"Covid-19 Cuts U.S. Life Expectancy by a Year in First Half of 2020, Biggest Drop Since WWII": What Exactly Does That Mean?

The headline is from NBC News, and the underlying CDC report (which has also been covered by other outlets) is here.

But I'm not sure just what this means. Under 0.2% of the U.S. population has died of COVID, even through all of 2020 and early 2021. If the average years of life lost as a result of each deaths was 10 years (see this estimate, which is focused on the first half of 2020 in the U.S.), that would itself reduce the life expectancy by just 0.02 years (0.2% of 10).

As I understand it, by now about 9% of the population has survived COVID, and even if we expect the survivors to have an average 1-year reduction in life expectancy as a result of long-term effects (but how would we know that at this point?), that would reduce the life expectancy by another 0.09 years, for a total of 0.11 years.

So where does the remaining life expectancy reduction come from? Did I do the arithmetic wrong here? Or does the great majority of the life expectancy reduction stem from other factors, such as neglected care for other illnesses stemming from the lockdown, from occasional lack of ICU beds, or something else like that (or perhaps suicides or drug overdoses stemming from the lockdown?)

Or is the CDC just projecting from the COVID deaths in the first half of 2020, assuming the death rate would remain constant going forward? Someone on a discussion list I'm on suggested that in response to my query. That seems odd, since it doesn't seem to take into account vaccination, the prospect of herd immunity being reached at some point, and the like.

I'm genuinely not sure about all this; health statistics certainly aren't my main field. But I do think there's something going on here besides just accounting for years of life actually lost to COVID during the epidemic. I'd love to hear what people who do know health statistics think about this.

Comments

[tuphat]

Although health statistics are not my field either, it's worth noting that they're talking about life expectancy at birth, "the average number of years that a group of infants would live if they were to
experience throughout life the age specific death rates prevailing during a specified period." Whether it's directly, indirectly or not at all related to COVID, more people dying at earlier ages -- as has been the vase recently -- has shifted the LEAB number downward, by about a year according to this report. I assume the statisticians know what they're doing.

[DaivdBehar]

Most of the excess deaths came from deaths of despair, under treated cancer and heart disease. All these excess deaths were caused by the Democrat Governor economic shutdown. Around the world, the shutdown, not the virus, caused millions of deaths by starvation. This is the biggest, fastest mass murder in history. It is 100% the fault of the left wing politicians. They are the agents of the tech billionaire owners of the media and of the Democrat Party. They did very well, with an extra $1.3 trillion in extra profits, from the unjustified shutdown.

[Nige]

'caused millions of deaths by starvation'

Are you sure they didn't all just get turned into newts?

Interesting upshot is, people who hold Trump responsible for the covid debacle in the US think he's an horrific failure, and people who think Democrats succesfully used covid restrictions to murder a few million people just because also obviously think Trump is an horrific failure for allowing them. Trump's an horrific failure in reality AND alt-reality.

[Brett Bellmore]

"Not acting like a dictator by seizing control over matters the states are constitutionally entitled to control " equals "Being a horrific failure "?

[Nige]

Can't remember any of his tweets saying 'I'd love to organise on the federal level to provide help and support states with their covid response but that would make me literallly Hitler, sorry.' Do remember him saying states weren't nice enough to him.

[WuzYoungOnceToo]

What Trump did/didn't tweet is irrelevant with regard to what authority he did/didn't have. We do however have many statements by Democrats asserting that he did not have the authorities in question.

[Nige]

Praising him for not exercising the authority he didn't have while he fucked up the authority he did is the saddest form of Trump apologism.

[DaivdBehar]

The shutdown, not the flu like virus took out $4 trillion from the economy. Millions of people living on $2 a day died. Tech billionaires took in an additional $1.3 trillion in 2020 over their profits of 2019. They whipped a hysteria in their media and by their Democrat Party agents. The net worths of Democrat politicians also shot up 40%. This is the biggest mass crime in history.

The dipshit lawyers here are oblivious. Child like bookworm naifs.

[Nige]

Thank God the Republians threw themselves in front of the bullet that was covid relief, think of how many millions of lives would have been needlessly saved.

[Alphonse Tomato]

Millions of people (in the US) were living on $2 a day? Cite? Perhaps I live a sheltered life, but I know a lot of poor people, including a few street people, and doubt that any of them live on $2/day ($30/month). Yes, corporate and bankster billionaires of all stripes raked in a lot of money for themselves. That includes a lot of Republcians, obviously.

[DaivdBehar]

Alphonse, take it easy. The poor in the US live on up to $65000
yearly budgets after counting the value of welfare benefits. The poor on $2 a day are living elsewhere. Their lives are just as valuable to them and to their families as ours.

[Armchair Lawyer]

So, I don't think you can assess the majority of the expectancy lost due to other causes. The below CDC graph shows the excess deaths. While there are a lot of excess non-COVID deaths, it's only about 50% of the number of COVID deaths.

https://www.cdc.gov/mmwr/volumes/69/wr/mm6942e2.htm#F2_down

What I think the issue is, is that 10 year number you're assigning is too low. A large number of COVID deaths occurred in people who had 20 or 30 years to live. 55 and 65 years olds. Even those over 85 had on average another 6 or so years to live.

It's the excess death rate in the "younger" (55-75) population that drives a lot of it.

[jdgalt1]

To get realistic numbers at all, we first have to deconstruct the reports so far, all of which vastly overstate deaths caused by Covid, mostly because hospitals get paid extra when they treat someone whom they can list as a Covid case.

For instance, if you believe the official numbers, flu deaths in 2020 were less than 20% of the usual number. I submit that the usual number happened and nearly all of them got misclassified as Covid deaths.

[Life of Brian]

For instance, if you believe the official numbers, flu deaths in 2020 were less than 20% of the usual number.

That's generous. So far in the 2020-2021 flu season, there has been one reported pediatric death. One. That's in the range of 1% of usual.

Data available here.

[Tilted]

I suppose that's possible, but to me it seems like the zebra explanation rather than the horse explanation. The drastic measures we've taken to reduce the spread of SARS-CoV-2 were likely to reduce the spread of influenza as well. So shouldn't we expect lower than typical influenza deaths this season? With all the isolation (particularly among those most likely to die from influenza) and distancing and mask wearing, among other things, I would expect an unusually low number of influenza deaths this season.

That said, seasonal flu deaths vary a great deal anyway. I just looked at the last 10 years and the relative standard deviation is 40%.

As for the reported number of deaths, yes, the typically reported numbers probably overstate the number of people who have actually died from Covid-19. The typically reported numbers don't seem to be limited to deaths where Covid-19 was an underlying cause of death.

But looking at the increase in the total number of deaths in the U.S. in 2020 from 2019, it (though I think numbers are still preliminary) was around 380,000. If we adjust the total number for 2019 deaths (2.855 million) by the average increase over the last 10 years (1.6%), then the increase for 2020 was around 330,000 more than what we might have expected. That's a little less than the number of deaths in 2020 with Covid-19 as an underlying cause of death, as reported by the CDC.

Total deaths in the U.S. increased by around 13% in 2020. The next largest annual increase that I see from the last 35 years is around 4%.

[Squirrelloid]

Masks, at least, don't have an impact on flu transmission. We have 100 years of studies to that effect.

Social distancing, otoh, matters a great deal. But less than 1%? Social distancing doesn't have that big an effect size.

As far as comparing year over year death totals goes, the problem is that Covid-19 isn't the only unusual change in 2020. There's also the lockdowns themselves, and all the unintended consequences (social isolation, increased stress because of unemployment, inability or unwillingness to get checked/treated for cancer, heart disease, strokes, etc...). So if there's only about 330,000 excess deaths, not all of those are due to covid-19. The UK is already estimating the impact of untreated or late diagnosed cancers due to the lockdowns will be at least the same order of magnitude as covid-19 deaths. (https://www.dailymail.co.uk/news/article-9276335/In-heartbreaking-pictures-human-face-cancer-epidemic-threatening-100-000-patients.html )

[Eugene Volokh]

Armchair Lawyer: The 10 year number is the one I get from the paper I cited; but even if the average years of life lost is 20, that still increase the diminished life expectancy stemming from COVID deaths from 0.02 to 0.04, nothing near 1.

[Armchair Lawyer]

Professor Volohk,

The other consideration is, as people note below, it's likely that it is the period life expectancy. The statistics extrapolate the current death rate, and need to assume that increased death rate going forward for the next 80 years (or whatever long range number you're using) for each age category.

So, if you see an increase in the death rate of 40% for 45-54 year olds, the statistics assume that increased death rate every year for the next 80 years. IE "perpetual COVID"

[Squirrelloid]

That's, ofc, a stupid assumption. Pandemics are blips, 1-3 year events, not perpetual.

[DaivdBehar]

If you are in a nursing home, you are moribund with a 25% chance of dying a year, instead of a 1% for others. Most of the COVID deaths were among moribund people. Perhaps, 100000 died from COVID, not with COVID. It is like a regular flu year. The flu deaths were markedly decreased at the same time in 2020.

The main purpose of shutdown was to elect a Democrat President to further the interests of the tech billionaires. Biden got a call, Harris. She could not get any votes on her own and was worthless to the campaign, politically. Black females hated her for being a mass incarcerator of their men. She is from the San Fran area, and is the prostitute of the tech billionaires.

[Joe_dallas]

fwiw

The median life expectancy for a male entering a nursing home is approx 3-4 months, , the median life expectancy for a female entering a nursing home is approx 7 months. (based on stats I recall from around 2010-2015). the average life expectancy is around 15-18 months due to few that hand around for 2-3 years.

But bottom line with nursing homes is that people go there to die.

[Dr. Ed 2]

Drug overdoses in the 25-35 population have gone way up.

[WuzYoungOnceToo]

A large number of COVID deaths occurred in people who had 20 or 30 years to live. 55 and 65 years olds.

Assuming that those people had those many years left to live based on the average life expectancy is invalid given the high % of those deaths that occurred in people with underlying serious health conditions. It's a case of extreme selection bias.

[Kazinski]

COVID is over but they don’t know it yet. Case rates hospitalization rates, death rates have all collapsed over the last 4 -6 weeks.

My surmise we’ve reached heard immunity by a combination of vaccinations(12.5%), case rates (probably at least 4x official numbers, 32%) and natural immunity (25-50% according to serology studies). If it were just because the holiday surge is over, the rate of growth would go down, but the daily case rate would still go up, but the daily case rate is down 73% (7 day average since Jan 11).

[AmosArch]

Probably overwhelmingly due to natural infection and immunity. The outcome was set in stone the moment China let those flights out of the country and phenomenon most responsible for ending it...herd immunity started months ago. Of course with President Harris in office the timing of the official recovery is adjusted a bit to conveniently coincide when she assumed power.

[aluchko]

You're adding recovered cases and "natural immunity" together like they're separate non-overlapping populations but they're actually the same thing, people having immunity due to a previous infection.

Either way, the collapse in rates has nothing to do with herd immunity. It's because the massive winter + holiday surge scared government and individuals into modifying their behaviour, and as a result infections declined.

If people start acting on the assumption that they've reached herd immunity it will simply cause another surge (especially with the more contagious variants spreading).

[Kazinski]

You got any evidence that people modified their behavior after the Holiday surge?

Because both Apple and Google show mobility activity is in a narrow range throughout Jan and Feb, and the drop is fairly consistent across all regions of the US, whether restaurant and bar friendly, or lockdown hell.

[Life of Brian]

You got any evidence that people modified their behavior after the Holiday surge?

You already know the answer to that. If it rains after the dance, it worked. If it didn't, you just need to dance harder.

[aluchko]

Mobility did drop in the US after the holiday surge, as did the new infection rate.

https://ourworldindata.org/covid-mobility-trends

And yes, rocks don't keep Tigers away, but spraying your house with water does make those big flames on your roof go away.

The connection between lockdowns and drops in community spread is both entirely expected and very well backed by evidence.

[DaivdBehar]

Many of the COVID deaths were presumed and fraudulent. Many were with COVID, not from COVID. For example, guy is shot in the head, or crashed a motorcycle, was heard to cough. The hospital got $13000 from Medicare if they put down COVID on the death certificate, $35000 if on a respirator.

A sample of all COVID deaths should be carefully examined to determine the extent of health care fraud. Then the fraudulent Democrat states should be forced to disgorge that fraction of fraudulent COVID claims from Medicare.

[Dr. Ed 2]

He is -- actually right.

Florida dealt with this a few months ago...

[Ben_]

Perhaps the answer is:

"It needs to be a year so it will make a compelling news story and get reported. Can you come up with an analysis where it somehow says a year?"

Some time later:

"It’s a year if you exclude ... and only count ... and extrapolate... Here. See? 0.96 years here at the bottom."

[Longtobefree]

"But I'm not sure just what this means."

Most likely, it means you can't trust numbers from the CDC.

[Tilted]

If I understand life expectancy estimates correctly, they’re based on current mortality rates. They’re saying something like (I’m oversimplifying, I’m sure):

If people from ages 0 to 4 die at a rate of X%, and people from ages 5 to 9 die at a rate of Y%, and… people from ages 95 to 99 die at a rate of Z%, then the life expectancy for someone 0 years old is A and the life expectancy for someone 62 years old is B. The X, Y, and Z rates are from whatever time period the life expectancy estimates are made for - e.g., for 2020.

So a substantial change in mortality rates in a given year - even if there isn’t good reason to believe that change will persist and even if there’s good reason to believe rates will revert to more or less what they were previously - can cause a meaningful change in the life expectancy as determined for that year.

IOW, IIUC, the CDC isn’t saying: The deaths which have been caused so far by COVID-19 and which are expected to be caused by COVID-19 will end up having a net effect of lowering the life expectancy in the U.S. by a certain amount.

Rather, the CDC is saying: If mortality rates stayed were they are now, which is in large part due to the effects of COVID-19, then life expectancy in the U.S. would be lower by a certain amount.

[Josh R]

Bingo! The 1 year drop assumes COVID deaths will continue at the rates we saw in 2020 into perpetuity. Thus, life expectancies will jump back up after the we beat COVID.

[Brett Bellmore]

Exactly. We lose a year of life expectancy if this pandemic never ends.

Not from a year of pandemic.

[bernard11]

Yeah. I think this is right.

They seem to be using death rates by age for the first half of 2020 to calculate the expectancy. So when the pandemic ends the lost year will probably come back.

It's more, "This is how the pandemic affected life expectancy in 2020," than, "This is how it affected life expectancy on an ongoing basis."

Of course the latter is wildly implausible unless you can show very serious long-term impact on covid survivors. There seems to be some such impact, but likely not nearly enough to move the numbers this way.

[Ben_]

So yet another completely fake news drama masquerading as science.

[Krayt]

Note it's the lousy reporting profiting from the scary headline, not the science itself.

[ragebot]

Putting aside the famous meme that 87.45890% of statistics are made up on the spot it needs to be noted that a lot of statistics really are meaningless.

One of my favorite examples of this was when the dems were pushing Obamacare it was endlessly pointed out that the infant mortality rate in the EU was much lower than the infant mortality rate in the US and this was ascribed to how wide spread socialized health care was in the EU. It turns out in the US if an infant is born and literally ten seconds later dies that counts in determining infant mortality. In the EU any infant that dies a week or less after birth is counted as a still birth and does not affect the infant mortality rate.

Another Obamacare argument was that life expectancy in the US was lower than in the EU; again ascribed to socialized health care. Turns out if you discount US military deaths in the Middle East and drug war turf war deaths in big cities the US life expectancy is on par with the EU.

Truth be told I could not make heads or tails of what was being said in the links. The first one claimed 2020 was the deadliest year in US history with over 3,000,000 deaths and a little under 500,000 due to COVID-19 (remember this was early on in the COVID-19 pandemic). So a lot more were due to something other than COVID-19. The second NIS link was to hispanic deaths only and looking at the graph it seemed the steep decline started in fairly early in 2019 well before COVID-19 was a gleam in the pols eyes. Not to mention it was only for hispanics and it was not clear to me was limited to US citizens.

Bottom line is even if peeps show their work in how they derived statistics it is often very easy to find issues with their work. In this case I see almost no rigor in how the underlying statistics were derived being cited.

One of the biggest issues I have seen being discussed is what is called 'excessive deaths'. Every year a certain number of people are expected to die (from stuff like old age, gang turf war murders, fatal social diseases, cancer, and a host of other causes). There have been a lot of folks who claim in 2020 there were not really a lot of excessive deaths, contrary to what the CDC claims. While there were a lot of deaths ascribed to COVID-19 the number of deaths due to ordinary flu dropped massively. There were also other causes of death that saw a significant decline.

I am not trying to minimize the serious issue I think COVID-19 is, rather trying to point out that it is still early on and a lot of statistics may not really mean much.

[gormadoc]

Excess deaths, not excessive deaths. The long history we have with the drug war gives us a good range of deaths to expect each year. Only deaths above that range are considered excess.

[Á àß äẞç ãþÇđ âÞ¢Đæ ǎB€Ðëf ảhf]

The drug war has nothing to do with it.

[gormadoc]

Obviously not, it was just a weird claim from that post.

[ragebot]

Only on my fifth cup of coffee when I wrote that. The point I was trying to make is there are a significant number of deaths in the US as a result of turf wars between drug dealers. If you look at Chicago as an example drive by shootings and collateral damage from them are the major cause of gun deaths there. Not trying to argue for or against the war on drugs just saying there are a lot of deaths in the US/big cities that don't happen in the EU due to fights between drug dealers.

[Á àß äẞç ãþÇđ âÞ¢Đæ ǎB€Ðëf ảhf]

There are similar hincky crime statistics. Britain, AIUI, does not tally a murder until they have a conviction; so a body with a dozen bullet wounds and no suspects does not count in the murder stats. I suspect there's something similar in violent crime stats. Last I remember, the least violent crime rate in the entire EU was double or triple the US, and the highest was Britain, with four times the US rate. Maybe that bullet riddled corpse simply adds to a different column.

[Davedave]

This is complete nonsense. It isn't nonsense to note that different countries gather different data, but the examples given here are absurd fictions.

[Á àß äẞç ãþÇđ âÞ¢Đæ ǎB€Ðëf ảhf]

I had also read about that difference in counting live births vs still born, and another accounting said the US tries to save earlier premature babies than most other countries, which also skews the statistics. These are all mysterious. It would be interesting to see a country-by-country description of the differences, with normalized statistics at the end.

[Life of Brian]

One of the biggest issues I have seen being discussed is what is called ‘excessive deaths’. Every year a certain number of people are expected to die (from stuff like old age, gang turf war murders, fatal social diseases, cancer, and a host of other causes).

Right. And for the past several years, expected deaths (the baseline off which "excess deaths" are measured) have trended up like clockwork -- about 100k or 3% per year. Growing population, largely constant life expectancy -- unsurprising.

But as if by magic, for the first time in many years, the CDC's baseline expected deaths for 2020 actually went down by about that same amount. That's a ~200k, or ~6%, swing in the baseline against which it's been reporting "excess deaths" for 2020.

Data available here. If anyone knows of any attempted explanation for this... interesting anomaly, I'd love to hear it. I've not seen one as of yet.

[ragebot]

I am not aware of any accepted explanation for this but have seen plenty of speculation that it was a deep state attempt to make things look worse than they really are. What is sad about this is with all the things talking heads have been blabbering about things like this seem to be being swept under the rug.

[Á àß äẞç ãþÇđ âÞ¢Đæ ǎB€Ðëf ảhf]

My simple-minded guess is that (a) the count of "deaths from covid" includes way too many "deaths with covid", and (b) too many "deaths from/with covid" are just quick guesses, and include a lot of, for example, flu deaths, because the symptoms are too easily confused.

But I haven't looked at the data itself.

[gormadoc]

I know that some countries saw a decrease in premature births but I don't know if it's the same for the US. Premature births depress life expectancy at birth so I wouldn't be surprised if the "correct" number (which we wouldn't necessarily have access to until much later as it takes a while to compile) is actually negative, since the other factors of COVID (COVID itself, lockdowns, isolation) are not affecting the cohort considered as much as any other cohort.

[TwelveInchPianist]

It would be nice if the folks at NBC and other news outlets who were writing these stories asked these questions, and included them in the reporting.

Instead what we get is, "It's already known that 2020 was the deadliest year in U.S. history, with deaths topping 3 million for the first time. Most were due to Covid-19, which has claimed the lives of more than 490,000 Americans since the start of the pandemic..."

[TwelveInchPianist]

Instead of spending bandwidth figuring out what their story was actually saying, they spent a bunch of time taking "health equity researcher" about what color the people who were dying were.

[OyVey!]

The article is misleading clickbait. The CDC does not say that life expectancy dropped by a year. Rather, the “period life expectancy” dropped by a year for the period January 2020 – June 2020. Period life expectancy does not describe the actual anticipated life expectancy of the population. Rather, period life expectancy is an estimate of how long the population will live if they were to experience the death rates observed in a given period. You could read how period life expectancy is computed at the link below.

Incidentally, while Covid-19 reduced the period life expectancy for 2020, it may increase the period life expectancy for 2021, and definitely increase it for 2022 and 2023. This is because a lot of the people who died in 2020 had a short life expectancy. Thus, a lot of the people would have otherwise died within the next few years died in 2020.

https://www.cdc.gov/nchs/data/vsrr/VSRR10-508.pdf

[Commenter_XY]

Thanks....your link explained it. Glad we fund the CDC and NIH.

The Results section, and then viewing Figure 1 in the document brought home your point. Net Net: Figure 1 will show the line moving up and to the right for the next couple of years.

I will be interested to read how this will be reported by the MSM in two years time.

[Brett Bellmore]

OTOH, a lot of people who were denied "elective" medical treatment last year, such as heart surgery or cancer surgery/chemo, will probably be dying this year thanks to that. So it may be a wash.

[Number 2]

Geez.

Thank God we have the mainstream media gatekeepers like NBC to protect us from fake and misleading news stories.

[Actuary77]

Life expectancy is calculated from the mortality rates at each age, or, more accurately, from the probability of survival at each age (1 - mortality rate). As of any age, it is the probability of surviving that year + the probability of surviving that year x the life expectancy for the following year. So the calculation works backward from the end of the table.

Example:
At end of table (say, age 120), last mortality rate = .7 => probability of survival = .3 so expectation of life = .3
Next youngest age (119), mortality rate = .4 => survival rate = .6,
expectation of life = .6 + .6 x .3 = .78 (survival rate in that year + chance to survive the year and then be able to have the following year's life expectancy.

Without knowing the details, it is impossible to fully understand what they did. Note also that there are several sets of mortality rates, which vary by race and gender.

[Life of Brian]

From the underlying CDC report:

"Provisional life expectancy estimates were calculated using abridged period life tables based on provisional death counts for the first half of 2020 from death records received and processed by NCHS as of October 26, 2020; provisional numbers of births for the same period based on birth records received and processed by NCHS as of October 27, 2020; and, April 1, 2020 monthly postcensal population estimates based on the 2010 decennial census."

So they basically just extrapolated the death rate in the first half of 2020 to infinity. Good enough for guv'ment work, right?

[NoVaNick]

Statistics will give you the answer you want if you torture the data enough, as they seem to have done here. Stats are very useful and have legitimate applications, but as for COVID and so many other things, when they are reported by the media, they often come from advocacy groups who have an interest in terrifying the public. I suspect the decrease in life expectancy will be followed by a call for more trillions to be spent on healthcare to raise it back up to where it was pre-COVID, which is going to happen anyway.

[Don Nico]

Bluntly speaking the NBC reported headline is at best misleading and at worst an outright lie. There really is no excuse for misinterpreting and/or misreporting scientific data so badly.

[Don Nico]

Have a look at National Vital Statistics Reports, Vol. 69, No. 12, November 17, 2020 that gives the numbers for 2018.
While the expectancy in 2018 was 1 year more at birth, by the time one get s to 70 the expected remaining life span has decreases by 0.3 years and if one makes it to 75, the difference between 2018 and 2020 is only 0.1 years.
Therefore the 1 year difference quoted by NBC is due to increased deaths is mostly due to a difference in death rate between 30 and 60. Given that the mortality due to COVID shows a rapid increase after 60 years of age, attributing these difference to COVID is unlikely to be correct.

[Mostly True]

Life expectancy at birth is the average age a person will live to if they experience the mortality rates at each age in a mortality table. The headline is saying that if a newborn experienced the mortality at each future age that people were experiencing at some point in 2020 at that age under COVID, their life expectancy would drop a year. Of course, that change in mortality cannot be permanent since, roughly speaking, even without a vaccine in few fears everyone will have had the disease, the remaining population will be immune, and mortality rates will go back down.

[mydisplayname]

This topic was discussed on Ezra Klein's Washington Post "Wonkblog" a number of years ago (and I'm too lazy to find the reference): the CDC's incarnation of vital statistics is a classic example of "GIGO" (garbage in, garbage out). It is absurd to reach any conclusion based on the CDC data other than (a) this is the approximate number of people who were born in settings where births were counted and reported to the CDC and (b) this is the approximate number of people who died in settings where deaths were counted and reported to the CDC.

In particular, only a smattering of death certificates are prepared by forensic medical examiners who have the slightest clue regarding the actual cause of death: all other certificates report the best guess of the preparer... and the preparer might be a Physician Assistant (with two whole years of medical training!) or a funeral director ("John Doe died peacefully in his sleep as a result of his battle with [respectable and feared disease XYZ]").

Applying formulae to the garbage -- stirring the pot of garbage -- is fun and profitable. Attempting to digest the garbage, though, is unwise.

[Stephen Lathrop]

I'm guessing there is some fairly reasonable explanation for the statistic. There is very little reason to pay attention to the statistic. It is a distraction from something anyone can by now understand—Covid-19 is an extremely dangerous disease—the most dangerous pandemic to strike the nation in a century—but Covid-19 falls short of being horrific. It is more important to focus on the implications of that than it is to focus on this statistic.

Reasonable imputations from the world history of mass disease suggest a likelihood that this nation will someday be challenged by a novel pandemic many times deadlier than Covid-19. That day could be as soon as next week. Covid-19 shows the nation is utterly unprepared for such a challenge.

Two considerations for policy:

1. What are the most efficient policies to end the Covid-19 pandemic and get the nation back to normal?

2. What does the Covid-19 pandemic teach about what the nation needs to do now, to be prepared to face down a truly horrific disease when it arrives later?

We know that Covid-19 policy has been a mess. Almost worse than anyone could make it by trying to screw it up on purpose. That is the premise the nation starts with.

Despite evidence, many policy makers will be extremely reluctant even to admit that, let alone take the next step and imagine an organized response to a pandemic that spreads faster, and threatens impartially to kill everyone.

The nation has just learned that life in politics dis-suits office holders for pandemic response. When a natural phenomenon as implacable as a mathematical equation appears, political leadership asks immediately, "How can we compromise?"

It does not ask, "What must we do to reduce the price we must pay." It asks instead, "How can we avoid paying full price." Disaster follows.

That is a flaw in our political system. But perhaps it will prove less dangerous if the nation undertakes to answer the questions before it falls into the grip of the next threat. There is a lot to be said for starting that effort while memories of this experience remain fresh.

[NoVaNick]

Stephen, there very well could be another pandemic that is deadlier than COVID, but, short of ordering everyone to literally not leave their home until a vaccine is available, there must be some sort of compromise to strike a balance between keeping people sane and allowing them access to their livelihoods, and protecting public health. The mask mandates attempted to do this, and while some idiots gleefully flouted them, most complied. The thing that really spread it was small gatherings at family events like the holidays. Unless the government shuts down all travel, there is no way to stop that.

[Stephen Lathrop]

NoVaNick, if I suggested your attribution to small gatherings at family events was conjecture, wouldn't we find folks on both sides of that? What is the point of your remark? Is it to discourage better-informed policy against more-threatening plagues? They will be whatever they are. Our choice will be to get ready or not. Which is better?

[Absaroka]

"if I suggested your attribution to small gatherings at family events was conjecture, ..."

... you'd be arguing with the CDC:

"But what we’re seeing as the increasing threat right now is actually acquisition of infection through small household gatherings.”

[Squirrelloid]

The organized response didn't do a damn thing. The only relevant actions taken were taken privately.

1. Citizens voluntarily limited their movement before governors ever issued shutdown orders. In fact, by the time any order was issued, the number of people 'staying home' had pretty much already topped out. (see https://fivethirtyeight.com/features/americans-didnt-wait-for-their-governors-to-tell-them-to-stay-home-because-of-covid-19/ )

2. Pharmaceutical companies diverted massive resources to research, testing, and vaccine development. The federal government's 'help' wasn't really relevant, and the government was often downright wrong-headed. They delayed testing by over a month. They delayed vaccine approval. (They literally sat on the submitted documents for approval for a week before even meeting about it). Government rules related to the vaccine rollout are overcomplicated and create poor incentives - increasing wastage and reducing how quickly the population is vaccinated.

If there's a lesson from Covid-19, it's the government should let non-partisan scientists give the public facts (and only facts), and let individuals determine what response makes most sense for them.

And all that's before even talking about how government's unnecessary and inept handling of the pandemic unreasonably politicized it.

[Doug Huffman]

Statistical literacy is uncommon and liars take advantage of that fact. Mathematical literacy and skepticism demands trust and verify. Media is impeached for verified dishonesty.

Falsifiability is the boundary demarcation of science from nonsense (non-science) and the narrative ad-hockery denies falsification; “But wait, more news tomorrow!”. The witch doctor’s tool. Faux FauXi

I was not a statistician, but neutron multiplication is ‘statistics’ and I was qualified to direct control of neutron multiplication.

The conspiracy of ignorance MASQUErades as common sense. Two chimp copywriters of a room full agreeing.

[Voize of Reazon]

I was qualified to direct control of neutron multiplication

EV should appoint you moderator of this blog.

[eyeroller]

Wikipedia says, "Period life expectancy at birth is the mean length of life of a hypothetical cohort assumed to be exposed, from birth through death, to the mortality rates observed at a given year."

[AtR]

The data and statistics reported by the press (and the CDC and WHO!) have been all over the map and are an excellent demonstration of how good people can be manipulated be even mild incentives.

There are significant political and financial interests at play and even the most experienced of scientists can get swept up by the expectations of their peers.

[James K. Polk]

I don't understand it either, though I suspect your guess about including the effects of lockdowns and economic downturn is correct.

If you do find the answer please post an update in a future VC email.

[B-Man]

As eyeroller and several other commenters have pointed out, life expectancy is a type of summary statistic that describes mortality rates observed in a given year (or other specified period of time). It is intentionally agnostic about the causes of death. It is also intentionally agnostic about what might happen in the future.

Absent some kind of "Ground Hog Day" scenario, life expectancy does NOT describe the expected length of life for any actual human being, because our actual lives are spread over multiple periods of time. This can lead to considerable confusion, as is evident in the comments above. But the rigorous definition of life expectancy is also a virtue, since it is easy to observe with precision and (unlike the attribution of mortality to particular causes) it resists attempts to "spin" the story.

That makes life expectancy a much more reliable metric for the benefits of public health interventions than, say, "lives saved." The latter simply can't be turned into a well behaved statistic that can be used outside a very narrow context. Banning coal, for example, would certainly save some lives. Would it save "net lives"? That's a difficult question to formulate, let alone answer. But it would be meaningful and useful (and observable) to know whether banning coal would raise or lower life expectancy.

So, don't be dismissive of life expectancy studies, but don't get misled by them either. The temporary effect associated with the onset of the pandemic is not something that can last, vaccine of no vaccine.

There is another temporary phenomenon that illustrates the limitations of life expectancy statistics: the alarm clock effect. Heart attack deaths tend to peak early in the morning -- especially on Monday mornings. So "period" life expectancy on Monday mornings is considerably lower than on the weekend. But during the pandemic fewer people are using alarm clocks so, FWIW, we can celebrate the reduction of the alarm clock effect!

[TutorFrank]

I have no particular knowledge here, and the argument suggested by other commenters that this is someone mechanically applying a formula that doesn't work in this context, rather that someone genuinely trying to figure out how long people will live, seems quite plausible.

Another guess I have is that this may reflect a change in our evidence, rather than a change in the world. So, for example, in any given year, there is a certain risk of a pandemic emerging. That risk may not have been changed much by covid. However, having seen covid, my estimate of that probability has gone up, and therefor the number of deaths of other pandemics that I expect in 2030 and 2040 and 2050 has gone up.

[ButWhatDoIKnow]

What about drug overdose deaths? Did opioid abuse just disappear?
Well..., not really. ...just from the media headlines.
"Over 81,000 drug overdose deaths occurred in the United States in the 12 months ending in May 2020, the highest number of overdose deaths ever recorded in a 12-month period, according to recent provisional data from the Centers for Disease Control and Prevention (CDC)."
https://www.cdc.gov/media/releases/2020/p1218-overdose-deaths-covid-19.html

[ButWhatDoIKnow]

It might be interesting to see how those drug abuse deaths break out by age groups, thereby affecting actuarial expectancy calculations. But I don't know how / where to find that.

[Rev. Arthur L. Kirkland]

"I'd love to hear what people who do know health statistics think about this."

Someone who (1) genuinely wanted to know and (2) is associated with an institution that employs plenty of relevantly knowledgeable people could probably get a reliable answer relatively quickly and easily.

Someone who wanted to rile some easily lathered rubes, however . . .

[Don Nico]

Spoken like a true rube and know-nothing.
In mathematics ability EV can run circles around you.

[John F. Carr]

As others have said, the assumption is that the pandemic never ends. Importantly, that includes two assumptions: (1) we never get immunity (by vaccination or otherwise), (2) reinfections are as deadly as first infections. We have reason to believe both are false.

This is a case of media freakout caused by extrapolating data past reasonable limits.

[Eric VonSalzen]

1. Thanks for some intelligent discussion of this issue by many commenters here.

2. I got my first covid vaccine shot during the Trump Administration (it was the morning of Innauguration Day). Biden got BOTH his shots during the Trump Administration (he lied that Trump didn't have any vaccine available).

3. On March 1 I turn 78. When am I scheduled to die? (Please don't bother with either birthday greetings or sympathy to my loved ones. Just facts. Be as specific as possible -- month day and year.)

[arch1]

Eugene, nothing in the CDC report or AFAICT in the article proper conflicts with anything you’re saying. Your beef is with whoever wrote the headline, which seriously misconstrues the message of the CDC report (the latter having made no attempt to quantify COVID’s mortality impact).

[arch1]

As for commenters’ complaints that non-COVID deaths are being misattributed to COVID, that the expected number of COVID deaths was mysteriously revised downwards, or that the CDC’s calculated life expectancies lack predictive value: These claims are all irrelevant to the findings, assumptions, and messaging of the cited CDC report, as you’ll see if you give it a quick look.