Documentation of tests and symptoms needs to be robust in the ER in order to mitigate audit fallout.

Featured speakers during a recent Monitor Mondays broadcast said it’s open season for audits on providers – and as we know, denials are always coupled with those audits. Notably, denials are increasing across the healthcare industry, and the emergency department (ED) is a prime target. Identifying the “what” and “who” will allow you to focus on triggers for the denials. 

What we’re seeing denied or adjusted by the payers are the higher ED levels of 4 and 5 (for both ED physicians as well as facilities), and certain services that allegedly could have been provided in another outpatient setting. Earlier this year, UnitedHealthcare was proposing to make denial determinations after the ED encounter, and if the visit was considered non-emergent, the claim would be subject to “no coverage or limited coverage.” After industry reaction, UnitedHealthcare deferred the plan. Aetna tried a similar practice a few years ago. The fact is that payers are not in the ED when the patient presents. Hindsight is always 20-20, but hindsight is not available when the ED providers are using their clinical and cognitive skills, analyzing symptoms, and rendering the actual treatment.

ED Levels

The April 2000 Federal Register introducing the Outpatient Prospective Payment System (OPPS) required healthcare facilities to create their own leveling approaches. It said an approach will be acceptable “as long as the services furnished are documented and medically necessary, and the facility is following its own system, which reasonably relates the intensity of hospital resources to the different levels … and CMS (the Centers for Medicare & Medicaid Services) would not expect to see a high degree of correlation between the code reported by the physician and that reported by the facility.” That last clause tells us that the facility level cannot be identical to the physician’s level, at least not all the time. Furthermore, the flexibility CMS afforded us leaves each organization open to defending its approach.

Meanwhile, payers have also defined their own leveling criteria, and we should anticipate that those payer-created criteria will collapse some of the services that would bump your visits up under your criteria. This is why I recommend to healthcare organizations the use of a facility leveling criteria developed by an authoritative entity, such as the American College of Emergency Physicians (ACEP:

The level of service billed must be based on the intervention(s) performed in relationship to the medical care required by presenting symptoms, resulting in diagnosis of the patient. This statement came from a Blue Cross plan that published its ED leveling rules.  Interestingly, it appears that this plan is using ACEP’s criteria, or a version thereof. However, the clause “resulting in diagnosis” hints to the fact that tests performed when results do not become available until after the patient is discharged may not be considered by the payer as medically necessary to contribute to making the diagnosis. Therefore, it’s important for ED physicians/providers to link test results to their impressions.

Professional (physician/provider) level codes are based on complexity of work performed, including the cognitive effort. Facility codes reflect volume and intensity of resources used by the facility to provide care.

Service Denials

For denial of services rendered during the ED encounter, this needs to be monitored closely. Remember, the payer has the advantage of hindsight. Computed tomography (CT) scans are a target because of their cost, as well as resource utilization. It’s an easy CPT code to monitor, across providers and nationwide. It’s also easy to correlate the CPT with the ED diagnoses. We’ve seen CTs documented as one service often denied, because the record lacks documentation to support ordering a CT, such as a reason that correlates to the LCD (a fall resulting in hitting the head, severe headache, etc.) It requires assessing whether the services were medically necessary and if testing results were made available during the ED encounter to help the provider make a treatment decision. This assessment may require the assistance of your physician advisors. 

Don’t be surprised if you find system issues when you’re investigating these test-related denials. At one organization, we found that the order “reason” area was too short and truncated the entry by the providers, and therefore, it was not being captured on the final report as well. Regardless, the cost of CT denials can be sizeable, so ensuring that providers note the conditions or symptoms for which they are ordering an exam is a worthwhile effort.

Artificial Intelligence Deployed

Recognizing the drivers of the levels, payers today are using artificial intelligence (AI) programmed to analyze both the electronically received records and claims for patterns, such as:

  • Whether services/tests ordered are consistent with the diagnoses reported on the claims;
  • Whether the services performed required facility resources;
  • Whether the test results were used to make the diagnosis;
  • Whether the levels assigned are consistent with the diagnoses reported, tests ordered, and treatment provided;
  • Whether services (that is, the CPTs) vary by physician/provider National Provider Identifier (NPI); and
  • The bell curve of the levels, sorted by NPI.

A recent RACmonitor article said some level 4 and 5 visits have conditions that are typically not denied, including when:

  • The patient was admitted to inpatient or observation, received critical care, or had outpatient surgery during the ED visit;
  • The patient was under 2 years old; and/or
  • The patient expired in the ED.

Collecting Data

So the logical next step is to trend denials by payer and ED physician. The ED physicians may need to establish or refine treatment and documentation protocols. If the denials are for facility levels, then compare your criteria to ACEP’s, or other commonly used criteria, to determine if yours are out of sync.

Similar to all types of encounters, documentation is paramount. For the ED, providers need to document conditions they considered, including chronic conditions that complicate the patient’s care. Linking those conditions to the treatment provided and the services and testing ordered helps support the levels claimed by both the provider and the facility. We need thorough documentation to avoid denials. Recognize that often, narrative descriptions carry more weight than checklists when payers review the records.   


If your organization is experiencing denials for the entire ED encounter, it may be due to something other than clinically related issues. Similar to other patient care settings, other factors that require attention in the ED include collecting accurate patient demographic and insurance information. This also means that insurance verification should take place for each visit.

A front-end question and triage assessment should determine if the patient is presenting with a liability or work-related problem or injury. When this is the case, billing the patient’s health insurance payer rather than workers’ comp or employer could result in a denial. Additionally, workers’ comp may have pre-defined providers and procedures. Similarly, patient care provided for a motor vehicle accident billed to the patient’s health insurer rather than the auto insurer may trigger a denial.

Keep in mind that it is important to know your state’s regulations for payer adjudication of ED claims. Most states have regulations that protect ED claims, because going to the ED is typically a patient’s choice. 

Lastly, defend your criteria, and don’t hesitate to quote the Federal Register listing that allowed you to establish it.

One additional note: just as payers are using artificial intelligence, healthcare providers can install these applications to validate insurance eligibility and retrieve insurance information, such as type of coverage, eligibility information, co-pays, and services that might not be covered. Additionally, such software can be focused to identify non-specific or outdated codes, unbundled charges, and medical necessity for the services provided versus the diagnoses associated with the orders for the services – as well as whether the treatment is clinically appropriate for the setting.

There are a number of AI firms in the revenue cycle arena offering this software. It’s probably worth your time to look at a few demos.

Share This Article