Spatio-temporal variability of temperature effects on mortality in Sao Paulo, Brazil
File(s)
Author(s)
Roca Barcelo, Aina
Type
Thesis
Abstract
Non-optimal temperatures pose significant health risks, yet understanding the nuanced relationship between temperature and mortality within cities, including vulnerability among specific populations and neighbourhoods. The lack of temperature data at the necessary spatial and temporal resolution, albeit not the only cause, an important one. Addressing this research gap is crucial for identifying shortcomings in current strategies and prioritizing interventions for high-risk groups/areas. My thesis focuses on São Paulo to characterize temperature-related all-cause mortality risk over time and across population and area-level characteristics.
To analyse adaptation patterns, I examined temperature-mortality trends in São Paulo State from 2000 to 2018. Using Time-Varying Distributed Lag Non-linear Models (tv-DLNM), I captured the non-linear and delayed effects of temperature on mortality, adjusting for various factors and stratifying the analysis by gender, age group, and ethnicity. Significant trends in Minimum Mortality Temperature (MMT) and Cumulative Relative Risk (cRR) were found, with variations across population groups indicating disparities in adaptation capacities.
Spatial variability in temperature across São Paulo was assessed using a Random Forest (RF) model, predicting daily mean temperature at 500m2 resolution. The model integrated data from ground measurement stations and satellite-derived products, exhibiting good overall performance and outperforming linear models. With these temperature estimates and individual mortality records, I explored differences in temperature-mortality associations among population groups based on individual and area-level factors. Findings revealed elevated mortality risks among older individuals, women, and those with limited education for both heat and cold, and among married or widowed individuals and those of mixed ethnic backgrounds for cold temperatures. Results for contextual factors were inconsistent, warranting further research.
This thesis delves into São Paulo's temperature-mortality link, stressing the need for continuous public health monitoring and updates. The development of a spatiotemporally resolved temperature dataset, first-of-its-kind in Latin America, minimizes exposure bias and facilitates small-area studies.
To analyse adaptation patterns, I examined temperature-mortality trends in São Paulo State from 2000 to 2018. Using Time-Varying Distributed Lag Non-linear Models (tv-DLNM), I captured the non-linear and delayed effects of temperature on mortality, adjusting for various factors and stratifying the analysis by gender, age group, and ethnicity. Significant trends in Minimum Mortality Temperature (MMT) and Cumulative Relative Risk (cRR) were found, with variations across population groups indicating disparities in adaptation capacities.
Spatial variability in temperature across São Paulo was assessed using a Random Forest (RF) model, predicting daily mean temperature at 500m2 resolution. The model integrated data from ground measurement stations and satellite-derived products, exhibiting good overall performance and outperforming linear models. With these temperature estimates and individual mortality records, I explored differences in temperature-mortality associations among population groups based on individual and area-level factors. Findings revealed elevated mortality risks among older individuals, women, and those with limited education for both heat and cold, and among married or widowed individuals and those of mixed ethnic backgrounds for cold temperatures. Results for contextual factors were inconsistent, warranting further research.
This thesis delves into São Paulo's temperature-mortality link, stressing the need for continuous public health monitoring and updates. The development of a spatiotemporally resolved temperature dataset, first-of-its-kind in Latin America, minimizes exposure bias and facilitates small-area studies.
Version
Open Access
Date Issued
2023-08-04
Date Awarded
2024-03-01
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Fecht, Daniela
Pirani, Monica
Vineis, Paolo
B. Piel, Frédéric
Sponsor
Imperial College London
Publisher Department
School of Public Health
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)