VDH Cleans Up COVID Deaths Stats

by Carol J. Bova

Anecdotes and social media comments have insisted that the Virginia Department of Health (VDH) has counted deaths from suicide, gunshot or motor vehicle accidents as COVID-19 deaths.

On March 19, VDH announced on its COVID-19 Dashboard that it had reviewed more than 10,000 reported COVID-19 deaths using the Virginia Case Definition for COVID-19 Associated Mortality. The announcement said of the reviewed deaths:

Among these, less than 1% (99 deaths) were determined to not qualify as “COVID-19 associated” deaths per the case definition and [were] re-classified to be a COVID-19 case that did not result in death. Today (3/19/21), a decrease in total net number of COVID-19 deaths is being reported on the VDH COVID-19 dashboard (more COVID-19 deaths were removed than added on 3/19/21).

Since the net change was -90, that would mean only 9 deaths were recorded on March 18.

Share this article


(comments below)


(comments below)


10 responses to “VDH Cleans Up COVID Deaths Stats”

  1. This was a useful first step by VDH. It should address some of the skepticism regarding the COVID death statistics.

    Next step: Distinguish between people who died with COVID and those who died because of COVID.

    1. VDH does say “VDH is counting people who died due to COVID-19, not people who died of some other cause while also having COVID-19.”


      The weak spot in their review is in their definition of how they count deaths. The first point is:
      1. During the case or outbreak investigation, the case investigator determined that the patient passed away due to COVID-19. This may occur through medical record review, talking with the patient’s healthcare provider, or talking with their family.

      Not sure why they didn’t stick to the death certificate and medical records and instead use talking to family.

      1. Stephen Haner Avatar
        Stephen Haner

        So, local doc or medical examiner signs a death certificate, and it either mentions COVID-19 as at least one cause or it doesn’t. If it doesn’t, why would there be any question? I remember reading on the CDC website that about 9% of the deaths it counts as COVID were people who died WITH but not OF, people who likely would have expired anyway. Are those the people we are talking about? Or were there people on the list with no mention of COVID on the death certificate?

        I’d lost confidence in this tracking long ago when we saw so many death certificates getting added so late, like over the past few weeks here in Virginia. I suspected intentional efforts to, pardon my use of the phrase, flatten the curve.

        Virginia’s per-capita deaths had been in the bottom ten among the states, DC and PR. With those additions in recent weeks, I think we rose and now are barely in the bottom third.

        1. VDH says they do count acute respiratory distress syndrome, viral pneumonia, and hypoxic respiratory failure, but don’t say whether those DC’s also say COVID.

          I am mystified how VDH reviewed over 10,000 death certificates when they added so many that were backlogged just a few weeks ago.

  2. LarrytheG Avatar

    Carol – you do a good job keeping up with this. Do you know WHO – the folks are out in the hinterlands that are submitting data. Is it an array of health people, doctors, funeral homes, etc or is it curated by only certain folks who can submit data?

    1. See pages 15 to 19 for details on “12 VAC 5-90-90. Those Required to Report” in the link below.


      1. LarrytheG Avatar


      2. LarrytheG Avatar

        Went to the referenced area but apparently there is a difference between reporting communicable diseases, and deaths from causes.

        It appears that once data gets to VDH that it is curated but not sure what they do if data is missing (like race) and/or contributing factors, etc.

        I’m largely ignorant of much of this while you have spent time and become much better informed.

        1. The reports go in through electronic portals. You’d have to ask VDH how they handle blank entries when they create datasets and reports.

          1. LarrytheG Avatar

            I’m assuming if they don’t have data, i.e. they show it as unknown or some such that the source that submitted it did not fill it in… and VDH accepted that form with no answer in that block.

            Even then, they still seem to be doing curating of other stuff… but when there is a crapload of folks submitting data – a certain amount of it is “dirty”.

Leave a Reply