Have the protests started the second wave of Covid-19? (update after 11 June data)
Updated: Jun 14, 2020
This will be my last update on USA George Floyd protest outcomes. It has now been 17 days since George's death and about 2 weeks since the large initial protests in the cities I am looking at. I think this is enough time to see if there are any changes in disease incidence. Any changes after this are most likely due to other issues.
Although total USA daily incidence increased again today, this is consistent with the seven day pattern of peaks and troughs. More cases are reported towards the end of the week than on the weekends. Average daily incidence appears about the same as the day of George's death marked by the green circle.
There was an increase today but this is consistent with the seven day cycle. Overall trend is still down.
Overall trend continues downward in Washington DC.
A spike again in Los Angeles. I may have been premature in calling a new downward trend. However, rates have plateaued over the last 14 days or so so there has been no increase in incidence that could be associated with the protests.
In Atlanta there appears to be no significant changes, with daily incidence still around 50.
New York continues its gradual decline and has shown no changes in this trend since George Floyd's death.
One article caught my attention today.
This article was reprinted in the Sydney Morning Herald and no doubt elsewhere. It is so misleading and biased I am calling it FAKE NEWS! Their main error is that they are concerned with total new cases rather than CHANGES in new cases. Here are a few points:
In fact, an Associated Press review found that demonstrations have taken place in every one of the 25 US communities with the highest concentrations of new cases.
It is nonsensical to merely look at total new cases without comparing these numbers to before the protests.
And in some of the communities, such as Minneapolis, the number of people hospitalised with COVID-19 has been rising.
I have been covering Minneapolis and this has shown one of the most consistent downward trends of any city on my list in terms of new cases. If we are to look at effects of riots, the relevant statistic to look at is daily new cases (called incidence in epidemiology). Hospitalizations are a separate issue completely. They lag new cases as people get sicker. We can see from the Hennepin graph there was a peak in new cases just before George died. The hospitalizations are a natural result of this peak. I can predict with great confidence that in the next few days hospitalizations will start to decrease as the peak incidence was passed about 20 days ago.
Ramsey and Hennepin counties, home to the Twin Cities, ranked seventh and eighth for the highest per capita new cases in the AP analysis. Each has seen more than 250 cases per 100,000 population in the past two weeks, together reporting nearly 2000 new cases in the last week alone.
As I have already explained, it is completely wrong to solely use daily new cases as a measure without comparing it to what occurred prior to the event. There was a peak in incidence around day 50 BEFORE George died and incidence has been falling since then.
Crowding protesters who have been arrested into jail cells can also increase the risk of contagion. An AP tally found that, thus far, more than 5600 people have been taken into custody.
At last something I agree with!
Osterholm and other public health experts note, however, that the protests aren’t necessarily as alarming as other events that could fuel new cases because they take place outside and many people are wearing masks
Michael Osterholm is one of the experts I listen to and I am pleased he raises exactly the same points I have been pushing in this blog. Pity the authors of the report did not run their analysis by him first.
Public health experts said it will take two to three weeks to know whether the protests cause a surge in coronavirus cases. And even then, they can’t definitively tie it to the demonstrations.
The unrest is happening in tandem with the reopening of gyms, hair salons, restaurants, parks and beaches. It also comes on the heels of the Memorial Day weekend, when many people attended large gatherings, so experts already were bracing for a case increase, said Jennifer Nuzzo, an epidemiologist at Johns Hopkins University.
I agree with this as well, which is why this is my last post on this matter. We are moving beyond the window where changes can be attributed to the protests and not other circumstances.
Hundreds of people also were arrested in Chicago, where Cook County has had among the highest per capita rates of new cases of any large county in the nation, with 283 new cases per 100,000 population in the past two weeks.
This is more nonsense. I have not been covering Cook county but here is the graph for Cook, Illinois.
This is in a similar situation to Minneapolis but the recent downward trend is even stronger. They are also coming off a peak about 40 days ago.
Unfortunately reporters get to publish nonsense like this and there is little recourse. I have already written a letter to the editor of SMH about this report and will try and contact the original authors. Reporters generally have no background at all in epidemiology or data science so they should not be in a position to write pieces like this that seem authoritative but have little basis in fact. I strongly dislike Trump but this is FAKE NEWS!
EDIT: 14 June 2020
This graph is showing hospitalizations in Minnesota. Note that they are currently decreasing and have been since May 23, before George's death. The article I am criticizing by MICHELLE R. SMITH and NICKY FORSTER was dated June 3 so they have no excuses to incorrectly claim that hospitalizations were increasing in this state since George's death.