Integrating demographic and epidemiological transitions into climate impact projections: a case study with temperature-mortality association in Antananarivo

Details

Supervisors

Masquelier, Bruno

Faculty

Faculté des sciences économiques, sociales, politiques et de communication

Degree label

Master [120] en sciences de la population et du développement, à finalité spécialisée: démographie

Abstract

Background: The well-established effect of global warming on human health underscores the urgency of addressing this relation on the most vulnerable populations. However, the scarcity of studies in highly vulnerable regions, such as Low and Lower Middle-Income Countries (LLMICs), raises methodological questions about measuring this association using methods that have mostly been applied in different epidemiological and demographic contexts. Methods: Using death notification data from Antananarivo from 2000 to 2015 and daily mean temperatures, we conducted multiple sensitivity tests for our chosen Distributed Lag Non-Linear Model (DLNMs). These tests aimed to assess variations in the temperature-mortality association by age, cause of death and temperature data choice. Additionally, we controlled for changing mortality rates and confusion from the seasonal patterns in the model. Results: We observed marked disparities in the temperature-mortality association for different age groups, although with a high degree of uncertainty. Projected changes in mortality rates, resulting from epidemiological and demographic transitions, appeared to contribute more to variations in the numbers attributed to non-optimal temperatures than climate change itself. The measured association varied markedly for different causes of deaths, depending on the inclusion of seasonality as a confounder and the choice of the exposure variable. Conclusions: Contextualised studies measuring the effects of temperature on health are necessary to understand the burden of climate change. Our findings underscore the importance of considering population structure and health profile when using DLNMs in LLMICs. While our study has limitations, we express its relevance, especially for providing a basis for future research to assess potential adaptation.