Study: Physical exams alone unreliable for diagnosing childhood pneumonia

A child arrives at the emergency department with a cough and fever. The doctor listens carefully with a stethoscope, hears something concerning in the lungs, and diagnoses pneumonia. But what if a second doctor, examining the same child minutes later, heard something different—or heard nothing at all? A new national study published in JAMA Network Open in August 2026 found exactly this problem: physicians examining identical patients frequently disagreed about the lung sounds that form the backbone of pneumonia diagnosis in children.

The research, led by Todd Florin and colleagues at Ann & Robert H. Lurie Children's Hospital of Chicago and the Pediatric Emergency Care Applied Research Network, enrolled 252 children ages three months to seventeen years who had been diagnosed with pneumonia across seven pediatric emergency departments nationwide. Each child underwent physical examinations by two independent doctors within an hour of each other. The physicians recorded their findings separately, without consulting one another. What emerged was a troubling pattern: the very sounds doctors have long relied upon to identify pneumonia—crackles, decreased breath sounds, and rhonchi—were not consistently heard or interpreted the same way from one examiner to the next.

Wheezing and signs of respiratory distress were somewhat more reliably identified, but even these findings fell short of what researchers considered strong agreement between examiners. The inconsistency held true whether children were sent home or admitted to the hospital, suggesting the problem was not tied to disease severity but rather to the inherent limitations of the physical exam itself. As Shubhada Hooli, an assistant professor at Baylor College of Medicine, noted, findings traditionally considered crucial for diagnosis were being identified unpredictably across different providers.

The stakes of this diagnostic uncertainty are real. Pneumonia ranks among the most common respiratory infections in children, generating nearly two million outpatient visits and 375,000 emergency department visits annually in the United States. Current clinical guidelines recommend that doctors diagnose pneumonia in otherwise healthy children based on symptoms and physical examination alone, without routine chest X-rays, when the child can be treated at home. This approach makes sense from a practical standpoint—it avoids unnecessary radiation and keeps costs down. But if the physical exam itself is unreliable, the foundation of that entire diagnostic pathway becomes shaky.

Misdiagnosis carries consequences in both directions. A child incorrectly diagnosed with pneumonia may receive antibiotics they do not need. A child with actual pneumonia who is missed may go untreated. The antibiotic question is particularly pressing: roughly two-thirds of children treated for community-acquired pneumonia experience side effects from the medications. If doctors could more reliably distinguish which children truly have bacterial pneumonia and which do not, they could spare many children from unnecessary drug exposure while ensuring that those who genuinely need treatment receive it.

Florin and his team argue that the stethoscope, while remaining a valuable tool, cannot be the sole arbiter of pneumonia diagnosis. They point toward more objective alternatives: electronic stethoscopes that record and digitize lung sounds, computer algorithms trained to analyze those recordings, advanced imaging techniques, and laboratory tests that detect markers of infection. None of these would replace the physical exam entirely, but each could add a layer of consistency and accuracy that human interpretation alone cannot provide.

The research underscores a broader challenge in pediatric medicine—the gap between what we think we can detect at the bedside and what we actually can detect reliably. A doctor's ears, trained and experienced as they may be, are still subject to variation in perception, interpretation, and even the ambient noise of a busy emergency department. The study does not suggest that physical exams are worthless; rather, it suggests they are incomplete. Moving forward, the question is not whether doctors should stop listening to children's lungs, but how to augment that listening with tools that can provide the consistency and objectivity that diagnosis demands.