You can’t fix a leak by mopping the floor. Medical billing’s most expensive mistakes may begin long before a claim is submitted—when critical information is missed, misplaced, or lost between systems.
“Prior authorization won’t be fixed by technology alone.
It requires the entire healthcare system to work together to solve real-world
challenges.” — Dr.
Mehmet Oz, CMS Administrator, May 13, 2026
The claim went out. The denial came back. Everybody
blamed billing.
Picture a familiar scene.
A physician sees a patient. The visit goes well. The
documentation is completed. The billing team submits the claim.
Then comes the plot twist.
The claim is denied because authorization was missing,
coverage information was incorrect, documentation didn’t support the service,
or a required detail never made it into the workflow.
Now someone must investigate, call the payer, open a portal,
search the chart, ask the clinical team for clarification, correct the
information, resubmit the claim, and wait again.
Congratulations. Your practice has just purchased another
round of administrative work for a problem it may have been able to prevent.
And the most frustrating part?
The denial may not have started in billing at all.
The error could have entered the process days earlier—during
scheduling, registration, eligibility verification, referral intake,
authorization, documentation, or a handoff between departments.
By the time the claim reaches the billing team, the original
mistake may be hiding behind several layers of software and human workarounds.
Billing gets the blame because billing sees the failure.
But seeing where a problem becomes visible is not the same
as identifying where it began.
The contrarian take: Your billing team may not be the
problem
Healthcare has become remarkably good at building systems to
clean up after other systems.
One team collects patient information. Another verifies
coverage. Someone else checks authorization. A clinician documents the
encounter. A coder interprets the record. A biller translates the encounter
into a claim. A clearinghouse checks the submission. A payer applies its rules.
Each participant may be doing their job correctly.
Yet the overall process can still fail.
Why?
Because a workflow is only as reliable as the information
that survives every handoff.
A patient’s insurance changes. A referral arrives late. An
authorization is associated with the wrong service. A clinical note lacks a
necessary detail. A scheduling change never reaches the person responsible for
verification.
The system keeps moving, but the information underneath it
has changed.
Then the practice spends time reconciling the difference.
This is the uncomfortable question healthcare operations
should be asking:
How much of the work we call “billing complexity” is
actually the cost of information that was incomplete, inconsistent, outdated,
or disconnected upstream?
Not every denial is preventable. Payer policies,
medical-necessity determinations, contract disputes, and legitimate coverage
limitations are real. But treating every denial as an isolated billing event
hides an important opportunity: identify the failures that repeatedly create
avoidable rework.
A faster denial-management process may improve recovery.
Preventing the underlying error may eliminate the need for
recovery in the first place.
The numbers make this more than an administrative
annoyance
The Centers for Medicare & Medicaid Services (CMS)
estimates that requesting prior authorizations costs providers approximately $20–$50
per hour and takes an average of 13 hours per week. CMS translates
that into roughly $34,000 and 700 hours of administrative time annually per
provider.
Those estimates concern prior authorization, not all medical
billing denials. They should not be interpreted as the cost of every denial or
as a guaranteed savings opportunity for every practice. But they illustrate the
scale of the administrative burden surrounding information collection,
submission, follow-up, and coordination.
Now consider what happens when authorization work is
disconnected from scheduling, the clinical record, or insurance verification.
The practice may not simply spend time obtaining an
approval. It may also spend time discovering that an approval was needed,
locating missing documentation, correcting an order, rescheduling a patient,
and explaining the delay.
One information gap can generate several downstream tasks.
The real cost is not always the individual error. It is
the chain reaction the error creates.
A current signal from the industry: Digital connections
are not enough
In an interview published by Healthcare Dive on
October 26, 2026, Ryan Bohochik, Epic’s vice president of value-based care,
described the outdated ways healthcare organizations still exchange
information:
“There’s just all these ways that people have to exchange
information that feel outdated.”
That observation matters because digitizing an old process
does not automatically make the process intelligent.
A fax can become a scanned PDF. A phone call can become a
portal message. A spreadsheet can become a dashboard. A manual checklist can
become an electronic form.
But if the underlying information is still incomplete,
duplicated, or disconnected, the same problem can travel through a more modern
interface.
Technology can accelerate the exchange of information. It
cannot reliably compensate for information that was never captured, was
captured incorrectly, or cannot be matched to the right patient, service,
payer, and workflow.
The next step is not simply to make healthcare paperwork
digital.
It is to make the information dependable enough that
fewer people have to chase it.
The 2027 shift: Why practices should pay attention now
CMS has established requirements for certain impacted health
plans to implement standardized electronic prior authorization APIs beginning
January 1, 2027. These interfaces are intended to support electronic exchanges
involving coverage requirements, documentation, and authorization requests.
This is meaningful progress. Standardized interfaces can
reduce portal switching, manual submission, and unnecessary back-and-forth when
the participating systems are connected and the workflows are implemented
effectively.
But there is a catch.
An API can transmit the information it receives. It cannot
magically guarantee that the information is correct, complete, current, or
clinically appropriate.
A connected system can still send bad data efficiently.
A well-designed digital workflow therefore needs more than
connectivity. It needs reliable information capture, clear ownership,
validation, context preservation, and a way to identify missing requirements
before they become expensive problems.
For independent practices, the practical question is not
just, “Will our EHR support the new technology?”
It is also:
“Will our workflow be ready to use it correctly?”
Six places where a billing problem can begin before
billing
1. Patient registration
A transposed digit, incorrect demographic detail, or
mismatched insurance record can follow the patient through multiple systems.
What to do: Establish a consistent registration
checklist. Validate key fields at intake and define when information must be
rechecked rather than assuming that yesterday’s record is still accurate.
2. Insurance eligibility
An insurance card is evidence of coverage information—not a
guarantee that a particular service is covered on a particular date under the
applicable plan rules.
What to do: Verify eligibility at the appropriate
point in the workflow. Record the verification date, payer response, relevant
coverage details, and any unresolved discrepancies.
3. Referral and prior authorization
A service may require a referral, authorization, or specific
supporting documentation. Requirements can vary by payer, plan, service, and
clinical circumstances.
What to do: Define who checks requirements, who
submits the request, who tracks the decision, and who confirms that the
approval applies to the intended patient, service, provider, and date range.
4. Clinical documentation
A clinician may have provided appropriate care, yet the
record may not contain all the information needed to support coding, medical
necessity, or payer-specific requirements.
What to do: Use clinically appropriate documentation
guidance and targeted prompts. Do not turn every note into a giant checklist or
encourage documentation that is not supported by the actual encounter.
5. Handoffs between teams
The front desk knows one thing. The authorization team knows
another. The clinician updates the plan. The billing team receives an
incomplete snapshot.
What to do: Identify the information that must follow
the patient and encounter from one stage to the next. Assign clear ownership
for updates and exceptions.
6. Claim submission and feedback
Some issues can only be identified after a claim is
submitted or a payer responds. The mistake is not necessarily that the practice
failed to prevent every error. It is failing to learn from recurring errors.
What to do: Categorize denials by root cause, source
workflow, payer, service type, and responsible process—not merely by the team
that handled the final claim.
The goal is to distinguish a one-time exception from a
repeatable failure.
A practical playbook: Stop managing denials as
disconnected events
You do not need to replace every system in your practice to
begin improving the process. Start by examining how information moves through
the systems you already have.
Step 1: Establish your baseline
Review a meaningful sample of recent denials, rejections,
authorization delays, and claims requiring rework.
For each case, ask:
- What
was the immediate reason for the failure?
- When
did the underlying information first become incorrect or incomplete?
- Which
workflow owned that information at the time?
- Could
the problem reasonably have been detected earlier?
- Was
the information available but ignored, unavailable, outdated, or never
collected?
Avoid assuming that every denial has a single root cause.
Some cases involve multiple contributing factors.
Step 2: Separate the symptom from the cause
A denial code tells you something about the payer’s
response. It does not always explain the full operational history that led to
the response.
For example, an authorization-related denial may involve a
missed requirement, a delayed request, a mismatch in the authorization record,
a change in the scheduled service, or a payer-processing issue.
Build a root-cause category for each recurring failure. Keep
payer adjudication reasons separate from your internal operational findings.
Step 3: Map the information journey
Follow the process from appointment scheduling to claim
submission.
Document where patient, insurance, referral, authorization,
clinical, and service information is created, checked, updated, and
transferred.
Look for duplicate entry, manual rekeying, disconnected
portals, unclear ownership, and information that becomes stale without anyone
being notified.
A simple workflow map often reveals more than another
meeting about “improving efficiency.”
Step 4: Create checks at the point of risk
A useful control happens before the error becomes expensive.
Examples include a missing-information alert before
authorization submission, a verification checkpoint when insurance changes, or
a review of whether the authorization matches the planned service.
Use proportionate checks. If every routine task triggers an
alert, staff will learn to ignore alerts—including the important ones.
Step 5: Assign an owner to every exception
Automation should not create a new mystery about who is
responsible.
Define who reviews a mismatch, who contacts the payer, who
obtains missing documentation, who updates the record, and who confirms
resolution.
Escalation rules should make exceptions visible rather than
allowing them to disappear into an inbox.
Step 6: Close the learning loop
Review recurring failures regularly. When a root cause is
confirmed, fix the workflow, update the relevant guidance, train the people
involved, and monitor whether the failure rate changes.
A denial resolved today is useful.
A recurring denial prevented next month is better.
Measure the process—not just the money recovered
A practice can improve its denial-recovery numbers while
continuing to generate unnecessary administrative work. Recovery and prevention
are different objectives, so measure both.
Consider tracking:
- Initial
denial rate: The share of submitted claims denied initially, with a
clearly defined denominator and reporting period.
- Clean-claim
rate: The share of claims accepted without preventable correction or
rework, using a consistent definition.
- Authorization-related
failure rate: The frequency of failures associated with missing,
mismatched, expired, or otherwise problematic authorizations.
- Rework
hours: Staff time spent correcting information, researching failures,
and resubmitting claims.
- Time
to resolution: The elapsed time from identifying a problem to a
documented resolution.
- Repeat-cause
frequency: How often the same root cause returns after a corrective
action.
- Avoidable
appointment disruption: Cancellations or rescheduling attributable to
preventable administrative issues.
Break results down by payer, service, workflow, and root
cause where the data supports it. Establish a baseline before declaring
improvement.
Do not claim that an intervention saved money simply because
denial rates declined during the same period. Changes in payer mix, service
mix, coding practices, staffing, and other factors can affect results.
If you only measure how quickly your team cleans up
mistakes, you may never discover why the same mistakes keep arriving.
Five traps that make the problem worse
Trap 1: Buying another tool before understanding the
workflow.
Software may help, but automating a poorly understood process can preserve the
same failure at greater speed.
Trap 2: Blaming the last person who touched the claim.
The person who discovers the problem may have had no control over where it
originated.
Trap 3: Treating every denial as preventable.
Some denials are legitimate disputes or reflect payer decisions outside the
practice’s control. Prevention requires accurate classification, not wishful
thinking.
Trap 4: Adding endless alerts.
Too many alerts create fatigue and encourage workarounds. Prioritize checks
based on risk and clinical or financial significance.
Trap 5: Using AI without accountability.
An AI system can help extract information, flag inconsistencies, summarize
payer requirements, or route exceptions. It can also misread a document, rely
on stale information, or generate a confident but incorrect recommendation.
Human oversight, validation, auditability, privacy protections, and clear
escalation paths remain essential.
The objective is not to make staff follow more steps
forever. It is to remove avoidable steps while protecting the quality and
integrity of the work.
Myth buster: “If we have an EHR, our information problem
is solved.”
Not necessarily.
An electronic health record can store large amounts of
information and still leave gaps between clinical documentation, scheduling,
insurance verification, authorization, and billing.
Data can exist in the system without being structured
correctly for a particular workflow. It can be available but outdated. It can
be accurate in one place and inconsistent in another.
The question is not simply whether information is
electronic.
The question is whether the right information reaches the
right person or system, at the right time, in a usable form—with enough context
to support the next decision.
That is a workflow and information-quality challenge, not
just a software-storage challenge.
Where AI actually belongs
AI can be useful in identifying missing fields, comparing
records, extracting information from documents, flagging inconsistencies,
summarizing authorization requirements, and prioritizing exceptions for human
review.
But AI should not be treated as a magic denial eraser.
If the source data is incomplete, the system needs to
recognize uncertainty rather than quietly manufacture an answer. If payer
requirements change, the workflow needs a way to use current, authoritative
information. If an AI-generated suggestion affects a clinical or financial
decision, the practice needs appropriate review and accountability.
The best use of AI is not necessarily to replace a person at
the end of the process.
It may be to give that person a cleaner, more complete, and
more reliable starting point.
The goal should be less administrative detective work—not
a faster detective.
The bigger idea: Build a system that learns upstream
Imagine a practice that does more than record a denial and
send it to someone for follow-up.
It identifies the failure pattern, traces the problem to the
earliest reliable evidence, assigns responsibility for the correction, and
changes the workflow so the same error is less likely to recur.
That requires a continuous information process:
Capture → Structure → Preserve → Propagate → Act → Learn
- Capture
the relevant information at the point it becomes available.
- Structure
it so that people and systems can interpret it consistently.
- Preserve
the context that gives the information meaning.
- Propagate
updates to the workflows that depend on them.
- Act
when a required element is missing, inconsistent, or outdated.
- Learn
from the outcome and improve the process.
This is the difference between processing a claim and
improving the conditions that make a clean claim possible.
It is also a more useful way to think about healthcare AI:
not merely as a tool that handles paperwork faster, but as an intelligence
layer that helps practices detect and resolve information problems earlier.
That is the direction we are exploring with ONNX 2.0—upstream
healthcare revenue-cycle intelligence designed around the realities of
physician-owned practices and the people who keep them running.
The point is not that every billing problem can be
eliminated. It cannot.
The point is that practices should not have to accept
preventable information failures as the normal price of doing business.
Frequently asked questions
1. Are most medical billing denials caused by front-end
mistakes?
Not necessarily. Denials can arise from eligibility,
authorization, documentation, coding, medical necessity, payer policy, contract
terms, processing errors, and other causes. Practices should analyze their own
denial data before deciding which causes deserve priority.
2. Should a practice focus on denial prevention or denial
management?
Both. Denial management is necessary for failures that
occur, including legitimate disputes. Prevention targets recurring, avoidable
causes before they create rework. A mature process measures each separately.
3. Can AI prevent medical billing denials?
AI may help detect missing information, inconsistencies, and
workflow risks, but it cannot guarantee that claims will be paid. Outcomes
depend on data quality, payer rules, implementation, clinical context, and
appropriate human review.
4. Will electronic prior authorization solve the problem?
It can reduce certain manual steps and improve information
exchange. It will not automatically correct inaccurate source data, resolve
every payer-policy dispute, or ensure that disconnected workflows operate as
one process.
5. Where should an independent practice begin?
Start with a sample of recent denials and rework cases.
Identify recurring causes, trace them upstream, map the relevant handoffs, and
implement one measurable improvement. Expand after confirming that the change
works.
The question healthcare should be asking
We have invested years in getting claims submitted, tracking
denials, automating follow-up, and speeding up recovery.
Those efforts matter.
But what if the next major improvement in medical billing
does not come from getting better at fixing claims after they fail?
What if it comes from making fewer avoidable failures
reach the claim in the first place?
If your practice could eliminate one recurring
administrative failure before it reached billing, which would you choose—and
where would you look for its true source?
Share your experience in the comments. If this perspective
challenges how you think about revenue-cycle management, repost it so more
practice leaders can join the conversation.
About the author
Dr. Daniel Cham is a physician and medical consultant with
expertise in medical technology consulting, healthcare management, and medical
billing. He focuses on practical insights for professionals navigating complex
challenges at the intersection of healthcare, technology, and medical practice.
He is the founder of ONNX 2.0, an initiative focused on
upstream healthcare revenue-cycle intelligence and reducing avoidable
administrative and billing problems.
Connect with Dr. Cham on LinkedIn.
Free resource: A free resource is available in the
Featured section of my LinkedIn profile—no signup required.
Continue the conversation
Connect: LinkedIn | Website | YouTube
Disclaimer
This article is for general educational and informational
purposes only. It does not constitute medical, legal, coding, reimbursement,
compliance, or financial advice. Requirements vary by payer, plan, service,
jurisdiction, and individual circumstances. Practices should consult qualified
professionals and authoritative current guidance before changing clinical,
billing, authorization, or compliance workflows.
References
1. The
Hidden Cost of Medical Billing: Why Clinics Are Losing Revenue in Silence —
Explores how fragmented workflows and administrative inefficiencies contribute
to revenue leakage in medical practices.
2. Healthcare
Saved a Stranger’s Life in Minutes. Why Does It Take Months to Fix a Broken
Claim? — Examines the contrast between rapid clinical action and slow
administrative processes, reinforcing the case for preventing recurring billing
problems.
3. Unpacking
the Medical Billing Process: From Visit to Payment — Explains the medical
billing lifecycle and provides context for how upstream information gaps can
affect downstream reimbursement.
4. CMS:
Electronic Prior Authorization — Provides official information on prior
authorization burden, estimated provider time and costs, and electronic
authorization requirements.
5. Healthcare
Dive: Prior Authorization Takeaways from Epic Ahead of the Looming API Deadline
— Discusses healthcare information-exchange challenges and industry efforts to
improve prior authorization through digital connectivity and automation.
6. CMS:
Moving Prior Authorization into the 21st Century — Outlines CMS’s efforts
to modernize prior authorization and reduce administrative friction in
healthcare.
#MedicalBilling #RevenueCycleManagement #HealthcareAI
#PhysicianEntrepreneurship #MedicalPracticeManagement #PriorAuthorization
#HealthcareOperations #ONNX2
No comments:
Post a Comment