Visualizing Mortality Dynamics for Causes of Death in the United States

Roland Rau, University of Rostock
Christina Bohk, University of Rostock
Magdalena Muszynska, Warsaw School of Economics

Demography is blessed with a wealth of data of relatively high quality. Hence, the challenge for demographers is not to make inferences based on small samples but to make sense of the available data, capturing essential trends and not omitting major characteristics. Mortality surface maps have been used to summarize millions of events with several thousand data points in a single figure. Building on those surfaces, we argue that maps of the rates of mortality improvement provide an even better tool to understand the underlying mortality dynamics. Using multiple cause of death data from the NCHS and exposures from the Human Mortality Database, we show that the slow increase of US life expectancy can be attributed to a mixture of period and cohort effects. Mortality from circulatory diseases improved enormously due to strong period effects. This development is counteracted by cohort effects of malignant neoplasms, in particular lung cancer.