This article is more than 5 years old. Information may no longer be current.
Rapid influenza tests most accurate in high prevalence times
In lighter influenza seasons, interpret test results with caution.
If the prevalence of influenza is less than 10% in your community, rapid tests should be used cautiously because positive results will be correct less than 70% of the time, according to a study reported in Pediatrics.
“These tests are good but not perfect. You need to know the prevalence of the disease in order to get the appropriate interpretation,” Carlos G. Grijalva, MD, research assistant professor of preventive medicine at Vanderbilt University, said in an interview.
Influenza virus infections result in excess hospitalizations and emergency department visits every year in the United States. Because the clinical characteristics of influenza often overlap those of other common respiratory diseases, diagnosis often requires confirmatory testing.
Rapid antigen detection tests can provide timely identification of infection. Grijalva and colleagues assessed the times at which rapid tests are most predictive of influenza infection.
They extracted data from the New Vaccine Surveillance Network, which enrolled children younger than 5 years who were hospitalized with respiratory symptoms or fever from October 2000 through September 2004.
Nasal and throat swabs were obtained. Influenza virus was detected by culture and reverse-transcription polymerase chain reaction, and results were compared with the rapid tests to determine sensitivity and specificity. To determine influenza prevalence, researchers extracted data from the same database.
Grijalva said there were a number of rapid antigen detection tests on the market, but they obtained an overall estimate and did not compare the merits of one test over another.
More specific than sensitive
During the four years of surveillance, 2,797 children were hospitalized with respiratory symptoms or fever and were subsequently tested for influenza by viral culture and rapid testing. Influenza infections were confirmed in 160 children.
Overall, 270 children had a rapid test for influenza. Of the 41 children who had influenza detected by culture and reverse transcription polymerase chain reaction, 26 were influenza positive by rapid testing, for a sensitivity of 63%. Among the 229 children who tested negative by the criterion standard, 223 had a negative test result, for a specificity of 97%.
Whether a positive rapid test result indicates influenza infection or not (positive predictive value) depends on the prevalence of influenza, according to the report. During the 2002 to 2003 season, when the prevalence of influenza was about 5%, the predictive value of a positive rapid test was approximately 50%. In other words, positive rapid test results indicated real influenza infections only 50% of the time. By contrast, a negative rapid test was accurate 98% of the time.
As the flu became more prevalent, the value of a positive test increased. When the flu peaked at 21%, the predictive value of a positive test was 85%. Conversely, about 9% of true influenza cases had a negative rapid-test result.
The predictive value of a positive test dropped as low as 25% when the influenza prevalence dropped to 1.6%.
Overall, throughout the entire 2002 to 2003 flu season, the predictive value of a positive rapid test was greater than 70% for only four weeks. – by Jeremy Moore
For more information:
- Grijalva CG, Poehling KA, Edwards KM, et al. Accuracy and interpretation of rapid influenza tests in children. Pediatrics. 2007;119:6-11.