The spatial distribution of mortality in a large metropolis: A methodology for identifying geographical patterns and contributing factors

Abstract

Objective: The distribution of mortality in large urban areas, such as major metropolises, is normally highly heterogeneous. Disadvantaged populations often present distinct health patterns that might be neglected by studies that focus on privileged areas of the city. Understanding the varied causes of death across city regions can greatly aid public health management and inform evidence-based policies.

Study design: Retrospective observational study.

Methods: In this paper, we propose a methodology that combines statistical analysis, clustering, and geographical visualization to analyze mortality patterns in large metropolises. It determines clusters of districts based on mortality profiles and identifies deviations from expected death rates considering demographic factors like age profile and Human Development Index (HDI).

Results: We applied the method to the city of São Paulo, analyzing 376,452 deaths across five years. The method could identify four clusters within the city, each with different characteristics regarding the causes of death. Unsurprisingly, these clusters are highly correlated with the district's HDI and age profile. However, the method identifies deviations from expected death rates based on HDI and age, potentially indicating deficiencies in public health services in specific regions.

Conclusions: The proposed methodology provides policymakers with tools to identify regions where causes of death exceed expected values for the population demography. It also offers insights into mortality patterns and their association with socioeconomic and demographic characteristics, enabling the development of targeted healthcare strategies to improve overall public health outcomes.

Keywords: Clustering; Evidence-based public policy; Geostatistics; Mortality; Spatial distribution.