Standing in line outside a grocery store in Seattle, almost at the front, the requisite six feet away from the person in front of me, I am beginning to panic a bit. From behind my mask, I try to hold back a cough. Pollen is thick in the air and I am convinced I can actually see the tiny dodge balls of pollen coming toward me.

My post-nasal drip and itchy sinuses are conspiring against me. The more I try not to cough, the more I feel my throat constricting and my lungs expanding as I anticipate the inevitable and hope one cough will suffice, the paroxysms will not overtake me, and I will not be forced out of line after having already waited for 20 minutes.

Coughing is a tremendously effective way to transmit SARS-CoV-2, the strain of coronavirus that causes Covid-19. Expelling droplets at speeds close to 50 miles per hour, cough is a spontaneous reflex that helps protect the airway and lungs from unwanted irritants and pathogens. It’s also an efficient way to spread disease, an advantage for respiratory viruses that need host cells in which to reproduce.


Around the world, cough is the most common reason for primary care visits and has a wide differential diagnosis which includes, but is not limited to, upper respiratory tract infections, asthma, allergies, gastroesophageal reflux, post-nasal drip, and lower respiratory tract infections such as pneumonia. And now, of course, Covid-19.

Cough is also a key component of diagnosing pneumonia in children in resource-constrained settings in low- and middle-income countries. If a child is coughing or has difficulty breathing, World Health Organization guidelines advise counting the child’s breaths and, if the rate is above age-specified cut-off points, the child should be diagnosed with pneumonia.


“Counting respiratory rate is hard to do and nobody does it,” a colleague from Lesotho reports. That is true. It’s not only notoriously difficult to do, but can be quite inaccurate, as several colleagues and I wrote in the American Journal of Respiratory and Critical Care Medicine.

“And what does counting breaths really tell you?” muses a pediatric anesthesiologist colleague who manages his patients’ breaths for a living. A single measurement of respiratory rate is unlikely to provide a clear diagnosis of pneumonia. There are many reasons why a child may be breathing faster than normal that have nothing to do with pneumonia, including other infections, exertion, anxiety, stress, and the list goes on. Even healthy children can meet the World Health Organization’s fast breathing threshold for pneumonia. Yet asking about cough or difficulty breathing and trying to count breaths is where we are with pneumonia diagnosis in resource-constrained settings.

Pneumonia is the leading infectious cause of death globally, in children and adults. It kills more children each year than HIV, tuberculosis, and malaria combined. Despite pneumonia’s global burden, the medical community is still unable to accurately and efficiently diagnose it. It remains perplexing why in this day and age — having mapped the human genome, contemplating colonizing Mars, and anticipating the development of a SARS-CoV-2 vaccine — we cannot come up with better ways to diagnose pneumonia than counting breaths in children.

On a call with a colleague who works in a pediatric emergency department in Chicago, I ask if the Covid-19 pandemic has changed how he is diagnosing pneumonia. He says that nothing has changed apart from the fact that he is now testing for Covid-19. Colleagues from Lesotho (the last country in Africa to record its first case of Covid-19, on May 13) and Malawi (first cases recorded on April 2) also note that not much has changed in their respective countries.

The challenges in diagnosing pneumonia should not be underestimated, in part because this infection can be caused by various species of bacteria, viruses, protozoa, fungi, and combinations thereof. Unlike malaria or HIV, for which rapid diagnostic tests exist that focus on detecting a specific pathogen, there is no rapid diagnostic test for pneumonia. Identifying pathogens in swabs from the nose, mouth, and throat could pick up infections limited to the upper respiratory tract. Obtaining an adequate sample of tissue or fluid from the lower respiratory tract in a child is not always feasible given that the procedure is invasive and, even when a specimen is obtained, there is no guarantee the disease-causing pathogen will be recovered.

Chest X-rays and other types of imaging aren’t ideal. Even when high-quality images are obtained and trained reviewers are able to agree on what they see, there is substantial overlap in the features of different lung diseases, resulting in potential misclassification. What’s more, imaging can’t identify the pathogen causing the problem.

So when a child has a cough or other respiratory symptoms, health care providers are often left to make a best guess as to whether the child has a pneumonia and, if so, which is the most likely pathogen given the presentation and the local epidemiology. Inevitably, pneumonias are missed, misdiagnosed, and mistreated.

Even in the United States, where pneumonia is a leading cause of hospitalization among children, with medical costs estimated at almost $1 billion (and exceeding $10 billion in adults), there is a need for new diagnostic tools. Despite the use of current comprehensive diagnostic methods in the Centers for Disease Control and Prevention’s Etiology of Pneumonia in the Community project, a large population-based study of community-acquired pneumonia in the United States, no pathogen was detected in most adult pneumonia patients. In children, multiple pathogens were identified (although primarily in the upper respiratory tract, not the lungs). These results highlight the need for sensitive, inexpensive, and rapid diagnostics to accurately detect and distinguish among pneumonia pathogens.

In resource-constrained settings, an inadequate supply or a lack of laboratory and clinical imaging capacity and, even more critically, of trained health care providers makes it difficult to diagnose pneumonia. Because of that, a child with fast breathing in the setting of a cough or difficulty breathing is said to have pneumonia and is empirically treated with an antibiotic.

This might work well if most of the infections were bacterial, but the epidemiology of pneumonia is changing and respiratory viruses are increasingly becoming the most commonly detected causes of pneumonia. Antibiotics cannot treat viral infections, and when used inappropriately can promote the emergence of antimicrobial resistance. Yet not treating bacterial pneumonia with antibiotics can result in death.

We need better pneumonia diagnostics for children and adults. Several groups are working on new ones, including biomarker panels, multimodal devices, and lung ultrasound assisted by artificial intelligence, but more innovation is needed. Pneumonia mortality is not decreasing at the same rate as mortality from other infectious diseases, largely because of decades of underinvestment, especially in improved diagnosis and treatment.

This time of Covid-19, when cough is like a perversely unreliable canary in a coal mine, should be an opportunity for us to raise awareness for pneumonia in general. By accelerating investment in the next generation of pneumonia diagnostics, Covid-19 could open the door to new ways to treat and prevent pneumonia.

Amy Sarah Ginsburg is a Seattle-based physician and epidemiologist who conducts clinical research in resource-constrained settings.

A roundup of STAT’s top stories of the day in science and medicine

Privacy Policy