Covid19 Update, May 23. Waiting for the turn.
Actual data, the state of things in a series of annotated charts:
Coming to the end of the week with slight declines in new infection cases. This may be a result of better testing, revealing what has been in the background all along. Or, as we move away from self isolation, it could be a rise in cases. Where I live in Central Oregon, an uptick in cases has been predominately among the 20-29 age group. These folks are now seen in the open pubs, packed cheek to jowl.
Now taking actual known data and forward casting it in models:
The data provider 1Point3Acres revised data back to March 25, but notably increased the number of infected cases by thousands for the last few days. That had the net effect of lowering mortality rates very, very slightly.
Here are the model runs:
As seen in the earlier charts, the saw-tooth pattern is from a drop in weekend reporting. Note also how this week in new cases is not significantly different from last week in magnitude. Mortality is down, but by an insignificant amount.
Looking at the relationship between known new daily cases and daily recorded deaths (now 96,483), see this superior correlation.
With experience and as cases rose, the slope of the modeled function declined. That decline represents a lowering of the mortality rate, here lagged by six (6) days from date of infection, that early in mid-March was over 9% and as high as 11% until the end of March. For May 22nd, the flawed lagged rate is 6.43%, and the 14-day average of the same is 6.48%. The six day lag yields the highest correlation, but it is determined using a denominator (actual infections) that is truly unknown. Some experts believe the true infection rate is 10 times, even 20 times higher than known. Given the number of new infections over the last week, we should expect daily deaths in coming days to be 1,250±300.
To view it another way, see this scenario chart:
This mortality calculation of 5.89%, -0.01% day over day, uses the straight across rate. That is, cumulative deaths to today, divided by cumulative infections.
A plot across the red dashed line, but employing the six-day lagged mortality rate, is seen here:
The two ways to consider mortality are useful, in my opinion. Also note that Italy and Germany have mortality rates (straight across) that are still rising, despite a significant drop in new cases.
That can be seen here in a comparison chart from 1Point3Acres for a suite of countries, including the US:
This turn is what I was referring to in my opening line. Notice how suddenly it happens. While the US has seen a slowing in daily new cases, we have yet to break the back on the spread of Covid19. We will in the years to come have significant opportunities to understand how to react if another pandemic comes upon us.
Finally, the cumulative and mortality mortality chart to June 28:
Have a fun Memorial Day weekend. But stay safe, and be safe.