Supporting Medical Necessity of Investigative Tests

The aim of documentation is to tell the true story.

Last year, I wrote a two-part series on clinical documentation integrity, specifically focusing on the emergency department (CDI in ED Part I; CDI in ED Part II). Today I will focus on how ED documentation supports medical necessity of investigative tests.

This article was prompted by a scenario posed on RAC Relief (a Google group discussion forum created to allow physician advisors to connect and ask each other questions). The situation was a patient presenting to the ED with shortness of breath (SOB), diagnosed with “anxiety” after work-up, including a B-type (brain) natriuretic peptide (BNP) level, presumably to explore heart failure as an etiology. The question was whether it is permissible to code the symptom of SOB in addition to the final diagnosis. The underlying issue is that the diagnosis of anxiety does not justify medical necessity for the BNP test.

All studies must have an associated, valid ICD-10-CM code that payers acknowledge as justification for that test. There may be some variability, depending on the payor and the geographic locale, although there is usually significant overlap of LCD/NCD (local/nation coverage determination) lists.

For our BNP example, acceptable diagnoses include hypertensive heart disease with heart failure; multiple cardiovascular conditions, including myocardial infarctions, atherosclerosis, cardiomyopathies, and all types of heart failure; and then symptoms like dyspnea, orthopnea, SOB (shortness of breath is subjective, while dyspnea is when the healthcare provider objectively perceives that the patient is having difficulty breathing). These can also include wheezing, tachypnea, and unspecified abnormalities of breathing.

Anxiety is not on the list, nor should it be. There is no need to measure a BNP level in a patient known to be presenting due to anxiety.

However, we often do not have the luxury of knowing the final or definitive diagnosis at the onset of an emergency department visit. The only things we have to go on are the patient’s complaints, their physical findings, and our clinical experience and judgment. The impetus for the BNP in this scenario was the symptom of shortness of breath.

The question posed was whether it was allowable to have SOB as a secondary diagnosis.

According to I.B.4. of the ICD-10-CM Official Guidelines for Coding and Reporting for FY 2019, if a “related definitive diagnosis has not been established,” codes describing symptoms and signs are acceptable. I.B.5. prohibits the inclusion of signs and symptoms that are “associated routinely with a disease process” as additional codes, and I.B.6. instructs us to use additional signs and symptoms diagnoses if they “may not be associated routinely with a disease process.”

In this case, we need to consider if shortness of breath is “routinely associated” with anxiety. Shortness of breath is not uncommon in panic attacks, but it is less commonly noted to be a component of generalized anxiety disorder. It is certainly not integral to the disease process, like dyspnea is integral to an exacerbation of chronic obstructive pulmonary disease (COPD) or heart failure.

In the ED, our mantra is to “rule out life threats.” If we knew conclusively from the beginning that the patient was having an episode of acute anxiety with symptoms of sympathetic nervous system activation, we would not pursue other diagnoses. But in this patient, the symptom of SOB elicited specific tests to rule out serious cardiopulmonary pathology.

There are several other factors to consider. Emergency physicians often juggle multiple patients simultaneously and are disinclined to list multiple diagnoses due to the time crunch. They also strive to find a single diagnosis that explains all of the symptoms and findings in a nice, neat package. Their impression list is often constrained by their electronic health record’s (EHR’s) available choices. They also operate in an outpatient environment in which uncertain diagnoses are not codable; only the inpatient technical side can capture uncertain diagnoses.

Emergency physicians (EPs) need to be taught to list all conditions, symptoms, and diagnoses relevant to each encounter. The fact that the patient is on chronic steroids for her well-controlled asthma may impact on the testing performed. Having “the worst headache of her life,” with what turns out to be a viral syndrome, may elicit a CT scan of the brain. This runs counter to our training to try to convert all signs and symptoms into a single unified disease process.

I recommend linking symptoms with definitive or uncertain diagnoses to explain one’s thought process. This may bolster the complexity of medical decision-making. If the EHR mandates selection of a diagnosis with an ICD-10-CM code, then put uncertain diagnoses in the ED course narrative paragraph (e.g., chest pain and shortness of breath, rule out acute pulmonary embolism). If you aren’t sure whether the symptoms or signs deserve their own listing, put them in the ED course (e.g., CT scan was performed for “worst headache of life,” and was negative for bleed).

Let me give you one more example. Say a patient presents to the ED for a sprained ankle and is noted to have an irregular heartbeat when their vital signs are checked in triage. They get an electrocardiogram, which reveals premature ventricular contractions (PVCs). My diagnosis list is the following:

  1. Right ankle sprain
  2. Premature ventricular contractions
  3. Accidental fall over uneven sidewalk

Let’s say the same patient turned out to be in heart failure from atrial fibrillation with rapid ventricular response, which caused them to fall and sprain their ankle. Their diagnosis list would then look like this:

  1. New onset atrial fibrillation with rapid ventricular rate or response (RVR) and acute heart failure (type to be determined)
  2. Dizziness due to No. 1
  3. Right ankle sprain
  4. Accidental fall over uneven sidewalk


I instruct EPs to put what they consider the most important diagnosis first. It is usually related to the chief complaint, but it may instead be the reason for admission or observation. Then I advise them to have a diagnosis or symptom to go with every procedure. You do not need a right ankle series for afib with RVR. You don’t need an EKG for a right ankle sprain.

Can’t the coder just cull the codes from the ED H&P without the provider having to formally list them? Yes; if the documentation supports the code and the diagnosis supports the medical necessity of the test/procedure, it does not have to be expressly listed in the final impression section. However, would there have been any question in RAC Relief if the diagnosis list had appeared like this? For example:

  1. Acute anxiety attack (which is F41.9, Anxiety disorder, unspecified. Personally, I probably would actually have diagnosed an acute panic attack, F41.0)
  2. Shortness of breath (R06.02)

My documentation practice went like this:

  1. Reason for admission, observation, or most important diagnosis first
  2. Additional signs, symptoms, or findings not routinely associated with the principal/first-listed diagnosis, and which explain why you did what you did and support the medical necessity of testing
  3. Diagnosis/es (could be only sign/symptom) related to chief complaint, if not already present above
  4. Comorbid conditions that might be relevant to today’s visit, including long-term medication use
  5. Pertinent personal or family historical conditions
  6. Social determinants and external causes codes

This constellation of impressions, diagnoses, and conditions usually tells the story of the ED encounter. And that is the aim of documentation: tell the story. Tell the truth.


Program Note:

Listen to Dr. Remer every Tuesday on Talk Ten Tuesday, 10 a.m. ET.

Comment on this article


Erica E. Remer, MD, CCDS

Erica Remer, MD, FACEP, CCDS, has a unique perspective as a practicing emergency physician for 25 years, with extensive coding, CDI, and ICD-10 expertise. As physician advisor for University Hospitals Health System in Cleveland, Ohio for four years, she trained 2,700 providers in ICD-10, closed hundreds of queries, fought numerous DRG clinical determination and medical necessity denials, and educated CDI specialists and healthcare providers with engaging, case-based presentations. She transitioned to independent consulting in July 2016. Dr. Remer is a member of the ICD10monitor editorial board and is the co-host on the popular Talk Ten Tuesdays weekly, live Internet radio broadcasts.

Related Stories

Measuring CDI Performance: A Truthful Conclusion

Measuring CDI Performance: A Truthful Conclusion

Task-based, outcomes measurement versus process improvement generally does not support sustainable long-term results. I was recently asked by a chief financial officer (CFO) what other

Read More

Leave a Reply

Your Name(Required)
Your Email(Required)

Featured Webcasts

Implantable Medical Device Credit Reporting for 2023 – What You Need to Know

Learn how to save your facility hundreds of thousands of dollars in repayments and fines by correctly following CMS requirements for implantable medical device credit reporting. We provide you with all the need-to-know protocols, along with the steps for correct compliance while offering best practices to implement a viable strategy in your facility.

January 25, 2023

Trending News