In the last two weeks (ending December 6), the outbreak of COVID in the central part of the country slowly began to ease but has been replaced with more robust spread in some of the areas that had earlier surges in the spring and summer.
These time-lapse maps from the New York Times show this pattern. The NYT changed its color coding between the second and third maps to accommodate higher levels of positive test results which makes it somewhat difficult to compare. It is also skewed by the fact that the maps are per capita and because counties in western states tend to be larger geographically. Nonetheless, the sequence shows the geographic pattern of the outbreaks.
Total hospitalizations, ICU bed usage and ventilated patients all increased over the last two weeks but the rate of increase slowed, especially in the last few days. Total hospitalizations and ICU bed usage were up by 22% and 24%, respectively. Ventilator use increased by 37%. All of these are about half the percentage increase two weeks ago.
Hospital conditions have eased considerably in some of the hot spot areas. Iowa and Wisconsin, for example, have dropped by 37% and 27%, respectively. Almost all of the Midwest states had double digit declines.
But other states, New York and California in particular, have more than made up for the declines in the Midwest. In New York, hospitalizations have increased four-fold in the last month but are still only about a quarter of their all-time high in the spring. California is up three-fold and set a new all-time high for that state.
These numbers are pretty stunning considering the degree those states are still in various stages of lockdown and to which they have been repeatedly lauded by the media for their restrictions. Interestingly, Texas, Florida and Georgia, which have been repeatedly panned for their responses to the pandemic, have all seen a much slower growth in hospitalizations over the last couple of months.
Average daily fatalities increased from 1,270 to 1,780. The increases were larger in the last few days of this two-week period, pushing the weekly average over 2,000 for the first time since the spring. This is probably a combination of both increased fatalities and some catch up from underreporting in the first week because of the holiday.
The 30-day lagging fatality rate is still hovering around 2%. So far, it has not declined as I have been expecting. Interestingly, the 20-day lag, which is the indicator the data analysts at the New York Times have suggested is the best indicator, has dropped by about a half of a percent in the last few weeks. There is a fair amount of noise in these numbers for this period because of the Thanksgiving reporting disruptions, so we will have to see what they look like over the next couple of weeks to get a better idea.
The IHME model has once again slightly raised its year-end fatality estimate, now pegging it at almost 346,000. That implies an average of about 2,200 fatalities per day between now and the end of year.
For all the hoopla about how testing would decline over the holiday, testing actually went up by another 8%, averaging nearly 1.7 million tests per day. Twice in the last two weeks daily tests topped 2 million. Positive test results increased by 13%, which was down dramatically from a 69% increase two weeks ago.
The ratio of tests coming back with a positive result ticked up slightly from 9.8% to 10.3%. Since the increase in positive test results barely outstripped the increase in tests, it is fair to assume that most of the increase in positive test results for last two weeks likely came from increased testing.
The last time I reported to you on the country’s COVID stats, I said the near-term future would turn on whether the outbreak in the middle part of the country would follow the traditional bell epi-curve and whether the areas that had seen earlier surges would be susceptible to a second wave. It appears that the answer to both questions is going to be in the affirmative. The question now becomes the relative magnitude of those opposing trends. Especially important will be the degree to which naturally acquired immunity, and soon vaccination acquired immunity, can blunt those second waves.