By Jerald Hughes
& Scott Robinson
We are entering yet another period of new developments in the progress of the COVID-19 pandemic, which means complicating factors that make prediction and therefore decision-making very difficult. Having already passed through a long holiday sequence for Thanksgiving that disrupted the available data, we are now entering an even longer double-holidays sequence, for Christmas and New Year’s. At the same time, we are at all-time record highs on cases, deaths, hospitalizations, and positivity rates, each of which brings problems to those trying to model the pandemic’s development. This post looks at multiple metrics, to try and get at least some kind of best picture of what’s currently going on.
The following chart shows the numbers of COVID-19 patients currently hospitalized:
Chart: Currently Hospitalized (for Dec 21)
This chart shows a steady climb in the burden on hospitals. On the right side of the curve we can see what might be a peak; we would like to see the curve lessen its slope, then flatten out, then descend, indicating recovery. Just the last few days of data on hospitalizations start to raise hopes that a peak may be developing. The following table shows the numbers, including the change in the daily number; that number has been negative for 3 days.
Table 1. Average daily hospitalization increases
However, there is another analysis of this phenomenon, which is that hospitalizations are no longer climbing because they cannot climb – too many hospitals have reached their maximum capacity. We have been warned about this by health officials for some time. The logic is sound: you cannot report higher hospitalization numbers than you have beds to offer. To check on this, we note that hospitalizations are a lagging indicator. We should expect to see cases level out first, followed in about 12 days by hospitalizations, followed in about six more days (18 total) by daily deaths. This means we could perform a check on the meaning of the hospitalization data, by looking at previous data on COVID-19 cases identified by testing. The following chart shows the reports of daily cases since December 1; this gives us data from the post-Thanksgiving period, including the nine days from Dec 1 – 9. Given the 12-day lag between cases and hospitalizations, if there is a peak that has developed, then we should see it in that data.
Chart: Pre-hospitalization cases counts
This data does not support the idea that COVID-19 spread is coming under control. The current hospitalizations are the result of previously identified COVID-19 infections, after enough time has developed for symptoms to become severe and potentially life-threatening. Instead, we see that the relevant time period saw a sharp increase in case numbers. The table of the underlying data shows the leap in average daily cases from 161,036 to 207,277: an increase of 28.7% in just nine days.
On the other hand, the daily cases number itself also shows a recent flattening, albeit at a very high level. Perhaps those cases are entering the hospital much more quickly? - that is at least theoretically possible, although we lack any medical/physiological reason why this might be different from previous infections. We need to ask the same question about daily cases that we asked about hospitalizations: where do these numbers come from, and what is our capacity? So, next we look at the numbers of tests being given. In previous waves we saw huge lines develop in testing centers, delays in completion of tests, lack of personnel to collect samples and carry out the processing, and a consequent flattening out of the reported tests given. The logic here is that you can’t measure more cases than your maximum ability to test for them. In the current outbreak we are again hearing about long lines for testing. Here is the data on daily tests given.
Chart: Daily Tests Given (for Dec 21)
Table 2: Average daily case changes
As in all the data measures of COVID-19, the Thanksgiving holiday produced a disruption in the data. However, we see that after the Thanksgiving period, the line charting the numbers of tests performed did not continue on past its previous peak on Nov 25. Instead, it climbed back up to pre-Thanksgiving levels, and has increased only slightly since then. This introduces the possibility of further unknowns. If it is true that we have nearly reached testing capacity, then it has become difficult for us to measure what’s actually going on with new COVID-19 infections.
Medical experts during the first two waves and now as well in the third wave have also pointed to positivity rates among the tests that are given, as a measure of how well or how badly we are doing. Ideally, we want to be in about the 4-5% range or below. The following chart shows the aggregate positivity rates for the USA since Nov 1.
Chart: Positivity Rates
This chart shows positivity rates far higher than what medical experts say need to be observed, if the pandemic can be considered to be under control. Looking closer at specific regions, we see that the positivity rates rose to even higher levels after the Thanksgiving period. These are numbers not seen since the worst days of the first wave. Considering again the 12-day lag between new cases and hospitalizations, we see in this chart that the positivity rates were increasing up through the period ending 12 days ago, Dec 9. This data also does not support the idea that COVID-19 is under control. On the contrary, positivity rates over 11%, especially with so many millions of tests being given, are indicative of uncontrolled spread. We are heading into a period of very dense fog, in the struggle against COVID-19. We are measuring what we can, but our ability to measure in multiple areas is constrained by lack of resources. This blog article concludes that it is unlikely that we have transitioned into a time of declining cases, and hopefully, thereafter declining hospitalizations and deaths. Finally, as mentioned before, the approaching Christmas holidays promise yet two more problems. We will likely see more disruptions in data collection and reporting, for an extended period; and we are in terrible danger of another spreading surge caused by seasonal gatherings and events. Acknowledging the very high levels of unknowns, we consider it on the whole more likely that currently seen flattenings in hospitalizations and cases are indicating that the USA is running out of capacity in these areas – a troubling development in its own right.
Unlike previous simple two-day weekend disruptions, the Christmas holiday data disruption may last for a week or more, followed immediately by another disruption from New Year’s. It could be a very troubling, very long time, before a clearer picture of COVID-19 development in the nation emerges again.
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.