Invited Commentary 

Less Is More

Acute Respiratory Infections in Primary Care—Precision Microbiology Fails to Bring Order to Diagnostic Disarray

Jeffrey A. Linder, Julia E. Szymczak

JAMA Intern Med Published Online: May 18, 2026

doi: 10.1001/jamainternmed.2026.1440

Collectively, acute respiratory infections are the most common symptomatic reasons for seeking ambulatory care, placing a major strain on health care systems. In many high-income countries, about half of patients who present with respiratory symptoms in ambulatory care receive antibiotics, but only about 50% of these prescriptions are appropriate.^1,2 Inappropriate prescribing can lead to serious complications, such as infections with Clostridioides difficile, adverse effects of antibiotics (eg, gastrointestinal, kidney), allergic reactions, secondary infections, and the emergence of antibiotic resistance.

Despite these risks, the belief that antibiotics might help and cannot hurt is a well-recognized driver of antibiotic overprescribing for acute respiratory infections in primary care settings.³ In addition, when patients have a history of receiving antibiotics to treat self-limited acute respiratory infections, it can be difficult for clinicians to counter patients’ belief that the antibiotics have helped in the past.⁴ To reduce inappropriate prescribing, precision microbiology has the potential to address the problem: if clinicians could reliably distinguish viral from bacterial infection and show patients what was causing their illness, prescribing would align with microbiology.

In this issue of JAMA Internal Medicine, Hay and colleagues⁵ describe a randomized clinical trial of the use of a rapid multiplex microbiological point-of-care test (RM-POCT) that detects 19 respiratory viruses and 4 atypical bacteria. The participants were primary care patients at 16 general practices in Southwest England 12 months or older with a clinician-diagnosed acute respiratory infection diagnosis for which the clinician or the patient thought antibiotics were, or might be, necessary. Control patients had a multiplex microbiological swab sample collected, but the results were not made available at the time treatment decisions were made. For intervention patients, clinicians were given RM-POCT results within about 45 minutes. On microbiologic testing of the 552 study patients, the control and intervention patients were nearly identical: about 51% had no organism detected, 42% had 1 or more viruses, 5% had atypical bacteria only, and 2% had a virus and atypical bacteria.

Although the investigators hoped for reduced antibiotic prescribing in the RM-POCT group, the antibiotic prescribing rate in the control and intervention groups was numerically identical: 124 of 276 (45%). In the control group, there was no difference in antibiotic prescribing between patients who did and did not have a virus subsequently detected (46% and 45%, respectively). In the intervention group, a decrease in antibiotic prescribing among patients with a virus detected was perfectly counterbalanced by an increase in antibiotic prescribing to patients without a virus detected (22% and 61%, respectively). So, testing was clearly associated with prescribing behavior, but why not in the way hoped for?

First, applying precision microbiology to a disarray of diagnoses is unlikely to bring clarity. Acute respiratory infections are heterogeneous and, in the study by Hay and colleagues,⁵ the clinician-diagnosed acute respiratory infections were organized in 14 disparate categories. Some diagnoses were clearly non–antibiotic-appropriate, like the common cold and acute bronchitis, for which no testing is necessary, or influenza and COVID-19, for which dedicated testing may have already been administered. Some of the diagnoses were potentially antibiotic-appropriate, like acute pharyngitis, for which focused clinical scores and specific testing identify patients likely infected with group A Streptococcus. Other diagnoses were nonspecific symptom syndromes or collections of other diagnoses, like acute cough, chest infection, or acute lower respiratory tract infection. The principles of diagnostic stewardship, which prioritize the right test, for the right patient, for the right diagnosis, to prompt the right action, dictate more judicious application of testing.⁶

Second, testing can compel action even if unwarranted. Once a diagnostic test is performed, a clinician is no longer just responding to a patient’s symptoms; they are responding to a result. Intervention clinicians might have increased antibiotic prescribing to patients without a virus detected because they interpreted such a result as indicating the possibility of an infection with other common, potentially pathogenic bacteria. The orientation materials provided to clinicians stated “[the multiplex test] does not test for the typical respiratory bacteria S. pneumoniae, S. pyogenes, H. influenzae, or M. catarrhalis since these can be commensally carried in the upper respiratory tract.” ⁵ Thus, clinicians in the intervention group, when receiving a virus-not-detected result, may have been confronted with the thought that they could be missing an infection with, for example, Streptococcus pneumoniae. The misapplication or overuse of microbiological tests can lead to overdiagnosis, unnecessary antibiotic use, excess cost, and, as may have happened in the trial by Hay and colleagues⁵ when no virus was detected, increased uncertainty.

Third, clinical encounters are not simply neutral diagnostic exercises: they are social encounters shaped by dynamics that have nothing to do with microbiology. In seeking care, the patient has determined that their symptoms warrant medical attention. Patients may expect (and clinicians may think patients expect) a tangible intervention in exchange for the time and money spent on consultation. Clinicians may anticipate patient dissatisfaction, leading to complaints, if no treatment is offered.⁷ Time constraints around the visit disincentivize discussion about why antibiotics are not needed. Improved diagnostic precision does not neutralize emotional, interpersonal, and organizational pressures that promote antibiotic use.

Another consideration raised by the study by Hay and colleagues⁵ is whether antibiotics help patients at all. For safety reasons, the investigators assessed clinical outcomes to make sure that a hoped for reduction in antibiotic use was not associated with an increase in complications. It was also possible that better targeting of antibiotics, away from patients with a virus detected to those without, might also have the potential to improve clinical outcomes by treating bacterial infections not captured by the study test.

It did not. There were no differences between the intervention and control patients in symptom resolution, worsening symptoms, or return to usual activity. The investigators even stratified day 2 to 4 symptom resolution between the control and intervention groups by virus detection, and there were no differences within or among groups. This trial suggests that changing the target of antibiotic prescribing within a trial of primary care–diagnosed acute respiratory infections does nothing to improve patient outcomes.

Improving antibiotic use in ambulatory settings has been a focus of research and public health initiatives for more than 30 years. Although much has been made about the promise of precision microbiology, as Hay and colleagues⁵ demonstrate, we are unlikely to test our way out of inappropriate antibiotic prescribing. Precision microbiology will not improve antibiotic prescribing or patient outcomes when applied to heterogeneous diagnoses for which antibiotics are mostly ineffective. Precision microbiology will not address the complex social, emotional, and organizational factors that are associated with care seeking and most inappropriate antibiotic prescribing.

Rather than technology, clinicians, health systems, policymakers, and researchers should attend to the many factors promoting antibiotic overuse. Multicomponent interventions informed by behavioral science, targeted at improving how clinicians communicate, and nudging them to make evidence-based decisions, are effective.^4,7Future research is needed to identify novel strategies to help patients with self-limited conditions manage their symptoms without resorting to an unhelpful clinical encounter, expensive diagnostic testing, and antibiotic prescriptions that will not change the trajectory of their illness.