Medical experts urged families to stay home and limit their gatherings during the Thanksgiving holidays. Some followed this advice, but the Thanksgiving weekend also saw a high volume of travel.
The indicators of increased infections during this period will be cases, then hospitalizations, then deaths.
The CDC says that the time between exposure and the appearance of symptoms is usually 5-6 days. If we allow 1 or 2 more days for an individual to decide they need a test, go get it, and receive the result, that puts the window of time for new cases to appear in the totals numbers, at 6 to 8 days.
After a positive test, the period until hospitalizations appear in the data is about 12 more days.
Also after a positive test, the period until additional deaths appear in the data is about 18 days.
This gives us a calendar of events to watch for:
Thanksgiving: Nov 25 (Wednesday) through Nov 29 (Sunday)
New surge appears in data: Dec 2 (6 days from Nov 25) through Dec 7 (8 days after Nov 29)
Hospitalizations surge appears: Dec 14 (12 days from Dec 2) through Dec 19 (12 days from Dec 7)
Deaths surge appears: Dec 20 (18 days from Dec 2) through Jan 6 (18 days from Dec 19). This is [very messy] event study. In an event study, you examine a time series of data (typically, stock prices, consumer prices, etc) both before and after a 'shock' event, which is the event you expect to change the series from what you would normally expect, to a new series of results.
In the case of COVID-19, we are watching for the shock of Thanksgiving gatherings and travel to increase the numbers of COVID-19 cases, hospitalizations, and deaths, over what we would have expected, if those gatherings and travel had not occurred. This is *far* from a rigorous event study, because we have a very poor idea of what the numbers *would* have been without the Thanksgiving shock event.
Nevertheless, we have now reached the first visible milestone, when the first period of surge, Dec 2 - 7, should now be completed, and with our data falling outside the typical weekend dip. December 8, Tuesday, should have the bulk of the new cases, whatever their numbers turn out to be, already completed, with test results in and reported. If the Thanksgiving holiday slowed test processing and reporting, it may take a few days longer. In order to take a first look at these results, it is important to compare apples to oranges. Comparing Tuesday's numbers to Monday's will create a poor indicator, because Monday's numbers are typically part of the weekend dip. To avoid this, we will look at the most recent Tuesday, Dec. 8, compared to the previous Tuesdays, going back several weeks. This will give us *some* sense of how the third wave is developing, before, during, and after the Thanksgiving holiday.

Chart: COVID-19 Tuesdays Thanksgiving
Here we see Dec 8 presenting the second-largest absolute increase and second-largest percentage increase, since the third wave began. The *best* interpretation we can draw from this one data point is that the upward surge has resumed. The remainder of this week will give a clearer picture. As a very tentative conclusion on the question in the title of this post, the immediate indication is that the Thanksgiving holiday activities have not helped reduce the spread of COVID-19, and likely harmed us to some extent; how much remains to be seen.
The most recent 7-day average of cases, including the Dec 8 very high 213,498 count of daily new cases, is 202,172. Taking a *conservative* prediction of the number of hospitalizations we should expect 12 days later, on Dec 20, using the ratio 66.05% of hospitalizations/cases, yields a prediction of 133,534 hospitalizations. This is far more than the most currently observed hospitalizations number (Dec 8) of 104,600.
Using an also conservative estimate of a fatalities ratio at 18 days lag of daily deaths/cases of 1.50%, we should expect a 7-day average of daily deaths of 3032 by Dec 26.
All opinions expressed in this blog are solely those of the authors, and do not reflect those of University of Texas Rio Grande Valley, or any organizations of which either is a member.
Comments