By Office for National Statistics, Natural Capital Team
The coronavirus (COVID-19) pandemic has led to a big drop in pollution levels – global carbon emissions per day were up to 17% lower than normal at one stage in early April, returning to levels last seen in 2006.
However, some studies have suggested that long-term exposure to air pollution before the pandemic is associated with severe symptoms from COVID-19 and a greater risk of death.
Studies in the United States (US), Northern Italy and the Netherlands all found that a small increase in pollution exposure raises the number of COVID-19 deaths, with the most recent study in the Netherlands pointing to a rise in the death rate of up to 21%.
Our analysis examines the impact of long-term exposure to air pollution on COVID-19 deaths in England, with results pointing to a smaller effect than previous studies.
What links pollution exposure and COVID-19 deaths?
Consistent exposure to air pollution is a known cause of breathing difficulties and other long-term conditions in the lungs and the heart.
Our data show that more than one-third (35%) of deaths involving COVID-19 in England up to the end of June had a respiratory or cardiovascular disease as the main pre-existing health condition.
We have published detailed analysis showing that deaths involving COVID-19 were more common in highly polluted areas, particularly early in the pandemic.
However, the correlation between pollution and mortality fell as deaths rose and lockdown was introduced, before levelling off in early May.
The link between early COVID-19 deaths and exposure to dirty air was partly down to the outbreak in London (where pollution levels are generally higher than the rest of the country).
Up to the week when lockdown began (week ending 27 March 2020), 45% of COVID-19 deaths in England had occurred in the capital. By the week ending 12 June (cut-off date for this analysis), this had fallen to 18%.
As the virus spread across the country and deaths became more evenly distributed, the correlation between air pollution exposure and COVID-19 mortality decreased.
Our analysis includes 46,471 deaths involving COVID-19 among usual residents of England between 7 March 2020 and 12 June 2020, registered by 22 June 2020.
We considered exposure to ozone, fine particulate matter (known as PM2.5), and nitrogen dioxide (NO2) and NOx. We did not account for changes in pollution levels since the pandemic began.
Polluted areas initially had higher rates of COVID-19 deaths, but this trend decreased as the death toll rose
- We focus on NO2 and PM2.5 as the strongest positive correlates with COVID-19 mortality. Ozone is excluded because it is negatively correlated with COVID-19 deaths, as there is no reason to believe that exposure to ozone would reduce the risk of dying from COVID-19. Instead, ozone is likely to be acting as a proxy for living in very rural areas (where it is more concentrated compared with other pollutants). This air quality map highlights areas of high Ozone exposure, compared with other pollutants.
- The purpose of this chart is to show how the relationship in the data between COVID-19 mortality and air pollution changed as the infection spread. This chart is not designed to provide any deeper interpretation of the correlation. However, to explain: the values here are the log natural of the odds ratio of death from COVID-19 for the whole population based on models in which nothing else is controlled for. It is important to remember that an odds ratio does not give an overall risk of death but rather the change in the risk of death from the population baseline; there is no adjustment for infection rate. Finally, we say “exposure” to make it easier to understand what we are measuring is the average concentration of pollutants around the home, which is a good proxy value but not quite the same thing.
The effects of long-term exposure to air pollution on COVID-19 deaths appear smaller than reported in previous studies
While death rates have generally been higher in polluted areas, this on its own does not establish pollution exposure as a cause of COVID-19 deaths.
To isolate the impact of air pollution, we developed a statistical model controlling for factors such as levels of deprivation, population density, public health (such as smoking rates) and pre-existing health conditions. This enables us to compare deaths among populations of similar health from the same sort of area, with the main difference being long-term exposure to air pollution.
We also controlled for ethnicity, thereby comparing deaths among populations with similar levels of ethnic diversity.
Our previous analysis suggests that a range of structural differences are resulting in higher COVID-19 death rates among people of Black, Asian and Minority Ethnicity (BAME). Air pollution exposure could be a contributing factor if it causes COVID-19 deaths, with the BAME population more likely than those of White ethnicity to live in inner city areas (that are more polluted).
Ethnicity is strongly correlated with pollution exposure, with ethnic minorities more likely to live in polluted areas
Ultimately, our analysis is inconclusive.
Without controlling for ethnicity, we find that long-term exposure to fine particulate matter (PM2.5) could increase the risk of contracting and dying from COVID-19 by up to 7% (comparable with results from a US study).
However, when controlling for ethnicity, air pollution exposure has no statistically significant impact on COVID-19 deaths.
As air pollution is just one of many factors that could be driving disproportionate outcomes for minority ethnic groups, the increased risk of dying from COVID-19 (found when ethnicity is not controlled for) is likely to be an overestimate of the true effect.
A recent study of 400 COVID-19 patients admitted to a Birmingham hospital said that BAME patients were “more likely to be admitted from regions of highest air pollution, housing quality and household overcrowding deprivation”.
Our approach in this study was not sufficient to pull all these factors apart. Achieving this is likely to require highly detailed individual-based modelling.
In this study, we have only been able to examine air quality and COVID-19 deaths at the population level, based on the availability of existing data. At this level, we have to make assumptions such as consistent exposure to pollution across a 1km grid square. In reality, exposure could vary from one end of a street to another, and individual exposure is affected by where someone works as well as where they live.
To disentangle factors such as pollution and ethnicity, and their contribution to COVID-19 mortality, we would need to identify detailed characteristics of the individuals who have died from COVID-19. While we hope to publish this kind of analysis in future, the data and modelling required take much longer to put together.
About the authors: This article was originally posted on the ONS website. It is reproduced here under a Open Government Licence v3.0. The main contact for the article is Adam Dutton (firstname.lastname@example.org). You can read the original article here.
Suggested further reading
Rowland N., Vlachos, S, The effect of exposure to air pollution on health and mortality Authors: Dr Neil Rowland, Dr Stavros Vlachos, Dr Mark McGovern, Professor Duncan McVicar, Professor Dermot O’Reilly Email: N.Rowland@qub.ac.uk Date: April 2020
Pearce, J., & Kingham, S. (2008). Environmental inequalities in New Zealand: A national study of air pollution and environmental justice. Geoforum, 39(2), 980–993. https://doi.org/10.1016/j.geoforum.2007.10.007
“How can we take the learnings of COVID-19 lockdown and turn them into a brighter environmental future for millions across the Global South?” https://www.birmingham.ac.uk/research/quest/sustainable-environments/COVID-19-and-air-pollution.aspx