COVID-19 Coding Dilemma: Z20.828 and Z03.818

ICD-10-CM codes are not robust enough to tell the patient story correctly.  

I was having a terrible time preparing for my COVID-19 coding webinar, held on the afternoon of April 29, because the current ICD-10-CM codes under the current guidelines are not robust enough to tell the story correctly. The American Hospital Association (AHA) guidance tried to eliminate some of the confusion, but I am trying to reconcile the guidance, the guidelines, and clinical facts.

Specifically, I am talking about the codes Z03.818, Encounter for observation for suspected exposure to other biological agents ruled out, and Z20.828, Contact with and (suspected) exposure to viral communicable diseases.

The ICD-10-CM Official Guidelines for Coding and Reporting for the 2020 fiscal year say:

• C.21.a. Certain Z codes may only be used as first-listed or principal diagnosis.
• C.21.c.1) Category Z20 indicates contact with, and suspected exposure to, communicable diseases. These codes are for patients who do not show any sign or symptom of a disease but are suspected to have been exposed to it by close personal contact with an infected individual or are in an area where a disease is epidemic … Contact/exposure codes may be used as a first-listed code to explain an encounter for testing, or more commonly, as a secondary code to identify a potential risk.
• C.21.c.6)…observation Z code categories (including Z03) … are for use in very limited circumstances when a person is being observed for a suspected condition that is ruled out. The observation codes are not for use if an injury or illness or any signs or symptoms related to the suspected condition are present. In such cases, the diagnosis/symptom code is used with the corresponding external cause code.

The observation codes are to be used as principal diagnosis (first-listed) only.

Therefore, both Z03.818, Encounter for observation for suspected exposure to other biological agents ruled out, and Z20.828, Contact with and (suspected) exposure to viral communicable diseases, are stipulated by the guidelines as being only for asymptomatic patients. Z20.828 may be a principal/first-listed or a secondary diagnosis, whereas Z03.818 is specified as only being allowable as a principal or first-listed diagnosis.

There are several issues that arise from these instructions and prohibitions. The AHA COVID-19 coding FAQs, revised April 16, added to my confusion (https://www.codingclinicadvisor.com/faqs-icd-10-cm-coding-covid-19).

This guidance states that Z03.818 should be used “if the patient is asymptomatic and there is possible exposure to COVID-19” and the test is negative. It also says that Z20.828 is used for a known or suspected COVID-19 exposure, if they have signs and symptoms consistent with COVID-19, and the test is negative, inconclusive, or unknown (e.g., pending). They base this on the guideline that “Z20 codes may be used for patients who are in an area where a disease is epidemic.” Does this guidance mean that the suspected COVID-19 exposure must be in a patient with signs and symptoms, or does it mean suspected COVID-19 exposure or if the patient has signs and symptoms?

I think it will be easier to understand the gap between the advice and the codes if we look at another disease, anthrax, and its coding.

In 2001, there were anthrax exposures, some of which led to fatal infections.

• Bob Stevens died of inhalation anthrax. His code was: A22.1, Pulmonary anthrax.
• A postal worker who was exposed to anthrax via mail would be given prophylactic antibiotics to prevent the development of the disease. Their code: Z20.810, Contact with and (suspected) exposure to anthrax.
• A nervous office worker who opened up a letter with white powder enclosed, but, after work-up, the Centers for Disease Control and Prevention (CDC) and the FBI determined that it was only powdered sugar. There was no exposure to anthrax, hence: Z03.810, Encounter for observation for suspected exposure to anthrax ruled out.

For my anthrax example, neither the postal worker nor the office worker was symptomatic. The difference between the two was whether or not there was exposure to anthrax. If the postal worker had a cough from acute bronchitis that was unrelated to anthrax, Z20.810 would still apply.

Let’s turn back to the coronavirus. If the PCR test is positive, the patient has COVID-19, with or without symptoms. If the test is negative, the patient may either have COVID-19, because there is a 30-percent false negative rate, or they may not have the disease. The provider must use their clinical judgment in determining which case they think applies, and must document it accordingly.

The AHA guidance implies that the distinction between Z03.818 and Z20.828 is whether the patient has symptoms or not. This is simply not correct, clinically speaking. A patient might have an adenovirus upper respiratory infection and seek medical attention because they have anxiety about whether it is COVID-19 without any known exposure. The diagnosis should be J06.9, Acute upper respiratory infection, unspecified, plus the code that best describes whether or not the provider believed there had been legitimate exposure to COVID-19.

For normal coding, we use the manifestation or signs/symptoms as the only codes. We don’t add Z20.828 if the patient reports that their sister has had a cold recently and sneezed on them. The issue with COVID-19, in particular, is that we are trying to surveil infections, exposures, and testing, and for that, we need a code to flag the encounter as involving COVID-19.

The AHA advice also states that “due to the current COVID-19 pandemic, when a patient presents with signs/symptoms associated with COVID-19 and is tested for the virus because the provider suspects the patient may have COVID-19, code Z20.828 may be assigned without explicit documentation of exposure or suspected exposure.” This is also not quite accurate.

First, the provider may be ordering the test to rule out the disease, precisely to prove that patient does not have it. Secondly, our country is wildly variable in how prevalent the disease is. Since there is insufficient testing and we have no idea what the extent of asymptomatic virus shedding exists, we have no idea what the likelihood of exposure is in an area not considered to be a surge site. The prevalence of COVID-19 in Westchester County, New York is not equivalent to the prevalence in Eddy County, North Dakota. A provider in New Rockford, North Dakota may still choose to order a COVID-19 test on their patient.

Different factors enter into whether we clinicians think there has been exposure. Did that patient with the cold in New Rockford have a cousin from New Rochelle, New York come stay with them? Is she a respiratory therapist or an emergency physician who drives an hour to work in an urban hospital setting? Has a patient in Westchester been quarantined in their room for 21 days and had no exposure to anyone even potentially with the virus? All the circumstances must be considered to decide whether there was exposure, potential exposure, or whether exposure can be effectively ruled out.

If you ruled out exposure, why did you do the test? Sometimes, you want the negative test to be able to declare that the exposure was also ruled out. It is too important to get this right, and doctors don’t like to be wrong.

Why can’t we use the code, Z11.59, Encounter for screening for other viral diseases?

Let’s use as our example a patient presenting for elective or emergency surgery. The provider or institution may want to know if they have COVID-19, because that may change their approach or the patient’s physical bed location. The patient would be tested to see if they have the virus.

This is not a “screening” test, per se. Screening is done to a population, not to a targeted specific patient. This admitted patient must have a code that justifies medical necessity for the PCR test. They may have symptoms unrelated to COVID-19, but they have no known exposure. The diagnosis that would support the PCR test needs to be related to COVID-19 signs, symptoms, or exposure. For this surgical patient, that diagnosis should be Z03.818, but as it stands right now, the scrubber would reject the claim. Z03.818 must be principal, and that would make no sense in this hypothetical scenario. The principal diagnosis and principal procedure should be related to the operative intervention. You shouldn’t be using the screening code because this is a diagnostic study, not a screening one.

In my opinion, symptoms should not be the distinguishing factor between Z03.818 and Z20.828. Z20.828 means that a patient is deemed to have had exposure or suspected exposure to COVID-19. Z03.818 means the likelihood of exposure to COVID-19 is felt to be low or nil, but the provider still believes that a work-up is indicated. Both codes should be able to be used with codes signifying symptoms that elicited the investigation, if any. Both should be permitted to be principal, first-listed, or secondary codes.

However, until the guidelines are hopefully changed, you should follow the AHA guidance. I think the only solution during the time of the pandemic is to use Z20.828 for any admitted patient who is tested for COVID-19, regardless of known or suspected exposure, and the test returns negative. If the test is pending at discharge, we are instructed to hold the bill until the results are back. If it returns positive, the coder should pick up U07.1 plus manifestations. If it returns negative and the provider made an uncertain diagnosis, consider whether a query should be generated. If not, the diagnoses would be the signs/symptoms/manifestations, plus Z20.828.

For asymptomatic outpatients, choose Z03.818 or Z20.828, depending on whether there was contact with or suspected exposure to COVID-19. For symptomatic outpatients who test non-positive (i.e., negative or pending), use Z20.828, regardless of contact status.

I hope they fix the guidelines soon. I am praying that this pandemic resolves, but COVID-19 isn’t going to disappear. It is going to be with us forever. Hopefully, there will be a vaccine and effective treatment in the future, but using Z20.828, Contact with and (suspected) exposure to viral communicable diseases, isn’t going to be the long-term solution. Clinicians are going to need to be able to test symptomatic patients even if exposure is ruled out, and we will need a way to record that in codes.

Programming Note: Listen to Dr. Erica Remer every Tuesday on Talk Ten Tuesdays, 10-10:30 a.m. EST.