Usefulness of death certificates for influenza surveillance

Background: Surveillance for influenza fatalities has long been part of the national system to gage the severity of an influenza “season.” Traditional surveillance, relying on individual case reports (ICR) from hospitals, physicians, and coroners can be slow and resource intensive. Surveillance using data from death certificates may provide more timely data and be less resource intensive which is important in a pandemic. Death certificates in Los Angeles County must be reported via the electronic death record system (EDRS). Reports are analyzed weekly for any case with “influenza” listed as a cause of death. Methods: We compared the cases of influenza fatalities reported by EDRS versus the fatalities reported by ICR during August 30, 2009-April 30, 2010. We compared the demographic variables and calculated the sensitivity and the positive predictive value of EDRS reported fatalities. Results: there were 102 fatalities reported by ICR versus 85 by EDRS. The % female (53, 54), % Hispanic (51, 56) and median age (47,49), respectively, were similar. However, only 48% of the EDRS records reported any secondary condition or risk factor for acquiring influenza versus 90% by ICR. Only 60 of the 85 EDRS reports were determined to be due to laboratory confirmed influenza. The sensitivity of EDRS was 41% versus traditional reporting and the positive predictive was 71%. EDRS identified 18 additional cases that had not been reported by ICR. Conclusions: EDRS provided a timely, automated method of tracking influenza deaths. However, it lacked sensitivity and cannot fully replace traditional surveillance methods.