Applying a Reconstruction Method to Cause-of-Death Series in Different Countries

Svitlana Poniakina, Insitut National D'Études Démographiques
France Meslé, Institut National d'Études Démographiques (INED)
Denys Dukhovnov, University of California, Berkeley
Magali Barbieri, Institut National d'Études Démographiques (INED)
Yui Ohtsu, Saitama University
Jacques Vallin, Institut National d'Études Démographiques (INED)

Cause of death data are essential for demographic and public health research, but classification changes over time pose analytical challenges. The Human Cause-of-Death Data series (HCD) addresses these inconsistencies by reconstructing historical time series of death by causes. The reconstruction method, initially developed by Jacques Vallin and France Meslé at INED, adjusts for ICD changes and has been applied to various national contexts. The method involves setting up correspondence tables between ICD revisions, creating fundamental associations to test consistency, and calculating transition coefficients. Mapping Alzheimer’s disease across France, Spain, and Japan demonstrates significant differences in how causes of death are categorized. In France, the association between Alzheimer’s in ICD-9 and ICD-10 is straightforward, while in Spain and Japan, the associations involve additional conditions, such as dementia, pneumonitis, and senility. These differences underline the necessity of tailoring the reconstruction method to each country’s unique mortality data. This study concludes that applying the reconstruction method enhances the accuracy of cause-specific mortality trends. However, the reconstruction between ICD revisions must be done separately for each country due to variations in medical diagnostics, comorbidities, and classification practices.

Keywords: Data and Methods, Mortality and Longevity

See extended abstract.