A Literature Review of the Effects of Air Pollution on COVID-19 Health Outcomes Worldwide: Statistical Challenges and Data Visualization

Abstract

Several peer-reviewed papers and reviews have examined the relationship between exposure to air pollution and COVID-19 spread and severity. However, many of the existing reviews on this topic do not extensively present the statistical challenges associated with this field, do not provide comprehensive guidelines for future researchers, and review only the results of a relatively small number of papers. We reviewed 139 papers, 127 of which reported a statistically significant positive association between air pollution and adverse COVID-19 health outcomes. Here, we summarize the evidence, describe the statistical challenges, and make recommendations for future research. To summarize the 139 papers with data from geographical locations around the world, we also present anopen-source data visualization tool that summarizes these studies and allows the research community to contribute evidence as new research papers are published.

Keywords: COVID-19; air pollution; data science; evidence synthesis; visualization.

Figure 1

PRISMA figure displaying the screening process for the 4,445 studies identified in the database searches run on May 21, 2021. In the first round, duplicate studies were removed. In the second round, studies deemed irrelevant by title and abstract review were removed. Finally, studies deemed irrelevant by full-text review were removed, resulting in a total of 139 studies included in the review.

Figure 2

(Figure appears on preceding page) Screenshots of the evidence synthesis tool dashboard (http://bit.ly/3hzcsbv). (a) Zoomed out dashboard containing information from most countries, including the three most studied countries (Italy, China, United States). (b) Zoomed into Central Europe. The air pollutant types, papers within each country, and statistical methods chart have changed accordingly. In addition, the Austria graduated circle and a specific paper from Austria have been selected and the associated text is displayed. The short video demonstrating the capabilities of the dashboard is available at https://vimeo.com/709361095.

Figure 3

Evidence synthesis results stratified by the type of air pollutant studied. There were a total of 139 studies (76 short-term, 63 long-term), and many studies studied multiple pollutants. The total number of studies (blue-green and pink), number of studies that reported a statistically significant positive association between the air pollutant and COVID-19 outcomes (blue-green), and percent of papers that report significant results (white numbers) are presented for each pollutant. (a) Long-term studies (AQI, n = 3; CO, n = 10; NO₂, n = 32; O₃, n = 21; PM₁₀, n = 27; PM_2.5, n = 52; SO₂, n = 14). (b) Short-term studies (AQI, n = 15; CO, n = 23; NO₂, n = 35; O₃, n = 32; PM₁₀, n = 40; PM_2.5, n = 55; SO₂, n = 17).

Figure 4

Screenshots of the fields in Survey123 (https://bit.ly/3hGdIti).