Journal Logo

Breaking News

Read breaking news and articles published ahead of print. Tell us what you think and comment on your colleagues’ views about EMN articles.

Monday, March 23, 2020

BREAKING NEWS: Clinical Acumen and COVID-19

There Won't Be Enough COVID-19 Tests,
So Use Clinical Symptoms as a Guide


The availability of COVID-19 testing is held as the linchpin for many of us waiting to know how to best protect our community, but we must be careful about thinking of COVID-19 testing as our only good option.

What do we do before testing becomes widely available in our communities? What if you practice in a remote or rural area where COVID-19 tests may not be readily available? What if multiple companies produce COVID-19 tests to meet demand, but they haven’t been sufficiently vetted or they simply have different testing characteristics that make them highly variable from each other and end up being confusing rather than clarifying?

The validity of the test will depend on the clinical pretest probability. What if we fail to make the distinction between screening asymptomatic people and using that as a piece of the picture for patients who are clearly sick?

What happens if we are testing, but some link in the chain breaks? We could run out of reagent, transportation lines could be disrupted, reporting could become significantly delayed, HIPAA violations could prevent sending results by phone, and the economic crisis could prevent having a phone or paying for a cellular plan.

What happens if those in drive-by testing spots are not trained well or not able to obtain a proper nasal specimen? Do we allow people to self-swab? Can the test be operator-dependent?

What about reinfection? Will the test lose some of its testing characteristics if the viral load is less? How many of the “positive” COVID-19 tests are asymptomatic carriers? What about false-positives? Eighteen percent of COVID-19-positive patients on the cruise ships were asymptomatic. Did they actually have the virus? Can a single “negative” test be trusted in a sick patient? The PCR measures viral genomic material, but doesn’t necessarily indicate contagiousness.

The CDC said a patient who has tested positive must have two consecutive negative COVID-19 tests to be considered negative because negative RT-PCR does not exclude COVID-19, especially early in the course. (Read a brilliant review of the statistics, including false-negatives, by Andy Chen in Towards Data Science. Feb. 15, 2020;

The amount of resources eaten up to retest results we don’t believe will be untenable. The increased testing, especially if we screen asymptomatic people driving through testing sites, may actually increase confusion. Some are calling for testing only hospitalized patients, select frontline staff, and certain households and institutions with very high-risk residents. (Figuring out this last one out will be difficult.) We should be recommending less COVID-19 testing instead of more, or at least targeted testing instead of mass screening.

Some of these questions and many others are critical to consider. What should we be doing in the ED if COVID-19 testing becomes less helpful than we are hoping? What should we do then? We should do what we have always done best: Use our bedside clinical acumen to know who to test and how to use the test.

Now, more than ever in our entire careers, we must not get swept up in blindly following someone’s protocol and turning off our brains in the process, especially when the “protocols” are changing every day. They are also not consistent from hospital to hospital or from clinics to minor emergency clinics. The protocols are not even consistent from the CDC to state health departments to local health agencies to the corporate rollouts that own the hospitals.

Everything, at least in the ED, begins with a triage decision based on the likelihood of disease, and that is not an automatic one-way door. It is like peeling an onion. The ED health professional back in the room is another layer that does triage (or sorting), again by using a detailed history and physical to make a differential diagnosis with high and low probability. We use clinical decision-making that may or may not require testing. We are the best prepared in this crisis to know how to drive in this lane.

I am not a virologist, an infectious disease specialist, or a pulmonologist; I am an average EP in Kansas. I am not offering a scholarly contribution as much as I am offering an emergency toolkit to break out in emergencies. A few clinical things that we and our patients should know mostly come from a single study (New Engl J Med. 2020 Feb 28.;

Many of the symptoms attributed to COVID-19 are rare:

  • Conjunctival involvement: <1%
  • Rash: <1%
  • Cervical lymphadenopathy: <1%
  • Diarrhea: <4%
  • Nasal congestion and runny nose: <5%
  • Nausea and vomiting: 5%

It should be mentioned that the American College of Gastroenterology just reported a small study from China in which diarrhea was present in 29 percent of 204 patients. (Am J Gastroenterol. 20 March 2020; (Someone smarter than me will have to tell us how to compare 29 percent of 204 compared with four percent of 1099.)

“Fever” was initially present in less than half of patients (44%), so its use as a dichotomous screening tool will fail. A total of 89 percent did end up having a fever at some time during their hospitalization, but the fever, contrary to hearsay, was not necessarily high—the median axillary temperature was 37.3°C.

Respiratory symptoms must be clearly defined to include cough and difficulty breathing as well as sore throat. This is important because not everyone had cough (68% did).

You shouldn’t be doing a strep test on everyone. Only 15 percent presented with a sore throat.

COVID-19 in most patients should still look like the flu. Severe malaise was present in 38 percent of patients. Myalgias were seen in 15 percent, particularly in the legs in children and in the back in adults, and were seen in extraocular muscles, which may present with pain behind the eyes or photophobia and headache (14%). (Iran J Med Sci. 2017;42[1]:2;

Use doxycycline, not azithromycin, if a patient has COPD or smoker’s bronchitis and a change in sputum production (34% with COVID-19). Secondary bacterial pneumonia in the 1957-1958 U.S. influenza epidemic was two to 18 percent. Staphylococcus aureus was a primary pathogen in the 1968-1969 influenza epidemic. Doxycycline will treat strep and staph, including MRSA; azithromycin won’t.

Don’t use steroids if you can avoid it. During the 2009 influenza pandemic, 35-55 percent of patients admitted to the ICU in Europe had been placed on steroids, and that may have been related to increased mortality; it was a retrospective study, which may suffer from acuity bias. (J Infect Dis. 2015;212[2]:183;

Use this critical time not just to mention but actually teach smoking cessation and about secondhand smoke exposure. If ever there was a time, it is now. It may have played a major factor in morbidity and mortality throughout the world. If you don’t know how to teach smoking cessation, you can learn it or hand out resources that can do it for you.

Let’s define high risk and low risk, and make a distinction between infection and mortality. The elderly are not at high risk for infection, but are at high risk for death if exposed. Young children are at high risk for infection but very low risk of death (0 in the 0-14 age category in the China study). Health care workers are also at high risk for infection but very low risk for death (0.3%) Pregnant women are apparently at very low risk of death. (This differs from some of our influenza outbreaks.)

Chest x-rays are not diagnostic. Eighteen percent of hospitalized COVID-19 patients had no radiographic abnormalities, but 95 percent of critically ill patients did. You are probably going to get a chest x-ray on them anyway.

Most of this information comes from one study in China of 1099 hospitalized COVID-19 patients, a select cohort from 7736 patients. Some of these percentages will likely change. Don’t sue me; we all know this is a moving target, and some of this will need to be tweaked even minutes after it is published!

Let’s use the scientific data we have. Let’s use the clinical expertise we have acquired throughout our professional training. Let’s not allow some uncertainty to lead to irrational “protocols” on a kneejerk-triaged patient. But there are some recommendations we cannot do enough: Wash your hands, and make those 20 seconds count.

Dr. Mosley is an emergency physician in Wichita, KS.