All of us who have been following the COVID data streams have known they are a mess.  We have known that there have been substantial delays in the reporting of both test results and fatalities, and differences between jurisdictions in their methodologies for collecting and reporting the data.

Texas made a major change in the way it counts COVID fatalities yesterday, that will provide greater clarity on COVID fatalities going forward but it has created a lot of confusion in the short term.  The details about exactly how the count was previously tallied and how the new system will work are still a little murky but here is what we know.

Under the old system, the State was apparently relying on daily reports from local health departments.  Historically, local officials have been responsible for collecting death certificates from doctors and funeral homes and reporting those to the State.  But last year, the State adopted a statewide electronic filing system that allowed doctors and funeral homes to file death certificates directly with the State, which must be done within ten days.  Once the State accepts the certificate it is available to local officials.  So, in Texas all officials should be on the same page about how many people have died from COVID.

But that is not what happened.

The new report shows fatalities through July 24 at 5,713.  The report the State issued on July 24 only showed the count at 4,717.  So the new count is a 996 (21%) increase.  We have known that there is a built-in delay in reporting fatalities because of the time it takes to complete and file death certificates, so no one should be surprised that a count based on the actual date of death is higher.  But that delay does not entirely explain this increase because it varied dramatically between the counties.

The largest discrepancy is in Harris County.  Its count jumped from 614 to 1100, an 82% increase which accounted for over half the statewide increase.  By contrast, Dallas County’s count only increased by two fatalities.  Somebody needs to explain why the Harris County count was so far off base.

Another head scratcher about the new count is that a number of counties showed reductions.  The largest was Hidalgo County whose count fell by 189 fatalities (433 v. 244).  I have no idea what can explain the count going down.

Another challenge with the new methodology is that it will make comparisons to other states more difficult.  The CDC has recommended to change the reporting to the actual date of death, but I have not been able to find anything to tells us how many other states have made this change.  For those that have not, we really will be comparing apples to oranges.

There is no question this will be an improvement going forward.  Also, the State has promised to begin using the death certificate data to improve reporting on fatality demographics.  But it also will underscore what many of us have known for some time – drawing conclusion from daily data is a fool’s errand.  The new count will not at all reflect what is actually happening on any given day for about two weeks.   Here is what Texas’ fatality chart looks like based on the new count.  However, the data for everything within about two weeks (right of the red line) will only show a fraction of the fatalities for those days.