Were excess deaths merely shifted?

Posted on by

A reader asks whether the excess deaths we are seeing now are mostly people who would have died later in the year, so that for 2020 as a whole the epidemic will not stand out. My thoughts.

1. For what it’s worth, the models say otherwise. Experts who look at models expect another wave of deaths as social distancing is relaxed.

2. I am willing to speculate against the models. What I am seeing is a very gradual decline in the death rate. The death rate will remain high enough over the next month to ensure that 2020 will be a bad year for excess deaths. But we may very well see months later in the year in which “excess deaths” are negative, because of the death in March-May of people who would have died later in the year.

3. During the first two weeks of April, we saw cumulative deaths rising at 10 percent per day or more nearly everywhere in Western Europe and in the most hard-hit U.S. states. Now, we are seeing increases of 10 percent per day almost nowhere. Not in Sweden. Not in Florida. Not in Texas. There is a very broad based slowdown. Too broad based to be accounted for by lockdowns, which were imposed in different degrees and at different dates across all of these regions.

a. Perhaps the slowdown is an illusion, and we are going to get a second wave. That is the story that the models want to tell.

b. Perhaps the virus has already attacked the weakest victims. Recall the Avalon Hill model, in which there are offensive factors and defensive factors. Perhaps the people with weak defensive factors have already been attacked, and what remains are harder targets: elderly in well-run nursing homes that are keeping the virus out; or people who themselves have a decent defense against dying of the virus.

c. Perhaps the offensive factor is down because the virus has mutated to be less deadly. I remember an early Weinstein-Heying podcast where Bret said, “The virus doesn’t want to kill you. It wants to get into the future.” Their point was that the virus would increase its survival chances by mutating into a less deadly form. They did not give a time frame for this benign mutation trend to emerge, and it may be improbable that it has taken place already.

By the way, here is a long NYT article on excess deaths.

31 thoughts on “Were excess deaths merely shifted?”

  3. Here is a viewpoint that while virus-related deaths are real and tragic, deaths as a whole are not increased at all. Auto-accident deaths may be way down due to less driving.

    This article is from RedState, which is not impartial at all.

    Still, the article does slow down the drumbeat about us being in a plague that requires semi-permanent lockdowns.

  5. I believe i saw a study out of Italy which concluded that the average person who died from covid had roughly 10 years of life expectancy remaining.

    • I saw the same study, and it was basically bull****. The problem was that they assumed the dead were just like the composite of their age demographic. For example, they just assumed that a 75 year old man that died from COVID-19 had the average remaining life span as the age demographic itself had pre-COVID. Seriously- that was all it had. This is exactly the same as saying all 75 year old men who died of heart attack had 9 remaining years of life.

      If you want to find the actual remaining years of life that COVID-19 victims actually had in absence of the disease, you have to divide them up better based on co-morbidities. As for actual data, you will eventually be left to studying all-cause mortality over the next several years. I think enough people have died in parts of Europe and the US that the signal will be detectable for the following 1-3 years, at least. When that signal vanishes into the noise is the point where you can probably estimate the actual life years lost by the COVID-19 victims.

  6. I’ve been wondering why some people get infected and some people don’t. Even within households, the infection rate in China was only like 10%. Given how easily the virus spreads before symptoms, that can’t *just* be isolation. The diamond princess only had 20% infected in a situation that seems about as bad as could be. We really don’t have any examples of a population where more than about 1/3 of the people were infected.

    I guess epidemiologists just model this as random. Which it probably is. But I can’t help but wonder if a lot of people (like say half or 2/3), while not necessarily immune, are much more difficult for the virus to infect.

    If we take that logic to eg NYC, let’s imagine the virus encountered half the population. Of those, 60% were “hard to infect” people and 40% were “easy to infect”. So ultimately you end up with 20% infected. A situation like this is very different from one where everyone is equally likely to be infected.

    • In certain environments, a much higher percentage get it. There was that infamous choir practice in Seattle, most got it. Nursing homes and homeless shelters and some other party events have higher infection rates. If you believe that 21% number from NYC as an average, then some places are lower and some are higher, probably with fat tails, so much higher. There were probably subway cars where nearly everyone caught it, but of course that’s impossible for us to figure out now.

  7. I actually think the excess deaths metric paints a decent picture of the overall impact of this virus, and it doesn’t depend on differences in testing or reporting. It aggregates everything like a price. Deaths from the virus, deaths from people not seeking or unable to obtain medical care for other reasons, suicide, homicides, highway fatalities, etc. are all baked into the pie.

    For example, this article points out that “in Ecuador’s Guayas province, just 245 official Covid-related deaths were reported between March 1 and April 15, but data on total deaths show that about 10,200 more people died during this period than in a typical year — an increase of 350 per cent.” So it’s possible the impact of the virus in Ecuador is much more dire than the official numbers indicate.

    • “I actually think the excess deaths metric paints a decent picture of the overall impact of this virus,…”
      This statement is perfectly correct, but for the last word in it. Because it’s very hard to say now how many of excess death are caused not by the virus but by our overreaction to it, panic and fear of seeing a doctor that led non-covid deaths to also increase. We have some preliminary evidence that this is quite significant
      eg Oregon.

  8. For “all US people”, the CDC says a 75 year old’s expected life is about 11.5 from an older table, less for men, less for blacks. “Around 10 years of life expectancy”.
    https://www.cdc.gov/nchs/data/hus/2010/022.pdf

    Of these people at any advanced age, some die sooner, some later. Most of those dying from COVID-19 have co-morbidities, often 2 or 3. The table doesn’t show this, and I’d guess it’s hard to find, but each co-morbidity probably reduces that life expectancy. (Big data will be giving better numbers in the near future, perhaps.)

    So I’m fairly sure an estimated life expectancy which took more medical states, including obesity, into account, would show the statistically least healthy are the most likely to die, if they get infected. And ventilators often do not save them (30% or more fatalities? 70%?).

    So if you take all the people who would die in the next 3 years from health reasons, and half of them die whenever they get the Wuhan virus, there will be some slightly(?) significant excess deaths now, but over a few years it will smooth out.

    • Merely to illustrate the problem, I’ll make up a bunch of numbers, but use one real number, the average life expectancy of an 80 year-old American female, which is 10 more years (!)

      Ignoring the costs of health harm to survivors:

      Let’s say you have 100 of these women, and the virus fatality rate is 25%, so 25 die.

      20% (20 of the 100) had underlying conditions (UACs), they had life expectancy of 2 years, and all of them die (20 deaths).

      80% did not have underlying conditions, have life expectancy of 12 years (comes out to that 10-year average) and only 6% of those die (5 of them).

      Ok, the naive estimate is that 25 deaths * 10 years/death = 250 years lots. The “fatality rate” is the same as the “life lost rate” (LLR).

      But in this case, 20*2 + 5*12 = 100, which is a 10% LLR; less than half.

      And note we’re not doing any “quality adjustment”. If you discount the UACs by half, then the QA-LLR is only 8%.

      Now, 25% is more than triple 8%, and that large factor is enormously important for any attempt to make a rational policy decision about the consequences of continuing Lockdown Socialism.

  9. Perhaps the slowdown is an illusion, and we are going to get a second wave. That is the story that the models want to tell.

    Is “models” your standard ‘pejorative’ here? Because, AFAIK, none of the ‘models’ that you’ve been criticizing cover anything more than a single first wave. Multiple waves of infection is a completely independent ‘prior’ of the epidemiologists and others – not a feature of the models themselves.

    • Actually, if you examine them in granular detail at the state level in the US, many of the models do want to show the second wave in the areas where deaths haven’t been off the charts already, but still appear to be on the right side of the bell curve. In other words, the models predicted, for example, far more deaths and cases in California, but those never happened, but the models, adjusted for the data in hand still try to predict far more deaths and cases in the near future.

  10. After the French 2003 heat wave caused an overmortality of about 15000, subsequent mortality rates showed a return to normal within 6 months. The summer dead were not going to die soon. They were fragile but could have survived much longer. You can live a long time with water up to your shoulders before the wave comes.

    https://www.inserm.fr/sites/default/files/2017-11/Inserm_RapportThematique_SurmortaliteCaniculeAout2003_RapportFinal.pdf

    • An interesting fact about France in the event- in the 18 years prior to the 2003 heatwave, in July and August, a average of 84,000-85,000 deaths were recorded, and it was tightly range bound between just short of 86,000 and 82,500. In 2003, it spiked to 102,000- an excess of about 17,000-18,000. In 2004 and 2005, the average for those two months was 79,000- a total drop of 8,000-9,000 for those two years. I think it could be claimed that the 2003 heatwave accelerated some deaths that would have been heat related in subsequent years. Additionally, it is quite likely that COVID-19 has accelerated the deaths of people who would have died in the the following 1-3 flu seasons, anyway, at a minimum.

  11. 3.d. How much of the trend may be attributable to the cold Northern hemisphere spring finally relenting and more people opening their windows and walking outside?

    One wonders too how much excess mortality will be due to governors prohibiting older Americans from walking in parks and at beaches? Steps per day is positively correlated with respiratory health. https://onlinelibrary.wiley.com/doi/abs/10.1111/ggi.12895

    • So people who would have died anyway very soon because they were weak died because they couldn’t leave their sickbed for their daily jog?

  12. All cause deaths 2020 minus 2019 or preferably an average of previous years might be negative later in this year but it might also be negative earlier this year. One of the Ioannidis videos says that Italy had fewer influenza deaths in the months before covid-19.

    Better yet: More informative than the average of previous years is a graph of the ‘n’ recent years max and min plotted underneath the 2020 points. I have not seen one.

  13. What I am seeing is a very gradual decline in the death rate.

    Actually what you’re seeing is an explosive growth in the death rate. A 3DDRR of 1.15 translates roughly to a death rate that quadruples every month if I’m doing my math right. This is somewhat slower than it was previously growing, but declaring victory seems premature (unless you’re calling it for the virus, of course).

  15. I’m very skeptical about the benign-mutation conjecture.

    Consider two types of the virus in evolutionary competition. Type A multiplies rapidly within its host’s body and releases a great many copies of itself into the environment in a short time before the host’s death. Type B multiplies more slowly within the host and releases copies of itself more slowly, during its hosts’s long life.

    As long as there’s a dense population of prosepctive hosts, Type A will have the advantage: it’ll spread quickly to many new hosts, and will out-compete Type B in hosts infected with both types. Only after a very large fraction of the hosts have died off will Type B have an advantage, in that Type A will kill its hosts before they can encounter and infect new hosts.

    Moreover, the development of more benign forms of a pathogen really represents co-evolution of pathogen and host. On the one hand, the rabbits that’re genetically most susceptible to myxomatosis will die off, leaving those genetically lest susceptible, or less severly affected. On the other, the most virulent and deadly strains of the virus die out as the decreased density of rabbits leaves them at a disadvantage relative to strains that don’t kill their hosts so quickly. The time-span for this process of co-evolution will be measured in generations of the host species: so, for human-pathogen coevolution, on the order of centuries rather than months.

    • Sorry: in that last paragraph, “least susceptible, or less severely affected”.

    • And, to continue on the subject, a hybrid AB strategy would be better still: Be benign while your host is healthy and has the potential for a long life, but shift gears to rapid virulent reproduction if it looks as though your host might not live much longer. This could explain the high death rate among people with comorbidities: it’s not that diabetes makes one especially vulnerable to the virus, but that it triggers the virus’s imminent-host-death response.

    • I agree. It’s true that viruses tend to evolve to a form that doesn’t entirely destroy their host populations, but coronavirus was never in danger of doing that. Ebola (85% mortality) would probably evolve to have a fatality rate similar to smallpox (20-30%) given a chance, but coronavirus (>1%) is not hitting this limit.

      Also, the mechanism of this sort of evolution is that excessively fatal viruses destroy their host populations. As a member of a host population, this is not comforting.

  16. The following paragraph is a quote from the article. The statement by Bret may be more than a bit insane; I quote him here; “The virus doesn’t want to kill you. It wants to get into the future.” OMG!!! That moron seems to think that the virus is sentient.

    [This totally distorts his point. He was talking about the evolutionary process. He was speaking metaphorically. Your commment is not appropriate.–ed.]
    c. Perhaps the offensive factor is down because the virus has mutated to be less deadly. I remember an early Weinstein-Heying podcast where Bret said, “The virus doesn’t want to kill you. It wants to get into the future.” Their point was that the virus would increase its survival chances by mutating into a less deadly form. They did not give a time frame for this benign mutation trend to emerge, and it may be improbable that it has taken place already. Me; I believe this puts a sane perspective on the debate

Comments are closed.