In this video, Jeremy Faust, MD, editor-in-chief of MedPage Today, outlines why COVID deaths are getting harder to count and recommends looking at an all-cause mortality framework to get a more accurate picture of mortality rates.
The following is a transcript of his remarks:
Hey, this is Jeremy Faust, medical editor-in-chief of MedPage Today. Thanks so much for joining us.
Today, I want to talk through an article on my newsletter Inside Medicine that’s about this question that we are so often asked about, which is: when a patient dies of COVID, did they die “of” COVID or “with” COVID? Was COVID the reason that they passed away, or was it just along for the ride or had some role in it?
Obviously, it’s very difficult because as many of us know, death certificates really just reflect people’s best guesses. There’s no real way to determine whether a death was primarily caused by COVID and therefore the death certificate should say “underlying cause of death, COVID-19,” or if that person actually had a heart attack and may happen to have COVID at the same time, so maybe COVID would be a contributing or multiple cause of that death.
Then the question that is totally unknowable: if they hadn’t had COVID at that time, would they have had the heart attack? So you get into this very difficult game in which you have to sort of try to figure out what happened in every case.
As a result of that, I and many others have been following not just COVID deaths, but also what we call “all-cause excess mortality” — the idea that we don’t worry about how many people die of COVID in a community — we just check to see if more people are dying than should of all causes and see if that tracks with COVID outbreaks.
Early in the pandemic, if someone had COVID on their death certificate, 95% of the time it was listed as the “primary” or “underlying” cause of death. This makes sense. We know that in early 2020, we were seeing the same kind of patient over and over. We were seeing patients gasping for air with terrible pneumonia all over their lungs; they had low oxygen levels, they needed oxygen support and mechanical ventilation.
As time has gone on and the vaccines have come and people have natural immunity, people still get hospitalized with COVID, but it looks less homogeneous. There are different kinds. It’s a patient with a heart attack with COVID. It’s a patient with COPD [chronic obstructive pulmonary disease] emphysema with COVID. Who’s to say how much of the blame COVID should have for each of those deaths?
As a result of that, if you track, as we do, the percent of COVID deaths that are listed as “primary,” that number has come down and down and down over the pandemic to the point where when we started, as I mentioned, it was 95% of the COVID deaths were listed and thought to be primarily due to COVID — the underlying cause — and now it’s closer to about two-thirds, 60%-65%. It has really fallen and it has fallen over time, and it’s because the kinds of COVID cases we’re seeing in hospitals are a little more complicated.
What’s interesting is that that doesn’t necessarily mean that the COVID deaths are not real — it just means that we’re a little less sure as to what gets the nod as the primary cause of death. It means the cases are getting more complicated.
What’s also interesting is that we still see COVID tracking with excess mortality and tracking with case counts. So, we know that the doctors and other people certifying death certificates are not too far off, because the numbers rise and fall with those outbreaks and with the end of those outbreaks. So what we see is that people are still correct, it seems like, in putting COVID on those death certificates, but no one could ever tell you what percentage [was due to COVID] — whether it was 100% or 0%.
Now, one question we get asked a lot and which I will add here is whether or not this is regional. What’s interesting is that if you look at the four regions of the United States the way the census bureaus break it down, the Northeast, the Midwest, the South, and the West, it’s actually quite interesting: there’s not a huge difference between those four groups. There’s some variations in terms of how much is considered “primarily” due to COVID, how many COVID deaths are listed as COVID being responsible. There are some differences, but some people are surprised to see that in the Northeast, sometimes that number is a bit lower.
The reason for that is actually because in the Northeast, people who are certifying these deaths are actually a little more likely to list COVID as a “contributing cause of death” than they are elsewhere. So, the “underlying causes of death” seems to be a little bit more straightforward. But then what happens is, and this is a little bit regional and a little bit age-related, is that when they’re not sure if COVID should even be on that death certificate, places like the Northeast are a little more likely to stick it on as a “contributing cause of death,” whereas in the South, they’re a little less likely to include that.
What that does is that creates an interesting artifact where it looks like more deaths in the Northeast are actually not COVID-driven, when in fact what that really probably reflects it that more people certifying deaths in the Northeast decide to include COVID as a “contributing cause” when it’s not the driving cause of the death, and the opposite being true in the South.
We’re also seeing a difference in age. As you get younger, from 18-49 to 50-64 year-olds, it’s also a little more likely to have variability. We saw at times during the pandemic that the percent of COVID deaths that are considered “primary underlying cause” of COVID has actually dipped as close to 50% in some places and in some age groups at certain times.
All to say: this is getting much harder. It’s not so straightforward as it once was where we really saw the same kind of patient over and over again. Now, it’s really difficult.
So, we have to track several pieces of information at once to make sense of where we are in the pandemic, because that’s what everyone wants to know. How many deaths is COVID responsible for? To what degree is it the “primary cause” — the underlying cause — or is it a “contributing cause”? And if it is a “contributing cause,” where does the needle go? Is it 20% related? Is it 80% a causality? We’ll never know.
We track that with all-cause excess mortality and we see how we’re doing. What we find out is that, when you compare today to previous waves, we are much better off in terms of all-cause mortality. At the worst of Omicron, for example, in the United States we had a 37% increase in all-cause mortality in January of 2022. That means that we had 37% more deaths of all-causes than we were supposed to have. In December of 2020, for example, right before the vaccines came on, nationwide 43% more deaths of all-causes than usual. Very, very bad.
Now, for this summer and even this fall we are beginning to be able to have confidence in the data we have, and we can say that excess mortality has been riding at around 8%-10%. Now, certainly we want it to be zero. And in fact, there were times in March, April, and May in various parts of the country where we actually had close to zero excess deaths, and in some cases maybe even fewer deaths than usual because of how bad Omicron was.
We actually had a pull-forward effect where the leftover cohort actually was already accounted for. In other words, people died a few months earlier, and therefore in March and April we had fewer deaths than we might have expected due to that pull-forward or harvesting effect, which is a tragic thing that we see in pandemics.
But overall this summer, things started to creep back and we’re still sort of in that 9%-11% range of excess mortality. That means that we’re still not at a normal place. We track that, we track the COVID deaths, we look at “of” versus “with,” incidental or coincidental, and we put it all together and we know that we’re better off than we were, but we’re not where we need to be.
So I hope that answers some questions that people have. Are people dying of COVID or with COVID? The answer is it’s a little bit of both, or it’s a lot of both and it’s hard to know, and we use excess mortality and COVID waves to try to put that into context.
All to say that we’ve made progress, but we still have room to go. For more, check out my article on Inside Medicine, and thanks for watching MedPage Today.