Changes in the Modal Age at Death of Ageing- and Behaviour-Related Diseases in the US: A Multiple-Cause-of-Death Approach

Jeroen Spijker , Centre d'Estudis Demogràfics (CED)
Paola Vazquez-Castillo, Interdisciplinary Center on Population Dynamics (CPop), University of Southern Denmark

Theoretical focus: As mortality patterns are shifting towards older ages, researchers have increasingly focused on studying the modal age at death (MAD) –the age that concentrates most deaths in a given year– because, unlike life expectancy, it’s unaffected by mortality changes at younger ages. Concurrently, at older ages, the mortality risk is becoming more heterogeneous due to comorbidities. Hence, analysing the age-at-death distribution within a multiple cause-of-death (MCoD) framework offers an excellent opportunity to enhance our understanding of longevity associated with multiple medical conditions. Data and method: MCoD and Underlying Cause of Death (UCoD) data for 1999-2019 were obtained for the US from https://wonder.cdc.gov and the population data from HMD. Multiple decrement life tables were first estimated for the UCoD to standardise the age distribution over time, so the MAD could be estimated using Kannisto’s quadratic approximation. Expected findings: Initial results show a notable variation in the MAD across different causes for both sexes. For female cancer the mode was 83.2 years versus 89.7 years for total mortality. For men, the MAD was respectively 80.7 and 87.5 years. Not all causes exhibited similar patterns: the MAD for female heart disease was 91.2 years, higher than the overall mortality mode. The changes in the MAD over time had peaks and troughs generally observed on the same moments with a generally upward trend. We still need to estimate MAD using MCoD data and compare it to the results for UCoD.

See extended abstract

 Presented in Session 65. Flash session Mortality and Longevity