In late August, the Centers for Disease Control and Prevention (CDC) updated their provisional death counts page to indicate that COVID-19 was the sole cause of death listed on death certificates in only 6% of cases.
This fact was interpreted by some as only 6% of reported fatalities, or around 10,000 people, actually died of COVID-19. This misleading claim became social media fodder for groups contending that the pandemic has been blown out of proportion.
Unfortunately, the CDC meant nothing of the sort. The CDC states, “COVID-19 should not be reported on the death certificate if it did not cause or contribute to the death.” Assuming accuracy of reporting, this statement means that COVID-19 at least contributed to 100% of the deaths listed on the Provisional Death Counts page. You can read the CDC’s detailed guidance here.
Moreover, COVID-19 is listed as the “underlying cause,” or disease or injury which set in motion the events leading to death, on more than 90% of death certificates.
Consider, for example, a death certificate that lists COVID-19, pneumonia, and respiratory failure as causes of death. In order to provide a complete picture of the death, the certifier listed additional causes. However, this patient had pneumonia and respiratory failure because they had COVID-19. A U.S. standard death certificate, in fact, would specify that COVID-19, the underlying cause of death, caused pneumonia, which in turn caused respiratory failure.
Philosophers of science sometimes make a distinction between proximate and ultimate causation. A proximate cause is the immediate means by which something occurs. So, the proximate cause in the case above would be respiratory failure. The ultimate cause is the overarching reason something occurred. In this case, COVID-19 is the ultimate cause. Like flicking the first domino in a line of dominoes, COVID-19 sets in motion a sequence of events, with the last event before death being the proximate cause. (Caution: the proximate-ultimate cause distinction gets messy, but is often useful.)
When the NRA says “guns don’t kill people, people kill people,” they are making this kind of point. The gun/bullet is the proximate cause of death, but the ultimate or underlying cause of death is the person pulling the trigger.
It is the same idea here — respiratory failure may have proximately caused the death, but COVID pulled the trigger.
None of this should imply that counting deaths is easy. It is not; there are cases involving COVID-19 where the underlying cause is unclear. For example, if a patient with end stage cancer gets COVID-19 and dies shortly after, the death certifier may list COVID-19 as a contributing factor as opposed to the underlying cause. But, we know from the death certificate data that such cases are rare.
Additionally, there are legitimate questions about how much clinical evidence is necessary to count a death as being due to COVID-19 in the absence of a positive test. Does one count presumed or probable COVID-19 deaths in the official statistics? This concern was especially relevant early in the pandemic, when the U.S. did not have infrastructure for testing.
One strategy that avoids these difficulties is to look at excess deaths. This approach considers the general number of deaths over the past few years to create an expectation of how many people would have died this year if COVID-19 hadn’t happened. Then, one attributes the deaths in excess of this amount to COVID. According to the CDC’s excess death model, if anything, we are undercounting COVID-19 deaths by thousands. Although even then, some of these deaths may be attributable to indirect causes rather than COVID itself (e.g., those who avoid going to the hospital for fear of getting COVID, or the effects of isolation).
The exact number of deaths attributed to COVID depends somewhat on data gathering and interpretation practices. However, there is good reason to think it is close to the
reported numbers and, quite likely, is higher.
Edited by Riddhi Sood and Joe Vuletich