate.
This is not a particularly intelligent way to run a modern
medical practice.
The revenue cycle is not a collection of independent
transactions. It is one connected operating system.
The industry needs to start treating it that way.
The Biller Is Not Always the Problem. Sometimes the
Biller Is the Messenger.
Let's talk about an uncomfortable subject.
Medical billing departments are frequently judged by the
number of claims they submit, the number of denials they resolve, and the
amount of revenue they collect.
But how often do practice owners investigate the quality of
the information their billers receive?
A biller cannot reliably correct insurance information that
was never verified.
A biller cannot create an authorization that was never
obtained.
A biller cannot legitimately document a service that the
physician did not document.
A biller cannot make a payer follow its contractual
obligations simply by submitting another claim.
Yet the biller is often expected to make everything work.
It is like asking a mechanic to repair a car while refusing
to let the mechanic open the hood.
The problem becomes worse when the most experienced biller
becomes the organization's unofficial operating manual.
Ask Susan.
She knows which payer requires which attachment.
Ask Linda.
She remembers the exception from six months ago.
Ask the office manager.
She knows which portal to check when the first portal says
something different.
What happens when Susan retires?
What happens when Linda gets sick?
What happens when the office manager leaves?
If your practice depends on someone's memory to function,
you don't have a reliable system. You have a very expensive memory test.
This is what I call institutional amnesia.
The organization repeatedly loses knowledge because the
knowledge was never captured in a dependable operational process.
AI offers an opportunity to change this.
But the objective should not be to replace the people who
understand the business.
It should be to stop making those people perform work that a
properly designed system should handle.
Statistics: The Numbers Behind the Information Problem
Healthcare data and administrative inefficiency are
frequently discussed using dramatic numbers.
Some are well established. Others are estimates or results
from individual studies.
Physicians should be careful about treating every industry
statistic as a universal truth.
Here are the areas that deserve attention.
Healthcare data volume: Healthcare is a major and
rapidly growing source of digital information. Clinical notes, diagnostic
images, laboratory results, medication records, claims, and administrative
documents create an enormous information-management challenge.
Unstructured information: A substantial portion of
healthcare information exists in narrative formats, including physician notes,
reports, correspondence, and scanned documents. This creates challenges for
systems that depend on consistent structured fields.
Patient matching: The Office of the National
Coordinator for Health Information Technology recognizes patient matching as a
critical component of interoperability. Incorrectly linking records or failing
to connect records can affect the completeness of a patient's medical history.
Claim errors: Eligibility, authorization, coding,
documentation, and demographic discrepancies can contribute to claim
rejections, denials, and rework.
Administrative burden: The financial consequences are
not limited to unpaid claims. They include employee time, follow-up costs,
delayed cash flow, and management attention diverted from patient care.
The statistic every clinic should calculate
Instead of relying entirely on national estimates, calculate
this:
How many staff hours does your practice spend correcting
preventable errors every month?
For example:
- Hours
spent correcting registration information.
- Hours
spent checking insurance eligibility.
- Hours
spent resolving authorization problems.
- Hours
spent obtaining missing documentation.
- Hours
spent correcting claims.
- Hours
spent following up on avoidable rejections.
Multiply those hours by the actual loaded labor cost.
Now you have something useful.
You have an estimate of the administrative cost of
preventable information failures inside your own practice.
You don't need a billion-dollar industry statistic to
recognize a problem costing your clinic thousands of dollars.
And you certainly don't need another software demonstration
to calculate it.
The Real Problem Starts Upstream
Here is a statement worth challenging:
A denial is not necessarily a billing problem. It may be
an upstream information problem.
Consider four examples.
Example 1: Eligibility
The patient provides an insurance card.
The staff member enters the information.
The practice assumes coverage is active.
The claim is eventually rejected because the coverage
information was inaccurate or outdated.
The biller discovers the problem.
But the original failure occurred earlier.
Example 2: Prior authorization
The physician orders a service.
The practice assumes the service is covered.
The authorization requirement is overlooked.
The service is performed.
The claim is denied.
Now the clinic is trying to resolve a problem that might
have been identified before the service.
Example 3: Clinical documentation
The physician provides appropriate care.
The note does not adequately support a particular coding or
billing requirement.
The biller cannot simply fill in the missing clinical
information.
The physician must review and clarify the record when
appropriate.
The process becomes unnecessarily expensive.
Example 4: Patient identity
The practice has two records for the same patient.
One contains the current insurance information.
The other contains the encounter history.
The staff member works from the wrong profile.
The resulting discrepancies travel downstream.
Different problems.
One common theme.
The information was not dependable when the next person
needed it.
This is why fixing the claim is sometimes the wrong place to
begin.
The better question is:
Where did the information first become unreliable?
AI: Are We Solving Problems or Just Making Them Faster?
The healthcare industry is understandably excited about
artificial intelligence.
AI can read documents, extract information, identify
patterns, classify records, and help automate repetitive administrative work.
These are valuable capabilities.
But there is a problem.
The industry sometimes treats AI as if it were a magic
eraser.
Put an AI model on top of an EHR, connect a few APIs, add a
chatbot, and suddenly the administrative nightmare disappears.
That is not how reliable systems are built.
Consider three situations.
Missing information: A patient's insurance member
number is absent.
AI can identify the missing field.
It cannot responsibly invent the correct number.
Conflicting information: The EHR identifies one
payer, while the payer portal identifies another.
AI can identify the discrepancy.
It should not arbitrarily choose the answer.
Ambiguous clinical documentation: The physician's
note is incomplete.
AI can flag the missing information.
It should not manufacture a clinical finding.
The distinction is important.
AI can assist with information management.
It cannot make unverified information true.
The garbage-in, garbage-out problem has a new name
For years, we have used the expression garbage in, garbage
out.
Now we have generative AI.
The garbage may be more eloquently written, but it is still
garbage.
A polished AI-generated summary can look authoritative even
when its underlying information is inaccurate.
This creates a new responsibility.
Healthcare AI must be evaluated not just by how impressive
its outputs appear, but by whether those outputs can be traced back to reliable
sources.
Accuracy without traceability is difficult to trust.
A system should be able to explain what information it used,
what rule it applied, what action it took, and when human review was required.
That is the difference between a clever demonstration and
dependable healthcare infrastructure.
FHIR: The Universal Translator Is Not the Entire Solution
One of the most important healthcare interoperability
standards is FHIR: Fast Healthcare Interoperability Resources.
FHIR provides a standardized framework for exchanging
healthcare information.
It helps different applications communicate using common
structures and definitions.
For example:
- A
Patient resource identifies a patient.
- An
Encounter resource represents a healthcare encounter.
- An
Observation resource can represent a laboratory result.
- A
Condition resource can represent a documented health condition.
These resources can be connected.
That matters because clinical information often exists in
multiple applications.
The EHR has clinical documentation.
The practice management system has scheduling and
demographics.
The clearinghouse processes claims.
The payer portal provides eligibility and authorization
information.
The billing platform tracks claims and payments.
Without effective interoperability, employees become the
integration layer.
They copy.
They paste.
They compare.
They reconcile.
They repeat.
FHIR can reduce these barriers.
But let's challenge another assumption:
Interoperability does not automatically mean
intelligence.
Two systems can successfully exchange information and still
exchange inaccurate information.
FHIR does not guarantee that the correct patient has been
identified.
It does not guarantee that an insurance policy is active.
It does not guarantee that a clinical note supports a
particular service.
It does not guarantee that the payer will reimburse the
claim.
A universal translator is useful.
But translating nonsense into another language does not make
it meaningful.
The real requirement is interoperability combined with
data validation, workflow ownership, and accountability.
Three Experts, Three Lessons for Independent Practices
The future of healthcare technology should not be determined
by software developers alone.
We need to learn from people who understand clinical
practice, medical errors, professional accountability, and healthcare systems.
Expert 1: Dr. Atul Gawande — Fix the System, Not Just the
Individual
Dr. Atul Gawande, surgeon and author of The Checklist
Manifesto, has spent much of his career examining why capable professionals
still make preventable mistakes.
His work on checklists and healthcare systems offers an
important lesson.
Competence alone does not guarantee reliable performance.
People need systems that support good decisions.
The OnnX takeaway:
Stop treating recurring administrative errors as individual
employee failures.
If the same registration mistake occurs repeatedly, examine
the registration process.
If the same documentation issue creates repeated claim
problems, examine the documentation workflow.
If one employee is the only person who understands a payer
exception, document the knowledge.
The solution is not always another employee.
Sometimes it is a better system.
Expert 2: Dr. John Whyte — AI Should Support Physician
Judgment
Dr. John Whyte, physician and CEO of the American Medical
Association, has emphasized the importance of AI supporting rather than
replacing physician judgment.
This distinction matters.
Healthcare is not simply an information-processing industry.
It is a professional environment involving clinical
judgment, patient relationships, accountability, and ethical obligations.
The OnnX takeaway:
Not every task should be automated in the same way.
Identifying a missing claim field is different from making a
clinical diagnosis.
Extracting information from a document is different from
deciding whether a treatment is appropriate.
Define what AI can do independently.
Define what requires approval.
Define what must remain a professional decision.
AI should reduce administrative friction, not create another
layer of clinical uncertainty.
Expert 3: The American College of Physicians — Ethics
Must Be Part of AI Design
The American College of Physicians has addressed the ethical
responsibilities associated with AI in medical practice through its Ethics,
Professionalism and Human Rights Committee.
The central issue is not whether AI can produce an answer.
It is whether the answer can be trusted, reviewed, and used
responsibly.
The OnnX takeaway:
A healthcare AI system should not operate like an invisible
employee whose decisions nobody can explain.
Physicians and practice owners should be able to understand:
- What
information the system accessed.
- Which
rules it applied.
- What
actions it performed.
- What
information was uncertain.
- When
human review was necessary.
- How
mistakes can be corrected.
If nobody can explain what the AI did, the practice has
gained automation but lost control.
Recent News: Healthcare Technology Is Moving, but Is It
Moving in the Right Direction?
The latest federal developments provide an important example
of why healthcare organizations must look beyond technology marketing.
October 26–28, 2026: HHS Finalizes the HTI-5 Rule
The Department of Health and Human Services finalized the
HTI-5 rule, which changes portions of the ONC Health IT Certification Program.
The rule removes 14 certification criteria and revises seven
others.
It also establishes a longer-term direction toward
FHIR-based APIs.
The stated objective is to reduce regulatory burdens on
health IT developers and support innovation.
Here is the interesting part.
The final rule also removes certain standardized AI model
disclosure requirements from the certification framework.
That creates an important question for physicians:
If some standardized disclosures are no longer required
through this certification framework, how will practices evaluate the AI tools
they purchase?
The answer should not be to stop asking questions.
It should be to ask better questions.
October 29, 2026: American College of Cardiology Examines
the Changes
The American College of Cardiology published an analysis of
the final rule.
Its discussion highlights the changes to certification
requirements and the implications for AI transparency and risk management.
This is important because physicians are increasingly being
asked to trust technology they did not design and may not fully understand.
October 29, 2026: American College of Radiology Raises AI
Transparency Concerns
The American College of Radiology also discussed the
implications of the final rule, particularly the removal of certain AI model
transparency requirements.
Its perspective reinforces the importance of obtaining
information about AI tools directly from vendors.
The contrarian lesson
Deregulation may reduce the burden on software developers.
But reducing regulatory requirements does not automatically
make a technology trustworthy.
A vendor's freedom to innovate should not become a
physician's obligation to trust blindly.
Independent practices need their own evaluation standards.
Before purchasing AI, ask about accuracy, intended use,
limitations, data handling, auditability, and human oversight.
Do not confuse a less burdensome certification environment
with a lower responsibility to protect patients.
Ten Practical Ways to Stop Creating Your Own
Administrative Problems
You do not need to build an artificial intelligence
laboratory in the back office.
You do not need eight Kubernetes nodes.
You do not need to hire a team of data scientists.
Start with the problems that repeatedly consume your staff's
time.
1. Stop assuming yesterday's insurance information is
today's insurance information.
Establish a consistent eligibility verification process.
Record when coverage was checked and the source of
verification.
2. Identify authorization requirements before the
service.
Create a process that assigns responsibility and tracks
authorization status.
Do not rely on someone remembering to check later.
3. Stop treating patient identity as a minor registration
detail.
Use consistent patient matching procedures.
Do not automatically merge similar records without
appropriate verification.
4. Identify recurring documentation deficiencies.
Review common clarification requests.
Work with physicians to improve documentation processes
without turning clinical notes into meaningless templates.
5. Establish clear coding and billing handoffs.
Define what information must be available before a claim is
submitted.
Make unresolved questions visible.
6. Categorize denials by their actual causes.
Separate eligibility, authorization, documentation, coding,
timely filing, and payer processing issues.
A denial report without root-cause analysis is just a
collection of bad news.
7. Assign an owner to every unresolved exception.
Every outstanding issue should have a responsible person, a
deadline, and a next action.
8. Automate repetitive administrative tasks first.
Begin with missing-field detection, eligibility checks,
document classification, and claim-status tracking where supported.
Do not begin with the most complicated clinical decision.
9. Preserve institutional knowledge.
Document payer requirements, recurring exceptions, and
resolution procedures.
Your experienced biller should not have to carry the entire
practice's memory in their head.
10. Measure whether the changes work.
Track claim acceptance, denials, days in accounts
receivable, manual rework, and exception resolution time.
If automation does not improve the numbers, it may simply be
expensive entertainment.
Step-by-Step: How to Build a More Intelligent Revenue
Cycle
Here is a practical implementation roadmap.
Step 1: Map the entire patient-to-payment journey.
Do not examine billing in isolation.
Include registration, eligibility, referrals, authorization,
documentation, coding, claims, payment, and follow-up.
Step 2: Identify the three most expensive recurring
problems.
Do not try to fix everything simultaneously.
Find the problems consuming the most staff time or creating
the greatest financial exposure.
Step 3: Establish baseline metrics.
Measure current performance before introducing new
technology.
Step 4: Find the original source of each problem.
Ask where the information first became inaccurate,
incomplete, or unavailable.
Step 5: Standardize the process.
Define required information, acceptable sources, responsible
employees, and escalation procedures.
Step 6: Introduce automated validation.
Identify missing and conflicting information before it moves
downstream.
Step 7: Introduce AI where it provides measurable value.
Use AI to extract, classify, prioritize, and assist with
repetitive tasks.
Step 8: Preserve human oversight.
Route uncertain or consequential decisions to qualified
people.
Step 9: Maintain an audit trail.
Record what happened, when it happened, and why.
Step 10: Review performance every month.
Measure the actual improvement.
A system that cannot demonstrate its value should not be
protected simply because it has an impressive interface.
Seven Pitfalls: How Clinics Accidentally Make Things
Worse
Pitfall 1: Buying AI before understanding the problem.
A software purchase is not an operational strategy.
Pitfall 2: Assuming more integrations mean better
information.
Connecting systems does not guarantee that their data is
accurate.
Pitfall 3: Allowing AI to fill in missing clinical facts.
Extraction is not invention.
Pitfall 4: Automatically merging patient records.
A false identity match can create serious clinical and
privacy consequences.
Pitfall 5: Automating a broken workflow.
You may simply make the same unnecessary steps happen
faster.
Pitfall 6: Treating every denial as the biller's fault.
Investigate the source before assigning responsibility.
Pitfall 7: Ignoring the people who actually perform the
work.
The front desk, medical assistants, physicians,
administrators, and billers understand where the system breaks.
Involve them before purchasing the solution.
Myth Busters: Seven Healthcare Technology Assumptions
Worth Questioning
Myth 1: AI automatically fixes bad data.
Reality: AI can identify inconsistencies and assist with
corrections. Verification remains essential.
Myth 2: FHIR makes every system interoperable.
Reality: FHIR supports standardized exchange. Implementation
and data quality still matter.
Myth 3: More automation means fewer mistakes.
Reality: Poor automation can scale errors.
Myth 4: Billing begins after the encounter.
Reality: Revenue-cycle performance begins with the
information collected before the encounter.
Myth 5: An experienced biller can compensate for any
broken process.
Reality: Experience is valuable, but constant manual
correction is not a sustainable business model.
Myth 6: AI will eliminate the need for medical billers.
Reality: The more likely opportunity is a shift toward
exception management, auditing, policy interpretation, and operational
oversight.
Myth 7: Data quality belongs to the IT department.
Reality: Data quality is everyone's responsibility because
everyone contributes to the information chain.
Legal Implications: AI Does Not Make the Practice Less
Responsible
Technology can change how work is performed.
It does not automatically change who is responsible for
performing that work appropriately.
Independent practices should consider:
HIPAA: Privacy, security, access controls, business
associate agreements, and appropriate handling of protected health information.
Documentation integrity: AI-generated or extracted
information must not introduce unsupported clinical facts.
Coding and billing compliance: Claims must accurately
represent the services performed and comply with applicable requirements.
Vendor agreements: Understand data ownership,
subcontractors, retention, permitted uses, security obligations, and incident
reporting.
Auditability: Preserve appropriate records of
changes, approvals, and automated actions.
Human accountability: Define who reviews exceptions
and who is responsible for consequential decisions.
The question is not whether the vendor promises that its AI
is compliant.
The question is whether the practice can demonstrate that
its own processes are appropriate.
Ethical Considerations: Patients Are Not Database Entries
There is something important missing from many healthcare AI
discussions.
Patients are human beings.
Their information may be incomplete, inconsistent, or
distributed across multiple organizations.
An AI system must not treat uncertainty as permission to
guess.
Five principles matter:
- Accuracy:
Use reliable and appropriately verified information.
- Privacy:
Limit access to legitimate purposes.
- Fairness:
Monitor errors and bias across patient populations.
- Transparency:
Make the use and limitations of AI understandable.
- Accountability:
Preserve appropriate human review and responsibility.
A patient should not have to suffer because two databases
cannot agree on a name.
And a physician should not have to trust an unexplained
algorithm simply because it produces a confident answer.
Tools, Metrics, and Resources: What Should Your Practice
Actually Monitor?
Technology resources
FHIR: Standardized healthcare information exchange.
USCDI: Standardized health information classes and
data elements.
HL7: Healthcare interoperability standards.
Open Policy Agent: A policy engine for evaluating
defined rules in software systems.
AI document processing: Extracts and classifies
information from clinical and administrative documents.
Practice management and clearinghouse systems:
Support eligibility, claim submission, payment posting, and status tracking.
Ten metrics worth monitoring
- First-pass
claim acceptance rate.
- Claim
denial rate.
- Days
in accounts receivable.
- Net
collection rate.
- Eligibility
error rate.
- Authorization-related
denial rate.
- Documentation
clarification rate.
- Manual
rework hours.
- Exception
resolution time.
- Duplicate
patient record rate.
Do not simply collect these numbers.
Look for relationships.
If eligibility errors increase, does the denial rate
increase?
If documentation clarification requests rise, does claim
processing slow down?
If exception resolution takes longer, does accounts
receivable age?
The value of metrics is not knowing what happened. It is
understanding why it happened.
The Future of Medical Billing: From Processing Claims to
Completing Work
Traditional software records transactions.
The next generation of software will increasingly help
complete operational tasks.
That distinction is significant.
A traditional eligibility system displays coverage
information.
An intelligent workflow can identify discrepancies, route
them to the appropriate person, track the result, and preserve the verification
record.
A traditional billing system displays a rejected claim.
An intelligent workflow can help identify the underlying
issue, prioritize the exception, assign responsibility, and track resolution.
A chatbot can explain a problem.
A properly controlled operational system can help resolve
it.
This is the shift from software that answers questions to
software that gets work done.
But the foundation must be dependable information.
The future will not belong automatically to whoever has the
largest AI model.
It will belong to organizations that combine reliable data,
useful automation, clear ownership, and measurable results.
OnnX: The Physician and the Medical Biller See the Same
Problem From Different Sides
This brings us to the reason behind OnnX.
OnnX is built by a physician and a medical biller.
That combination is intentional.
A physician understands the clinical realities of running a
medical practice.
An experienced medical biller understands the practical
realities of turning clinical services into properly documented, submitted, and
collected claims.
Both understand something that generic technology vendors
sometimes overlook:
The problem does not begin when the claim is submitted.
It begins much earlier.
At registration.
At eligibility verification.
At authorization.
At documentation.
At every handoff where information can become incomplete,
inconsistent, or disconnected.
OnnX is an AI-powered medical billing SaaS focused on
helping small and medium-sized clinics reduce unnecessary administrative
complexity and improve revenue-cycle operations.
The underlying philosophy is straightforward.
Data → Workflows → Revenue
Data: Establish dependable patient, insurance,
clinical, and operational information.
Workflows: Connect activities through clear
ownership, validation, dependencies, and escalation.
Revenue: Reduce preventable rework, improve claim
quality, and strengthen the financial operations of independent practices.
The contrarian OnnX position
We do not need another AI tool that creates another place
for physicians and staff to check.
We need technology that removes unnecessary work.
We do not need to make the biller faster at correcting
preventable errors.
We need to prevent those errors earlier.
We do not need software that simply explains why something
went wrong.
We need systems that help complete the work required to
resolve it.
And we should not confuse adding more software with
improving the practice.
The objective is not to put AI everywhere. It is to
eliminate the work that should never have existed in the first place.
That is the opportunity OnnX is working toward.
Final Thoughts: Stop Treating the Symptoms. Start Fixing
the System.
Healthcare has become remarkably good at documenting
problems.
We have dashboards for denials.
Reports for aging receivables.
Work queues for missing information.
Portals for checking status.
Applications for managing applications.
And, increasingly, AI tools to explain why all the other
applications are not working together.
Something is wrong with that picture.
Perhaps the next breakthrough in healthcare technology will
not come from building a more sophisticated chatbot.
Perhaps it will come from making information dependable
before it enters the next workflow.
Perhaps the most valuable AI feature will not be another
answer.
It will be an unnecessary task that disappears.
And perhaps the best revenue-cycle improvement will not be a
faster way to appeal a denial.
It will be a process that prevents an avoidable denial from
happening.
For independent physicians, this is more than a technology
discussion.
It is a business survival issue.
Independent practices cannot afford to spend their limited
resources repeatedly correcting problems that better systems could prevent.
They need predictable operations.
They need reliable information.
They need technology that works in the background rather
than demanding more attention.
And they need to remain in control of their practices.
A better revenue cycle begins before the claim.
A better AI system begins before the algorithm.
And a better healthcare system begins by asking why we
keep accepting preventable problems as normal.
Get Involved: Three Questions for Physicians and Clinic
Owners
I want to hear from the people who actually deal with these
problems.
Question 1: If you could eliminate one recurring
administrative headache in your practice tomorrow, what would it be?
Question 2: Do you believe medical billing technology
is solving the underlying problems, or simply becoming better at processing
their consequences?
Question 3: How much of your billing team's time is
spent collecting revenue versus correcting information that should have been
accurate from the beginning?
Share your experience in the comments.
Challenge my perspective if you disagree.
And if you know a physician, practice owner, administrator,
or medical biller who would recognize these problems, consider reposting this
article.
The future of independent medicine should not be determined
exclusively by software vendors.
It should be shaped by the people who provide care and
operate medical practices every day.
Let's stop normalizing administrative chaos. Let's start
building better systems.
About the Author
Dr. Daniel Cham is a physician, entrepreneur, and
healthcare technology consultant with experience in medical practice
operations, healthcare management, and medical billing.
He is the founder of OnnX, an AI-powered medical billing
platform built by a physician and a medical biller. His work focuses on helping
independent medical practices rethink the relationship between clinical
information, administrative workflows, artificial intelligence, and
revenue-cycle performance.
Dr. Cham writes about the practical challenges facing
physicians and the opportunities to build more efficient, sustainable, and
independent medical practices.
Disclaimer
This article is intended for general educational and
informational purposes. It does not constitute medical, legal, coding, or
compliance advice. Readers should consult qualified professionals regarding
their specific circumstances and applicable regulations.
Continue the Conversation
Explore more practical ideas, professional perspectives, and
firsthand observations about healthcare operations, medical technology,
artificial intelligence, and the future of independent medicine.
Connect: LinkedIn | Website | YouTube| Spotify
Knowledge drives progress. Question conventional
thinking, learn from experience, and help build a better future for independent
medicine.
References and Related Reading
1. The
Denial Isn't the Problem. It's the Clue.
Explores how seemingly minor information errors travel through registration,
eligibility, authorization, documentation, and billing—eventually becoming
costly claim problems.
2. Your
Medical Billing Problem May Not Be a Billing Problem: The Hidden Cost of
Creating Work You Should Never Have to Do
Examines how upstream operational failures create downstream administrative
work and why prevention is more valuable than repeatedly correcting avoidable
errors.
3. The
Best Claim Is the One You Never Have to Fix
Explains why reliable information, early validation, and preventive controls
matter more than simply becoming faster at denial management.
4. FHIR
Overview — HL7 International
The official FHIR specification overview, explaining how standardized
healthcare resources support electronic information exchange and
interoperability.
5. Artificial
Intelligence Governance in Health Systems: Systematic Review of Frameworks and
Integrative Model Proposal — Journal of Medical Internet Research
A 2026 systematic review identifying data governance, risk assessment,
validation, monitoring, and integration as important components of healthcare
AI governance.
6. Augmented
Intelligence in Medicine — American Medical Association
Explains the AMA's position that AI should enhance physician capabilities
rather than replace professional judgment, with attention to transparency and
responsible implementation.
7. HHS
Finalizes HTI-5 Deregulatory Rule to Encourage Innovation and Reduce Health IT
Certification Program Compliance Costs — U.S. Department of Health and Human
Services
October 26, 2026. Announces the HTI-5 final rule and its changes to the ONC
Health IT Certification Program.
8. HTI-5
Final Rule Overview — Office of the National Coordinator for Health Information
Technology
Summarizes the removal of 14 certification criteria, revision of seven, and the
direction toward FHIR-based API requirements.
9. Health
Data, Technology, and Interoperability: ONC Deregulatory Actions to Unleash
Prosperity — Federal Register
The published final-rule text for readers who want to examine the regulatory
language and specific provisions.
Free Resource: Visit the Featured section of my
LinkedIn profile for a complimentary resource. No registration or signup
required.
If this perspective resonates, consider reposting it.
Help more physicians, clinic owners, and medical billers question the way
healthcare administration works—and imagine something better.
#MedicalBilling #HealthcareAI #RevenueCycleManagement
#IndependentPhysicians #HealthcareTechnology #MedicalPracticeManagement
#HealthcareData #FHIR #HealthcareInteroperability #AIinHealthcare
#MedicalBillingAutomation #PatientSafety #DataQuality #PhysicianEntrepreneur
#OnnX
