The study examined 384 nasopharyngeal samples collected from symptomatic patients between March 2021 and August 2022, during the COVID-19 pandemic. Every specimen was tested using simplex reverse-transcription polymerase chain reaction, or RT-PCR, for SARS-CoV-2 and with the BioFire FilmArray Respiratory Panel 2.1, commonly called BioFire FA-RP2.1. Unlike a conventional assay designed to detect one pathogen or one small group of targets, the automated panel simultaneously searches for multiple respiratory agents, including SARS-CoV-2 and viruses such as rhinovirus. The researchers also used targeted next-generation sequencing to investigate samples in which the two primary tests disagreed.

The distinction between these methods is important. RT-PCR detects genetic material by repeatedly copying selected sequences until they become measurable, while multiplex panels use a cartridge-based system to test several pathogen targets in parallel. A result is often expressed through a cycle-threshold, or Ct, value: the more amplification cycles required before a signal appears, the less viral material was present in the original sample. High Ct values therefore generally indicate a low viral burden, although the precise interpretation depends on the assay, sample quality and timing of infection. Such borderline samples are particularly vulnerable to differences in test chemistry, sampling and detection thresholds.

The two methods did not produce identical results. Simplex RT-PCR identified 163 SARS-CoV-2-positive samples, whereas BioFire FA-RP2.1 detected 141. The multiplex system nevertheless identified 192 mono-infections and 20 co-infections across the respiratory panel. SARS-CoV-2 was the most frequent single infection, accounting for 129 cases, followed by rhinovirus with 48. The findings suggest that a patient presenting with what appears to be a typical COVID-like illness may instead have another respiratory virus—or may be carrying more than one pathogen at the same time.

Overall, the researchers identified 61 discrepant SARS-CoV-2 results. In 39 cases, or 64 percent of the disagreements, RT-PCR was positive while BioFire was negative. In the remaining 22 cases, BioFire was positive while the targeted RT-PCR result was negative. The discrepancies represented about 16 percent of the tested samples, a level large enough to matter for clinical care, infection-control decisions and surveillance systems. Yet disagreement did not necessarily mean that one test was simply wrong. Instead, it exposed the difficulty of detecting small quantities of viral material in complex biological samples.

The discordant specimens tended to contain less detectable SARS-CoV-2 genetic material. Their mean Ct value was 30.9, compared with 27.4 among samples in which results were more consistent, a statistically significant difference reported by the investigators. Co-infection also increased the likelihood of disagreement: samples containing multiple pathogens had a risk ratio of 2.1 for discordant SARS-CoV-2 detection, with a 95 percent confidence interval from 1.36 to 3.30. Several mechanisms could contribute. One virus might dominate the sample and interfere with extraction or amplification, different assays might target genetic regions present in unequal amounts, or the patient might be sampled at a transitional stage of infection when viral concentrations are near the limit of detection.

To determine which results reflected genuine infection, the team applied targeted next-generation sequencing to 23 discrepant samples. Sequencing confirmed SARS-CoV-2 in 22 of them. This approach does not merely ask whether a broad fluorescent signal is present; it reads pathogen-specific genetic fragments and can provide an independent molecular line of evidence. The sequencing results underscore the value of using complementary techniques when tests disagree, particularly in surveillance programs where a false-negative result could obscure ongoing transmission or distort estimates of the burden caused by different respiratory viruses.

The clinical data offered a less powerful route to identifying the causative pathogen. Arthralgia, or joint pain, was associated with SARS-CoV-2 positivity, with an odds ratio of 3.22, while cough was also associated with infection, with an odds ratio of 1.77. Dyspnea, or shortness of breath, was associated with negative SARS-CoV-2 results in this dataset. But symptom-based prediction models achieved only moderate discrimination, with areas under the receiver operating characteristic curve of approximately 0.72 to 0.74. An AUC of 0.5 would indicate performance no better than chance, while 1.0 would represent perfect separation. The results therefore suggest that symptoms may help guide suspicion, but cannot reliably distinguish SARS-CoV-2 from the many other viruses that produce overlapping respiratory syndromes.

That limitation is especially consequential in settings where laboratory resources are unevenly distributed. A syndromic surveillance strategy based on multiplex testing can reveal which pathogens are circulating, whether several viruses are rising simultaneously and how often patients carry co-infections. It can also reduce the tendency to interpret every fever, cough or sore throat through the lens of the most prominent epidemic. During the pandemic, the study notes, 19 percent of symptomatic patients in Antananarivo who tested negative for SARS-CoV-2 by RT-PCR were positive for other respiratory viruses. A negative COVID-19 result, in other words, did not mean that the illness lacked a detectable viral cause.

The researchers describe BioFire FA-RP2.1 as a useful tool for acute respiratory infection surveillance, while emphasizing that it should not be treated as infallible. Its lower SARS-CoV-2 detection count compared with simplex RT-PCR, together with the 61 discrepant results, shows why assay performance must be interpreted in context. Multiplex testing offers breadth, speed and automation, but targeted RT-PCR may detect some low-level infections that fall below the multiplex platform’s threshold. Conversely, a multiplex panel can identify pathogens that a SARS-CoV-2-only assay is never designed to find. The most informative strategy may therefore be sequential: first apply targeted testing, then use multiplex analysis to investigate negative or clinically ambiguous samples, or deploy the panel directly when broad surveillance is the priority.

The study also illustrates why molecular surveillance is not simply a matter of counting positive tests. Every assay is a measurement system with a detection limit, target design and susceptibility to the conditions of the specimen. Results can shift with the stage of infection, the quality of nasopharyngeal collection and the presence of other microbes. Sequencing can resolve some uncertainties, although it requires additional technical capacity and is not necessarily practical for every routine sample. Together, the methods create a layered diagnostic picture: targeted RT-PCR provides sensitive detection for a selected pathogen, multiplex testing broadens the search, and sequencing helps adjudicate difficult cases.

The findings have implications beyond Madagascar. Respiratory viruses continue to circulate in overlapping waves, and the symptoms caused by influenza, rhinoviruses, seasonal coronaviruses, respiratory syncytial virus and SARS-CoV-2 often converge. In hospitals, knowing the likely cause can influence isolation procedures and treatment decisions; in public-health laboratories, it can change estimates of transmission and disease burden. The Antananarivo results show that syndromic surveillance is strongest when it recognizes the limits of both clinical observation and individual assays. A cough may signal infection, but not which one. A negative result may reflect true absence—or viral quantities hovering just below a platform’s detection threshold.

Conducted as part of routine clinical monitoring, the investigation received ethical approval in Madagascar and included consent for the use of samples and associated data. The work was funded by the Fondation Mérieux in Lyon, France. By mapping respiratory infections with a combination of targeted amplification, multiplex detection and sequencing, the researchers provide a detailed snapshot of diagnostic complexity during the pandemic period. Their central message is clear: identifying the pathogen behind an upper respiratory infection requires more than matching symptoms to a familiar disease. In a world of co-circulating viruses and mixed infections, broad molecular testing can turn an apparently simple respiratory case into a more accurate—and more useful—epidemiological signal.