How things are going with the epedemic
By now most people are aware of the sequence of events that make up the covid-19 epedemic. It goes something like this; Testing = Cases. Cases are followed by Hospitalizations, Hospitalizations are followed by deaths. Deaths are not only the most important metric in determining the severity of the epedemic, deaths are also the most reliable metric. There is, however, another item that affects our knowledge of "how things are going," and it is involved in all of these aspects, that is reporting and the lag time involved in that reporting. The longer it takes to learn something the more out of date that information is, and all of this has to do with reporting.
Primarily using Florida as an example I have tried to show, here and here, how the lag time in reporting affects the daily coverage of deaths. The lag time of reportindg deaths not only skews the actual quantity of deaths on any given day, it delays our knowledge of "how things are going. In their Pandemic Planning Scenarios, the CDC put the reporting of deaths at betwen two and eighteen days with the mean number of days being seven days. This means that what we know about deaths nationally is about a week old. This also depends on the size of a state and how many different localities they depend on for their own reporting. Not to mention the efficiency or inefficiancy of their health departments.
Taking all this into account, it very much is looking like the "Sunbelt Surge" peaked in mid July. The only way that you can truly judge how thing are going is to look not at today but at the past. In addition as you look into the past you must look not at what was "reported" but rather at what was recorded, the actual date of death numbers.
Texas is a good example:
As you can see the most "actual" deaths in Texas occurred on July 14th. We are now three weeks removed so few if any cases will be added to that number. You will also notice that cases, which are reported quicker than deaths reached there peak at about the same time. Both cases and deaths began to fall and continue to fall.
A better example of this trend is Arizona:
Here you see that Arizona's date of death peak was on July 17th. Arizona is easier to track because it is a smaller, though not a small state with very few counties to "keep track of." There are only 15 counties in Arizona compared to 254 in Texas. Since Arizona really only has
two main population centers, Phoenix and Tucson it is more up to date than larger states like California, Texas or Florida. You can clearly see that Arizona's surge of cases peakes at the end of June. Since they record cases by date of test rather than date of confirmation, they actually show somewhat earlier peak in cases than most states would. Regardles whether judged by cases or deaths Arizona very much appears to have "passed peak."
A couple more examples are Florida which appears to have peaked on July 16th , though due
to lack of testing and reporting during the hurricane scare that is not as definite.
South Carolina which seems to have peaked around July 21st.
I would like to do California but their web site makes it difficult to find things and I cannot find their date of death chart, if they have one. It seems that there may indeed be other "flare ups" particularly in the MidWest and even in California but the sunbelt surge appears to be abating. The implications of this on the political landscape could be significant, but that will wait for another post.
