Over the last two weeks (ending 12/19), most of the country has seen COVID infections slowly climb, and in some areas actually begin to level off.  But California has been a stark outlier, especially as it relates to the all-important hospitalization metric.

Hospitalizations

Hospitalizations Total hospitalizations, ICU bed usage and ventilated patients all increased over the last two weeks but the rate of increase slowed significantly.  Total hospitalizations and ICU bed usage were up by 13% and 9%, respectively.  Ventilator use increased by 11%.  All of these are about half the percentage increases two weeks ago and all three declined in the last 3-4 days.

However, the stunning aspect of the increase during this period is that nearly 60% of it came from California (7,450 of 12,739).  And the growth in hospitalizations in California showed no signs of slowing down during the two-week period.  New York also added nearly 2,000 hospitalizations but its growth seemed to be moderating in the last few days.

Other than California (and to lesser extent Arizona) we have not seen the surge in hospitalizations after Thanksgiving that so many predicted.  While California hospitalizations more than doubled since Thanksgiving, in the rest of the country they only rose by about 10%.  Hospitalizations have declined since Thanksgiving in 19 states.

Fatalities

Average daily reported fatalities increased from 1,780 to 2,529.  Most of the jump was in the first week.  In the second week, the seven-day rolling average stabilized at about 2,500 per day.

The 30-day lagging case fatality rate, which had been trending slightly above 2%, fell sharply to just over 1.6%.  Interestingly, the 20-day lag, favored by some, worked its way back up to nearly 1.7% after previously dropping.

The IHME model has once again slightly raised its year-end fatality estimate, now pegging it at just over 349,000.  That implies an average of about 3,400 fatalities per day between now and the end of year.  Unless we see another leg up on hospitalizations in the next two weeks, I have my doubts about fatalities jumping that much.

Testing

It appears we may be reaching a plateau on testing.  The average number of daily tests only rose by 10% over the last two weeks, which followed similarly modest growth in the previous two weeks.  It is hard to know exactly what this portends.  Certainly, the decline is not a capacity issue as tests have never been more available.  A few people I have talked to who have been tested at the City’s sites say they are mostly empty.

Positive test results increased by 21%, a bit more than the previous two-week period but still far below the 40-60% increases we were seeing in October and early November.  The ratio of tests coming back with a positive result ticked up from 10.3% to 11.2%.

Summary

Many talking heads in the media claimed the increase in COVID metrics over the last two weeks was confirmation of the theory that Thanksgiving gatherings would result in a surge of infections.  But the variation in the states’ experiences actually undermines the theory rather than supporting it.  As I mentioned, 19 states have seen their hospitalizations decline since Thanksgiving.  Are we to believe that the folks in these 19 states were more careful about their Thanksgiving gatherings than Californians?

This somewhat schizophrenic article in the New York Times vacillates between a tortured attempt to square the data from Thanksgiving with the experts’ predictions while having to concede that it really doesn’t.¹  The holiday hypothesis is frequently repeated by public health officials and talking heads in the media, but the truth is that the data has never definitively supported it.

To me, this underscores what I have increasingly come to believe – that the standard orthodoxy about how COVID spreads and what can to done to control it, is, at a minimum, a grossly incomplete understanding.  I think that in the years to come we will learn there were other factors (e.g., susceptibility heterogeneity, cultural norms, demographic characteristics, and genetics) that affected on how the virus spread and that will make the effect of our interventions pale by comparison.

From a public policy response perspective, California has done everything the experts have prescribed.  It was one of the first states to adopt a mask mandate and has had probably the most severe lockdown orders in the country.  And yet the hospitals in southern California are on the verge of being overrun with COVID cases.  Somewhere along the way we forgot that non-pharmacological interventions are designed to mitigate contagious outbreaks, not control them.  Historically, their effects on pandemics have been limited.

This pandemic will only end when we get to herd immunity through both natural and vaccine-induced immunity.  The good news is that we now have two vaccines approved that will significantly accelerate reaching that goal.

More on that topic later this week.

_______________________________________

Note 1 – The article correctly identifies California as the outlier post-Thanksgiving but for some inexplicable reason gratuitously mentions Texas generally and Houston, specifically, including an idiotic quote from the Harris Health CEO that there was no “super-spreader” event other than Thanksgiving gatherings to explain an increase of COVID patients in his system’s hospitals from 33 to 66!  Hospitalizations in Texas and Houston are up since Thanksgiving, but they have been increasing since early November and the rate has actually slowed slightly since Thanksgiving.