Mortality Gaps & Cause Composition: A Methodological View

How have leading causes of death shifted by age and over calendar time, and what do the gaps between major causes reveal about population health? This talk unifies two distributional perspectives within an age–period setting: 

• First, Associate Professor Andrea Nigri models differences in death counts directly and shows that the Skellam distribution parsimoniously captures cause-of-death gaps (e.g., cancer vs circulatory diseases). Case studies from 1960 - 2015 illustrate when gaps widen or narrow across adulthood and old age.
• Second, he presents an Age–Period Dirichlet–Multinomial model that treats causes as a composition of total deaths, ensuring coherence by construction, cause-specific predictions always sum to the total. This avoids the incoherence of modelling causes independently, while accommodating over-dispersion for realistic uncertainty. 

These tools provide a transparent, robust lens on who dies of what and when, supporting population-level comparisons across ages and periods and giving practitioners demographically grounded insights for monitoring change and designing policy.

Andrea Nigri is an Associate Professor of Statistics at the University of Foggia. His work develops and refines statistical methodologies for statistical demography, bridging longevity analysis with statistical and machine-learning approaches. He has undertaken research visits at the Max Planck Institute for Demographic Research, the Interdisciplinary Centre on Population Dynamics (University of Southern Denmark), the Department of Clinical Epidemiology (Aarhus University), and the Department of Actuarial Studies and Business Analytics (Macquarie University). Andrea serves as Co-Editor-in-Chief of Communications in Statistics—Theory and Methods, and as Associate Editor for Population Health Metrics and the Journal of Population Research. 
His current research follows two strands: (i) methodological development in longevity-risk modelling; and (ii) the analysis, monitoring, and understanding of demographic phenomena.

Join Zoom Meeting

https://anu.zoom.us/j/83873550469?pwd=u5NmcuzrPEp4euXUkpxdW6n3VaywlW.1

Meeting ID: 838 7355 0469

Password: 232267