The diagnostic accuracy of flu tests is often overestimated by clinicians. So how sensitive are they for “flu-like illness”?
Rapid Influenza Testing
The diagnostic accuracy of influenza swabs is often overestimated by clinicians. The CDC notes that rapid influenza testing has a sensitivity ranging from approximately 50% to 70% — meaning that in up to half of influenza cases, the flu swab results will still be negative. Another study in Turkey showed that the sensitivity of rapid influenza testing for H1N1 ranged from 31.7% to 50%, again depending on the brand of test. A 2012 metanalysis of the accuracy of rapid influenza testing showed an average sensitivity for detecting influenza in adults of only 54%. Sensitivity in children is somewhat higher since they tend to shed more virus. In most studies, the specificity of influenza testing is 90-95%, meaning that there is a much lower incidence of false positive results.
Both the CDC and the FDA note that once influenza is established in a community, false negative rates of rapid influenza testing increase and therefore a negative flu swab should not be used to exclude influenza as a diagnosis. “Particularly at the beginning of the season or outbreak, negative results may not be relied on to guide management or treatment decisions.” The FDA also notes that inadequate specimen collection will yield a low quality specimen and will “likely yield false negative tests.”
Rapid influenza testing may be a good way to establish a diagnosis of influenza, but it is not a reliable way to rule *out* influenza when influenza is present in a community. And don’t forget to get a good specimen sample.
Clinical Diagnosis of Influenza
Influenza has a typical incubation period of 1-4 days. In other words, patients may not have flu symptoms for several days after they have been infected. When patients develop symptoms, the symptoms usually last 3-7 days, but the cough can last for up to 2 weeks. Can we reliably diagnose influenza with clinical exam?
The CDC notes that – when influenza has been confirmed in a community – with healthy patients between 13 and 65 years old “a simple clinical definition of influenza (acute onset of cough and fever) for laboratory-confirmed influenza infection has a positive predictive value of 79% to 88%.” The combination of headache, fever, and cough had a 75% positive predictive value for influenza in a British study by van Elden et al. A study by Stein et al. showed that neither rapid influenza testing nor a clinical decision rule was superior to clinician judgment in diagnosing influenza –
and that diagnostic accuracy increased significantly when symptoms were present for less than 48 hours.
Acute onset of fever and cough in children 5-12 years old had a positive predictive value of 71% to 83% while those symptoms had only 64% positive predictive value in children less than 5 years old. Younger children are more likely to have vomiting and diarrhea as secondary influenza symptoms.
Patients more than 65 years old are also more likely to have atypical influenza symptoms. Acute fever and cough had a positive predictive value of 53% for influenza in elderly patients. Headaches, malaise and myalgias were common findings in elderly patients with influenza.
When in the midst of influenza season, a clinical diagnosis of “influenza like illness” in healthy patients 13-65 years old is probably as accurate as the results of influenza testing and other clinical workups.