Coronavirus and Pollution – the Limitations of Observational Studies

Written by Argumentful

We’ve seen a recent research that has made the headlines of mainstream media like The Guardian, BBC News, The Independent or CNN: “Can atmospheric pollution be considered a co-factor in extremely high level of SARS-CoV-2 lethality in Northern Italy?“.

We’re convinced that such a paper would never have been picked up outside such “unprecedented” times where any additional piece of information that makes you feel more in control fills a huge void.

The reporting was satisfactory, with most authors pointing out that the study shows only a correlation between levels of pollution and increased mortality from Covid-19. However, we’re not fully satisfied with the tone of the articles which, similar to general approaches to media reporting, emphasised the fact that it’s a study (rather than some opinion), the fact that the authors are professors at important universities or that it was published in a reputable journal. We see that as important only to eliminate the suspicion of fake news, which would fall upon the study if it was published on some dodgy website. But it does appeal to authority which is not 100% scientific.

What we would emphasise for all studies, though, are two things:

  1. Methodology – in this case it was an observational study based on a relatively limited set of data
  2. Funding conflicts – in this case the authors declare no conflicts of interests of untoward funding, which is good.

As we’ve discussed in the previous post, the methodology of a study is very relevant because it determines what kind of conclusions it draws. All analytical studies aim to show relationships between two or more factors but they are fundamentally different:

  • Observational studies – quantify the relationship between two factors without intervention from investigator
  • Experimental studies – experimenters intervene to show causal links between two factors

More info here.

It is therefore important to put more trust in results derived from experimental research compared to observational or exploratory studies. This is how Argumentful is designed and we believe it provides a fair evaluation of arguments based on scientific research.

Observational studies don’t have the purpose to prove causal links. Things that show a correlation aren’t necessarily in a causal relationship. In our case, it does look like the pollution and the moratlity rates are correlated and the authors also show a potential mechanism of how air pollution produces inflamation in the lungs. While this is a valid hypothesis, it’s definitely not a proof and we know that many other things are correlated to pollution. Large cities and economically developed areas have, unfortunately, bad environmental records (take London as an example). So the agglomeration
of people in urban places, on mass transport, at football matches, could also be factor.

Of course, we haven’t posted this to “debunk” the study or belittle its impact but rather to place it in the right context. It’s very useful for people to validate good hypotheses (such as the one from this study) especially some that are grounded in existing research. The authors are actually very honest about their paper and mention in the end that “experimental and epidemiological studies are urgently needed to evaluate the role of the atmospheric pollution”.

You May Also Like…



  1. Coronavirus and Men – Yet Another Observational Study | Argumentful - […] as we were publishing the last post about the potential effects of pollution on mortality rates, the tabloids were…
  2. About the Higher Death Rates from COVID-19 amongst BAME People | Argumentful - […] we’ve mentioned in the previous posts, such research, which can be classified as a cross-sectional observational study, can’t explain…

Leave a Reply