Accurately assigning causes or contributing causes to deaths remains a universal challenge, especially in the elderly with underlying disease. Cause of death statistics commonly record the underlying cause of death, and influenza deaths in winter are often attributed to underlying circulatory disorders. Estimating the number of deaths attributable to influenza is, therefore, usually performed using statistical models. These regression models (usually linear or poisson regression are applied) are flexible and can be built to incorporate trends in addition to influenza virus activity such as surveillance data on other viruses, bacteria, pure seasonal trends and temperature trends.
Mortality exhibits clear seasonality mainly caused by an increase in deaths in the elderly in winter. As there may be substantial hidden mortality for a number of common pathogens, we estimated the number of elderly deaths attributable to common seasonal viruses and bacteria for which robust weekly laboratory surveillance data were available.