Study population and setting
The study set out to examine whether there was an increase in COVID-19 incidence following Black Lives Matter (BLM) protests in the US. The study crosswalked two publicly available data aggregator sites: one for the BLM protests that included the estimated number of protestors and their location, and one for COVID-19 statistics at the county level. BLM protests from May 25 (Memorial Day holiday in the US) through June 16, 2020 were included, and a control county in which no protest occurred within the same state was selected, selected based off of county population size and COVID-19 case numbers on the day of the protest. Multiple protests in a county were merged as a sum of protestors, and the differences in case rates were analyzed in the 3 weeks following the protest date, with separate analysis for week 1, 2, and 3 conducted. Additional cases estimated by each protestor was determined using quantile regression.
Summary of Main Findings
The study found that 326 counties had 868 demonstrations and an estimated 757,077 protestors in attendance over the 22-day study period. The median initial case rate in protest counties was 3.1 per 100 people, as compared to 2.9 per 100 people in control counties. While there was a statistically significant increase in the case rate where protests occurred, this was estimated as a median of 1.4 per 100,000 people in week 1, 5.4 per 100,000 cases in week 2, and 16 per 100,000 people in week 3. Using quantile regression to assess increases by protestors rather than by area with a protest, they found that each protestor added less than 1 case per 100,000,000 people in week 1, 2, or 3, which was not significant.
The study used publicly available data which aggregated across all counties. They calculated not only the by-protest rate of case increases, but the by-protestor rate as well.
By using county-level aggregation and the BLM protest aggregation site, there may be issues with misclassification, and the number of protestors may be misestimated. This is particularly challenging for determining the per-protestor change in case-rate. Additionally, the study could not control for other temporal trends that differed by county, and assumes that the change in case-rate were related to the BLM protests, which may not be the case. The study also summed multiple protests occurring in the same county together, while they may have implemented different social distancing and infection control protocols, or draw different populations. Finally, they attempted to define controls based on population size and baseline case-rate, but were unable to control for confounders that may bias estimates, such as differences in county-level mask mandates or lockdown requirements.
This study is an updated analysis following the BLM protests over the summer to determine if there were changes in COVID-19 incidence in counties.
This review was posted on: 19 December 2020