Historically, influenza outbreaks have been tracked by clinical findings and serology, and characterized by culture. In a study published in the latest issue of the British Medical Journal, Carman and colleagues compared the relative speed and sensitivity of three flu surveillance assays: serology, culture and polymerase chain reaction (PCR). Samples were collected from the nose and throat of 168 patients within a mean 5.3 days after the onset of flu symptoms; 112 patients were subsequently confirmed to have flu infection. The fastest laboratory results were obtained by PCR (36 hours), followed by culture (minimum, 1 week) and serology (minimum, 3 weeks). In 60% of patients with positive serology and negative PCR results, samples had been collected 8 or more days after the onset of symptoms. PCR did not yield any false-positive results. For samples collected within 7 days of the onset of symptoms, the sensitivity of PCR for any positive diagnosis (by PCR, serology or culture) was 94.2%; the sensitivity of serology for any positive diagnosis was 88.7%, whereas culture was comparatively insensitive. These findings show that PCR has advantages over serology for rapid, real-time flu surveillance, although the timing of sample collection for PCR analysis is important.