Influence of atmospheric parameters on human mortality data at different geographical levels

Human mortality data are often modeled using a demographic approach as a function of time. This approach does not present an adequate fit model for the number of deaths with great variability. For this reason, additional information (social, economic and environmental) is required for complementing...

Descripción completa

Detalles Bibliográficos
Autores: Sánchez Balseca, Joseph|||0000-0002-1741-3229, Pérez Foguet, Agustí|||0000-0002-2737-4710
Tipo de recurso: artículo
Fecha de publicación:2021
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/346437
Acceso en línea:https://hdl.handle.net/2117/346437
https://dx.doi.org/10.1016/j.scitotenv.2020.144186
Access Level:acceso abierto
Palabra clave:Air--Pollution--Health aspects
Air quality
Environmental statistics
ENSO
Volcanic
Negative binomial
Human health
Aire -- Contaminació -- Aspectes sanitaris
Àrees temàtiques de la UPC::Desenvolupament humà i sostenible::Degradació ambiental::Contaminació atmosfèrica
Àrees temàtiques de la UPC::Desenvolupament humà i sostenible::Desenvolupament humà::Salut
Descripción
Sumario:Human mortality data are often modeled using a demographic approach as a function of time. This approach does not present an adequate fit model for the number of deaths with great variability. For this reason, additional information (social, economic and environmental) is required for complementing and improving demographic modelling. This article evaluated the association between human mortality data (segregated by age and sex) with meteorological and air pollutant covariates at three geographical levels: country, macro-climate regions and county. The modelling was based on a generalized linear modelling framework and takes into account the common characteristic of overdispersion in human mortality data through the application of negative binomial distribution. The proposed approach improved the dynamic behavior of the Farrington-like model (basic demographic model) and took into account the extreme meteorological and natural air pollution events. Notably, the proposed modelling worked well in cases where the amount of data was scarce.