Tuesday, May 12, 2020

The inequality of COVID-19 deaths

The data and data analysis is rolling in on the inequality of COVID-19 and it doesn't paint a pretty picture about our society. I'm not going to dwell on it here because regular readers will know how I feel about that and how hard I am working using my limited bandwidth and my personal need for emotional distance from the enormity of it all.

Let's just discuss air pollution, statistics and bicycling--three perennial favorite blog topics.

First off, read the synopsis of Exposure to air pollution and COVID-19 mortality in the United States: A nationwide cross-sectional study. You can also read the full study on Medrxiv.org

Results: We found that an increase of only 1 μg/m3 in PM2.5 is associated with an 8% increase in the COVID-19 death rate (95% confidence interval [CI]: 2%, 15%). The results were statistically significant and robust to secondary and sensitivity analyses.

Conclusions: A small increase in long-term exposure to PM2.5 leads to a large increase in the COVID-19 death rate. Despite inherent limitations of the ecological study design, our results underscore the importance of continuing to enforce existing air pollution regulations to protect human health both during and after the COVID-19 crisis. The data and code are publicly available so our analyses can be updated routinely.
This is only county-level data and represent deaths only up to April 22, 2020.*
Fig 1: Maps show (a) county-level 17-year long-term average of PM2.5 concentrations (2000‒2016) in the United States in 𝜇g/m3, and (b) county-level number of COVID-19 deaths per 1 million population in the United States up to and including April 22, 2020

A risk ratio is the amount of risk you incur relative to some baseline or reference group.  The baseline can be the average, e.g. the average air pollution exposure of all people.  In the case of Black people, the ratio is computed relative to all people who are not Black.

Earlier estimates pegged an increase of 1 𝜇g/m3 in PM2.5 with an 15% increase in the COVID-19 death rate.  The analysis was recomputed taking into account confounding variables to separate out the risks of being poor, being black, density, etc.

It turns out that living with more pollution is deadly--even just 1 𝜇g/m3 more of long-term PM2.5 exposure increases your risk of dying should you catch COVID-19 by 8%.  For context, the national average is 8.4, 𝜇g/m3.  The California legal limit is 12 𝜇g/m3 and the Federal limit is 15 𝜇g/m3.

Table 3: Mortality rate ratios (MRR), 95% confidence intervals (CI), and P-values for all variables in the main analysis.
Being black in the US is deadly.  COVID-19 is no exception.  It increases your risk of dying by 45% (1.45 ratio relative to non-black people living with the same air pollution exposure.) P-value** is a measure of how sure we are of the finding.  A lower number means we are more sure.  In this case, there is zero doubt that anti-black racism kills.

In contrast, the association with density is less certain.  A p-value of 0.40 means there is a 40% chance the association isn't really true.  Better than 50/50, but still weak.  The association with home ownership is stronger (but still weak)--perhaps because older people are more likely to own their homes?

Anyway, I just want to point out that racism and environmental racism kills.

Because of the link between air pollution and COVID-19 mortality, governments around the world are trying to keep people from getting back into cars.  In Los Angeles, we have done nada, zip, zilch.

An emergency bike lane in Bogotá, Colombia, March 2020. Photo by Gabriel Leonardo Guerrero Bermudez/iStock
* I live in anomalous Los Angeles County, the most populous county in the US with 10 Million residents and 88 cities.  In contrast, NYC has 8.4 Million people in 5 counties that make up 1 city.  I look forward to more granular analyses of within-county differences later.

** Physics professor Eric has written an excellent guest-blogger series clearing up misconceptions about P-values.

I've got a busy day ahead so I'll save the rest for later.