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

Area of Science:

• Environmental Epidemiology
• Biostatistics
• Demography

Background:

• Traditional demographic models struggle with high variability in human mortality data.
• Integrating social, economic, and environmental factors is crucial for improving mortality predictions.
• Existing models lack adequate handling of extreme meteorological and air pollution events.

Purpose of the Study:

• To evaluate the association between human mortality (by age and sex) and meteorological/air pollutant covariates.
• To improve demographic modeling by incorporating environmental data at various geographical levels.
• To enhance the dynamic behavior of basic demographic models, like the Farrington-like model.

Main Methods:

• Utilized a generalized linear modeling framework.
• Applied negative binomial distribution to address overdispersion in mortality data.
• Analyzed data at country, macro-climate region, and county levels.

Main Results:

• The proposed approach improved the dynamic behavior of the Farrington-like model.
• Successfully incorporated extreme meteorological and natural air pollution events into the model.
• Demonstrated effective performance even with scarce mortality data.

Conclusions:

• Integrating meteorological and air pollutant data significantly enhances human mortality modeling.
• The negative binomial generalized linear model provides a robust framework for analyzing variable mortality data.
• This enhanced modeling approach is particularly valuable for regions with limited demographic data.