Effects of Drought on Mortality in Macro Urban Areas of Brazil between 2000 to 2019: Dataset
For the first time in Brazil, this study poses a comprehensive assessment of the short-term association between drought exposure and weekly non-external, circulatory, and respiratory mortality during the last two decades, including a stratified analysis according to sex and age ranges of the population. Five files are incorporated, which include three databases supporting the conclusions of this study and the R code used in the main analysis. The list of variables included in each database is indicated in a text document. Particularly, data sets consist of information on weekly (1) non-external-, (2) circulatory-, (3) respiratory-cause mortality for the thirteen major Brazilian macro-urban areas in Brazil across 2000-2019, including information segregated by sex and different age population groups. The mortality databases were obtained from official daily death certificates of the Brazilian Health System database “DATASUS” (https://datasus.saude.gov.br/transferencia-de-arquivos/) publicly available through the Data Science Platform developed by the Institute of Scientific and Technological Information in Health of the Oswaldo Cruz Foundation. Moreover, each file incorporates weekly time series corresponding to SPEI-1 (drought index calculated at short-term, which was obtained from the Climatology and Climate Services Laboratory (LCSC) website (https://global-drought-crops.csic.es/#map_name=all_spei_1#map_position=4)) and mean temperature (obtained from the ECMWF public datasets (https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-levels?tab=form)) for each Brazilian location. This database version is completely described in Salvador et al. (2022) Effects of Drought on Mortality in Macro Urban Areas of Brazil between 2000 to 2019, GeoHealth, and you should reference it directly.
Steps to reproduce
A two-stage time-series analysis was conducted accounting for multivariate complex associations and pooling methods (meta-analytical techniques) to estimate the short-term association, between drought events and weekly mortality in Brazil at the same exposure week, expressed as relative risk. In addition, impact measures were also estimated through the calculation of the attributable fraction to drought exposure for the 13 macro urban areas as a whole. The analyses were conducted with R software (version 4.0.1) using dlnm and mixmeta packages.