CDC’s Latest COVID-related Deaths Data

The vertical axis is deaths per million. Spot the outlier.

Click for a bigger image. Via Kevin Drum.  Note that COVID death counts are likely undercounts as they do not include all “excess deaths“. [CORRECTED]

This entry was posted in COVID-19. Bookmark the permalink.

11 Responses to CDC’s Latest COVID-related Deaths Data

  1. Eric says:

    According to Dr. Deborah Birx, the United States is taking “a very liberal approach to mortality . . . if someone dies with COVID-19, we are counting that as a COVID-19 death.”

    That could also explain how a person dying (in Florida) from a motorcycle accident was initially counted as a Covid-19 death as well.

    I wonder if mixing correlation with causation has anything to do with the United States being such an outlier.

    • Bill Wilson says:

      Eric, you might be surprised, but the same facts exist in other countries. Died in an accident, but the cause of death is covid 19.

      • Eric says:

        Many years ago, when I was still a bright-eyed and bushy-tailed law student taking a class on torts for the first time in my intellectual journey to legal enlightenment, I found the concept of causation to be unusually subjective and prone to the unforeseeable whims of the finders of fact.

        In a situation not unlike a jury trying to determine if Ms. Palsgraf’s injuries were caused by well-intentioned but overzealous railroad guards or merely the ever-present effects of gravity in an inherently unstable and chaotic situation called life, a medical examiner must determine if the coronavirus was the proximate cause of death or simply one of many passengers in a person’s body that was traveling along for the inevitable ride to the grave.

        Without thinking too much of the discrepancy between my analogy and the actual legal issues present in the holding of that case, the concept of causation is inherently ambiguous and totally dependent on a person’s arbitrarily chosen frame of reference. The issue will always depend on what a person deems to be relevant or irrelevant in relation to the seemingly infinite amount of variables in the equation.

        Boldly taking the analogy where no metaphor has gone before, assume that the coronavirus is a bad actor that invades the body of a reasonably prudent person minding his or her own business in the overpriced aisles of Publix. According to the data, the effects of this invasion are de minimis in the main but potentially disastrous in the extreme. In such outlying cases, these people have comorbid conditions that hearken back to the concept of the eggshell doctrine. But for the fragile condition created by the comorbidity, a person would have survived the relatively mild effects of the coronavirus.

        But then again, depending on how you frame the issue, any pre-existing condition could be viewed as the ultimate cause of death notwithstanding the introduction of the coronavirus. Phrased differently, could the coronavirus be considered a normal part of the environment like any other mild pathogen subject to other variables in play?

        Due to the one-way direction traveled by the arrow of time, human beings tend to view cause and effect in the context of past and present without regard to the feedback loop within life’s continuum. It seems reasonable therefore to assume that coronavirus was the cause of death in a patient with a weakened immune system, acting in a fashion akin to the proverbial straw that broke the camel’s back. The effect of gravity, up until the addition of the final straw, was not enough to cause the break. This viewpoint frames the issue in terms of weight increasing over time and gives ultimate significance to the straw itself when seen as a chain of similar events.

        However, is there anything inherently wrong with viewing gravity as the ultimate cause of the collapse?

        It seems to be no different than the concept of burden shifting in the law. On the one hand, we can start with the presumption that a person was living in perfect harmony with a pre-existing condition that was eventually destabilized by the coronavirus. On the other hand, we can also start with the presumption that a person was considered to be dying in the first place for the very reason that a coronavirus or other seemingly innocuous variable of life was now potentially lethal given the body’s current state of decline. It was not the virus itself that caused the death but rather the person’s ecosystem, much like gravity, which acted upon the pathogen and forced its expression within the environment in a radically different way.

        I suspect that the Roman philosopher Seneca would not list have listed the coronavirus as the cause of death in people with significant pre-existing conditions. Then again, taking his philosophy to the extreme, he probably would not have helped the analysis in any meaningful way by ironically listing the cause of death to be nothing other than life itself:

        “It is not the last drop that empties the water-clock, but all that which previously has flowed out; similarly, the final hour when we cease to exist does not of itself bring death; it merely of itself completes the death-process. We reach death at that moment, but we have been a long time on the way.”

        So in conclusion, it seems to me that the lethality of the coronavirus, like everything else in life, is ultimately subjective and wholly dependent on the way that society chooses to frame the issue.

        • I am not on board for the claim that — for legal purposes — it is in any way useful to trace all causation to the Big Bang, even if in some physical sense this is likely true.

          In law, we deploy the concept of causation — or, more precisely, the somewhat imprecise term ‘proximate cause’ — as a way of determining which human agency, if any, might fairly be charged with responsibility for an event. A simple example is the so-called ‘last clear chance’ rule that is often deployed in auto accidents. We assign substantial liability for the tort to the party that could, last in time, reasonably have prevented it. We don’t assign legally relevant blame for the accident to Henry Ford, even though in the cosmic sense he may have contributed to it.

          I also think I disagree with the final paragraph of your post. The lethality of a virus is an ascertainable fact, at least within some margin of error. Of course “lethality” will vary depending on how we define it: for example, if it turns out that there is a class of long-term illnesses that kill the host after, say, a year, we could choose to count them or not depending on how we wanted to use the final number. Or we might have a suite of ‘lethality’ numbers just like we have several measures of inflation or unemployment.

          The actually facts on the ground are, I believe, objective and substantially ascertainable. Any statistic we construct depends on the purpose for which we intend it.

          • I don’t follow this:

            As the population of the United States continues to increase by more than 4 million live births each year, is it not possible that the last few years simply created a critical mass of the elderly population that is suddenly reaching its end of life after being prolonged so long from medical advancements?

            I don’t see why an increase in the population of young people would cause more old people to die. I don’t see why, if the population is bigger, a constant number of old people dying would express itself as anything but a smaller percentage of the population. I don’t see why if medical science prolongs life X years then this runs out, we’d get a mass die-off since that would by hypothesis only effect people of a particular age cohort per year.

            And finally, while ‘causation’ has many meanings, there’s not much difficulty in saying that if a person dies from COVID while they have a co-morbidity, we might frequently identify COVID as the proximate cause or a joint cause.

            • Eric says:

              I was just implying that as new births have increased the population over decades at an ever increasing rate, it would make sense that more old people would be dying every year in general as these babies grow into the elderly population.

              (It was late at night when I typed it up so it’s not my best work by any stretch of the imagination. I am going to visit my mom to watch the weekly Sunday night family movie so I will try to explain my haphazard analogy about causation in a little bit.)

          • Eric says:

            I completely agree that my reference to gravity and by extension the forces of the Universe leading all the way back to the Big Bang in a fully extended chain of actual “but-for” causal reasoning does not reflect the current legal standard nor should it.

            However, in the context of a legal blog, I did think it somewhat appropriate to briefly reference the way that lawyers attempt to characterize the metaphysical problem of causation inherent in human reasoning. The focus on human conduct in a system, as opposed to cosmic conduct, simply highlights the fact that multiple methodologies exist in an effort to frame the relevant factors in such an intellectual endeavor. As to whether those cosmic forces are the direct cause of human conduct is a discussion on free will that unfortunately lies outside the scope of this current comment.


            But just like proximate causation attempts to refine the analysis in the context of law, the scientific method attempts to refine the analysis in the context of epidemiology. To that end, your statement about creating a framework that “depends on the purpose for which we intend it” mirrors the sentiment by Albert Einstein in one of his final works entitled “Physics and Reality.”

            “The whole of science is nothing more than a refinement of everyday thinking . . . In my opinion, nothing can be said concerning the manner in which the concepts are to be made and connected, and how we are to coordinate them to the experiences. In guiding us in the creation of such an order of sense experiences, success in the result is alone the determining factor. All that is necessary is the statement of a set of rules, since without such rules the acquisition of knowledge in the desired sense would be impossible. One may compare these rules with the rules of a game in which, while the rules themselves are arbitrary, it is their rigidity alone which makes the game possible. However, the fixation will never be final. It will have validity only for a special field of application (i.e. there are no final categories in the sense of Kant).”


            In other words, human beings are required to choose inherently arbitrary and subjective axioms as the starting point of further reasoning. Moreover, the concept of “success in the result” is also inherently subjective. In the context of proof, human beings tend to use induction in addition to deduction as their guiding light. But as David Hume discussed, there is nothing inherently rigorous about the inductive inferences used by people to identify with absolute certainty the repetition of future events from prior circumstances in the context of causation.


            It seems that people reason through subjective emotion that is hardwired into the brain; proof, therefore, depends on the varying individual thresholds of individuals. So when you state that the “lethality of a virus is an ascertainable fact” and that such facts are “objective and substantially ascertainable” in the context of Covid-19, I decided to review the underlying philosophy of the field. However, it seems that no clear standard exists: “Causation is an essential concept in epidemiology, yet there is no single, clearly articulated definition for the discipline.”


            I have almost completed my review of the medical literature from the very first detection of the virus in human patients to the replication of the symptoms in animal models. However, I will save that discussion for your most recent topic and comment accordingly when my research has been completed.

  2. Good science often comes with caveats. But Birx’s claim is missing an important one. The so-called “excess deaths” consist primarily of people who die outside a hospital setting. They are people who were not in the main tested in life or posthumously for COVID. So we don’t know to what extent those deaths are in fact COVID related. But the hypothesis that they are in the main COVID related seems strong.

    Let’s put it this way: something abnormal caused those deaths. What could it be? Maybe some fraction is due to knock-on consequences of COVID, e.g. depression, job loss. But surely far from all. Occam’s Razor suggests the cause of the lion’s share of those deaths is COVID. If so, the numbers Birx is lauding as “very liberal” are nonetheless missing cases that, in their own terms, they should catch.

    Similar arguments apply to those who want to attribute COVID+ cases to the +. Maybe that makes sense up to the normal rate of death for such people, But all those people on ventilators was not a normal thing.

    As to cross-country comparisons, they’re about as hard as cross-state comparisons given the differences in practices around the US; Keven Drum has written about this in the past. The is some wiggle in the number if you want to try to factor in national differences in accuracy and method (the accuracy being a bigger issue), but not enough to vastly influence the results at least in the industrialized world. I’m not sure anyone really knows what is going on in much of Africa and even some parts of South America, but I fear it’s not pretty.

  3. Eric says:

    I apologize if my link to Dr. Birx was not made with the proper context. The quote came from a press briefing on April 7 and described the way Covid-19 deaths were being reported. My comment suggested that the statistics were perhaps skewed by such an overly broad definition relative to the other countries on the chart. The example with the motorcycle accident was anecdotal evidence that other issues exist amongst the states that have not been caught and corrected, thereby implying the attribution of deaths to Covid-19 could be wrong.

    But I agree that the excess death rate for the year 2020 is out of proportion to the other years. When I explored the CDC’s website to explore mortality rates from 1999 to present and examined the months of January to July, I noticed that death rates were mainly flat over time with only occasional excess spikes ranging from an extra 60,000 to 80,000 a few times. As it concerns the 2020 mortality data for the months of January through July, there was a spike of about 200,000 extra deaths so far this year. Compared to the other spikes, that is an unusual number that is 120,000 more than the other spikes in the data.

    I believe that something is indeed causing more death than normal. But there are questions that need to be answered surrounding the circumstances of the deaths, such as New York potentially spreading disease in nursing homes back in March or the possibility that ventilators cause much more harm than good.

    But as to the lockdowns and masking mandates being employed by the states, I still have not adopted that policy standpoint. I think that the long term effects will be so utterly devastating to the majority of the population that no amount of good intentions can justify the collapse of the economy. While I do not wish for anyone to die, regardless of demographic, it makes me wonder if the world at large simply substituted one particular group of people for another.

    To that end, I will see your Occam’s razor and raise you the law of unintended consequences.

    • COVID-19 very disproportionately kills older people; we don’t know about the longer-term effects on younger survivors. As much as 40% of the death toll may be nursing home patients and staff, and there’s plenty of blame to go around for that, both in NY and in states like Florida where the first, second, and third impulses of nursing home operators and of the state governor were to coverup, deny, and obfuscate. Some of the blame may also lie with the federal government, which took no action to prevent this (e.g. not regulating, failing to making testing easier, for example), and indeed took actions that enabled it (e.g. controlling supplies of PPE and doling them out erratically).

  4. Eric says:

    Perhaps the answer to this problem is much simpler than we have considered.

    Upon further review of the mortality data from 1999 to 2019, it appears that deaths are gaining significant momentum. For instance, I noticed that the same number of people died in the 11 years spanning 1999 to 2009 (26,707,605) as the 10 years spanning 2010 to 2019 (26,712,753). While there were some years in the first range that had significantly negative excess deaths, every year in the second range showed positive increases.

    As a better way of viewing it, consider that the overall yearly mortality amounts numbered anywhere from 2.39 million to 2.47 million in the first range without changing much between 1999 through 2009. By comparison, the overall yearly mortality numbers started at 2.47 million in 2010 and kept increasing every single year to become 2.85 million in 2019.

    In other words, the maximum difference within the first range of 11 years was only 0.08 million while the difference in the second range was 0.38 million. Even assuming that there are 200,000 more deaths in 2020 than 2019, that’s only 0.2 million extra deaths compared to the 0.15 million relative increase of total deaths from 2015 to 2019.

    As the population of the United States continues to increase by more than 4 million live births each year, is it not possible that the last few years simply created a critical mass of the elderly population that is suddenly reaching its end of life after being prolonged so long from medical advancements?

    Perhaps the novelty of this particular strain of coronavirus is pushing them over the edge and represents the accelerating trend of this particular demographic. Moreover, it’s also quite possible that there are multiple respiratory viruses happening at the same time this year and our attention has unduly been focused on SARS-CoV-2 to the exclusion of others. Is it so impossible to believe we are having a bad influenza year and a bad coronavirus year at the same time without properly distinguishing between them?

    For reference, I am providing the figures below from the CDC website that shows the total deaths and increase or decrease from the previous year.


    2,391,399 (2.39)

    2,403,351 (2.40)

    2,416,425 (2.42)

    2,443,387 (2.44)

    2,448,288 (2.45)

    2,397,615 (2.40)

    2,448,017 (2.45)

    2,426,264 (2.43)

    2,423,712 (2.42)

    2,471,984 (2.47)

    2,437,163 (2.44)

    2,468,435 (2.47)

    2,515,458 (2.52)

    2,543,279 (2.54)

    2,596,993 (2.60)

    2,626,418 (2.63)

    2,712,630 (2.71)

    2,744,248 (2.74)

    2,813,503 (2.81)

    2,839,205 (2.84)

    2,852,584 (2.85)

Comments are closed.