Analysis of weekly mortality as a tool for the early detection of alerts in public health

Abstract

Background. There is a growing interest in public health in monitoring indicators in real time that permit immediate action to be taken. The aim of this study is to evaluate the utility of the analysis of weekly mortality for the early detection of alerts in public health. Methods. We analysed the early reception of mortality (ERM) from five computerised civil registers in Navarre and gathered 70% of total mortality (TM) for the 2003-2007 period. We identified the weeks when mortality exceeded the threshold calculated on the basis of the five previous years. We analysed the correspondence of these excesses in mortality with flu levels of incidence and with high atmospheric temperatures. Results. The was moderate concordance between the alerts detected from the information proceeding from the ERM and TM sources (Kappa index 0.59; CI 95% 0.42-0.76). Both sources behaved analogously in the detection of excesses of mortality when these occurred in weeks of severe flu, with 13 and 12 alerts respectively. In relation to high temperatures, TM only generated alerts in the years when heat waves were declared, while ERM also generated them in other weeks which were hot but did not reach the heat wave threshold. Conclusion. The early availability of results, the acceptable concordance between sources and the high percentage of excesses of mortality attributable to identified public health problems, show the potential utility of monitoring weekly mortality in epidemiological surveillance.