Data science

Using a database of multiple environmental pollutants to establish critical limits on exposure to air pollution in maternal and child health in Brazil

— University of São Paulo (USP)

Project Presentation

  • What it is?

    The project investigated the individual and joint effects of exposure to multiple pollutants (particulate matter, carbon monoxide, ozone, nitrogen dioxide, and sulfur dioxide) and fires on adverse birth outcomes such as prematurity, low birthweight, and congenital malformations and on the principal causes of hospitalization in the under-five population).  The proposed methodology used the new database of the Environmental Information System (SISAM), which contains daily measurements of exposure to the target pollutants in Brazilian municipalities and the national databases Information System on Live Births (SINASC), Information System on Hospitalizations (SIH), and Mortality Information System (SIM) in Brazil’s 5,572 municipalities from 2012 to 2018. The analyses performed were highly robust due to the amount of data analyzed and allowed the greatest contribution to the area, which is the recommendation for the revision of the critical limits on exposure to pollutants, which is important information for drafting public policies in the environment
    and health.

  • How was the experiment?

    1. Odds ratios (via logistic regression model using an ecological study design) for the associations between exposure to ambient air pollution (PM2.5, CO, NO2, O3) and fires during pregnancy and the target outcomes: prematurity, low birthweight, congenital malformations, and hospitalizations in children under five years

    2. Comparison of the results with the air quality limits set by CONAMA and recommended by the WHO (interim targets) for calculation of the percentage of attribution of outcomes and number of preventable cases in case of revision of the limits.

  • Main Results

    1. Mapping of the rates of exposure to pollutants and fires, mean exposure to the principal pollutants by municipality, thereby revealing the heterogeneity of exposure.

    2. Distribution of maternal-child health outcomes, analyses for the period 2012-2018, due to change in the tabulation of gestational age in SINASC in 2011. A total of 17,393,005 live births in Brazil were included in the study.

    3. Impact of exposure to multiple pollutants and fires on maternal-child health, individual and joint effects of exposure in the states and municipalities. The study’s results revealed the important impacts of exposure to pollutants, leading to adverse birth outcomes such as premature births, congenital malformations, and hospitalizations. Low birthweight correlates directly with prematurity, and since the critical limits set by CONAMA are much higher than the tolerable exposure levels, the values pursued by the WHO (interim targets) in the air quality guidelines revised in 2021. In the case of fires, there is no proposal suggested by the WHO or CONAMA. In addition to present the results as odds ratios and confidence intervals to facilitate the understanding and alert administrators and public policymakers on the problem’s seriousness, the project decided to present the results as the percentage of cases attributable to the pollutant’s exposure and the number of cases for each outcome that could be prevented with one unit reduction in the pollutants’ emission.

    4. Print and digital material, with the synthesis of the principal findings in accessible language for awareness-raising. A report containing all the findings aggregated by state and national territory for public policymakers; a folder containing  principal findings with accessible language for awareness-raising.

     

    5. Digital material for a workshop with the main stakeholders.

  • Brazilian Health implications

    • Establishment of critical limits for exposure to air pollution on maternal-child health in Brazil.
    • Recommendation for revision of critical limits on exposure currently adopted in Brazil.
    • Awareness-raising on the impact of exposure to pollutants and fires on maternal-child health

  • Global Health implications

    • The analyses’ robustness (data volume, time horizon, and territorial scope) allows their use by other countries.