Exponential growth rate
May 29, 2020 4:38 PM   Subscribe

Global deaths due to various causes and COVID-19 - a sobering data visualization that puts the pandemic in greater context
posted by They sucked his brains out! (51 comments total) 39 users marked this as a favorite
 
I am still blinking, but I did shut my mouth. I think I just needed the accompanying figures to help me get my head around the danger we are in. States that are opening are seeing their infection numbers rise. Once we open, we will start the whole thing again, only there are already hundreds of thousands of infections to start from, rather than a relative few.
posted by Oyéah at 4:49 PM on May 29, 2020 [4 favorites]


Doesn't tuberculosis kill more than a million people every year? It seems odd that it's not included on a chart like this.
posted by Umami Dearest at 4:52 PM on May 29, 2020 [4 favorites]


Deaths by motor vehicle collision are also not included. And probably other things.
posted by aniola at 4:56 PM on May 29, 2020 [1 favorite]


Definitely leaves out a few things. But the way in which COVID-19 comes out of nowhere and leaps to the top is scary.
posted by TedW at 5:07 PM on May 29, 2020 [8 favorites]


Yeah, where are the shower deaths? I was always told one is in more danger taking a shower than flying in an airplane.
posted by Oyéah at 5:09 PM on May 29, 2020


My takeaway is that COVID is the Little Virus That Could.
posted by GenjiandProust at 5:09 PM on May 29, 2020 [11 favorites]


the way in which COVID-19 comes out of nowhere and leaps to the top is scary.

That seems misleading though, because Covid-19 hasn't leapt to the top of anything except for this cherry-picked list of causes of death that have lower numbers than Covid.

Tuberculosis is relevant because there are concerns that the TB eradication effort has been set back years by the overwhelming attention being paid to Covid-19, which is overshadowing efforts to combat other diseases like TB that actually kill far more people.
posted by Umami Dearest at 5:23 PM on May 29, 2020 [27 favorites]


This is fascinating and terrifying, but It’s hard to tell what exactly the baseline is - do all the counts start from 0 on Jan 1st?
posted by STFUDonnie at 5:25 PM on May 29, 2020


Wait a minute; none of the bars for different non-COVID causes of death change size relative to one another, which means that all this guy did was find out how many people have died from each cause of death so far this year, then interpolated them all from 0. For example, it strains belief to suppose that about 100 people have died due to terrorism every single day of 2020. So the imagery of COVID deaths leaping over other causes of death is based on a misleading presentation of how the death counts for other causes actually increase over the year.
posted by J.K. Seazer at 5:29 PM on May 29, 2020 [27 favorites]


So the imagery of COVID deaths leaping over other causes of death is based on a misleading presentation of how the death counts for other causes actually increase over the year.

I was wondering about that as well; it seems like influenza deaths in particular would have seasonal variations. In fact charting flu deaths during flu season might look exactly like this, with flu leaping to the top of a cherry-picked list of causes of death lower than the flu.
posted by Umami Dearest at 5:34 PM on May 29, 2020 [1 favorite]


Deaths by motor vehicle collision are also not included. And probably other things.

The U.S. averages between 35,000 and 38,000 crash deaths every year. COVID's lapped it twice already and we're not quite halfway through 2020.
posted by tzikeh at 6:07 PM on May 29, 2020 [13 favorites]


COVID's lapped it twice already and we're not quite halfway through 2020.

No, road accident deaths are also far higher than Covid-19 (at least for now), standing at around 1.3 million people per year.
posted by Umami Dearest at 6:13 PM on May 29, 2020 [3 favorites]


Genji, I remarked, several times when I was on the Covid service at my hospital, "What can't it do?!"
posted by cobaltnine at 6:14 PM on May 29, 2020


No, road accident deaths are also far higher than Covid-19 (at least for now), standing at around 1.3 million people per year.

No they aren't. I teach Driver's Ed. The U.S. averages between 35,000 and 38,000 crash deaths every year. The U.S. has over 100,000 COVID-19 deaths so far this year and it is not yet June.
posted by tzikeh at 6:27 PM on May 29, 2020 [14 favorites]


The 1.3 million number is the global number of road deaths.
posted by DSime at 6:39 PM on May 29, 2020 [5 favorites]


The 1.3 million number is the global number of road deaths.

Yes, I'm aware. My original comment specified that I was only talking about the U.S. I was then told that I was wrong.
posted by tzikeh at 6:40 PM on May 29, 2020 [18 favorites]


CDC: leading causes of death per year in the US.

The top one is heart disease at ~650,000 per year.
posted by bonehead at 6:42 PM on May 29, 2020 [1 favorite]


The U.S. has over 100,000 COVID-19 deaths so far this year and it is not yet June.

and an honest count would probably be north of 160,000.
posted by srboisvert at 6:51 PM on May 29, 2020 [8 favorites]


I wish this was the real top causes of deaths too. But it is a striking visualization.
posted by latkes at 7:21 PM on May 29, 2020


the way in which COVID-19 comes out of nowhere and leaps to the top is scary.

That seems misleading though, because Covid-19 hasn't leapt to the top of anything except for this cherry-picked list of causes of death that have lower numbers than Covid.


This, this, this.

Look at this chart of global causes of death. The visualization in the link leaves out all the biggest causes of death, which still outpace covid by a lot. If the visualization showed covid leaping up to a spot below cancer, heart disease, diabetes, dementia, and diarrhea, it might seem less alarming.

I feel like this data visualization sort of misrepresents the story: The really serious/terrifying story isn't the numbers we've seen (though those of course are still terrible), it's the much, much greater numbers we *would see* (or might still see) if things get worse, if the virus is not properly contained, etc.... That's maybe a more complicated point, but also seems more accurate.
posted by ManInSuit at 7:30 PM on May 29, 2020 [16 favorites]


The fact that this chart shows how much COVID deaths are outpacing influenza deaths is enough to make it worthwhile, and useful for rubbing it in the face of people who say "bbbut the flu!1!1!1"
posted by JohnFromGR at 7:35 PM on May 29, 2020


bbbut the flu!1!1!1

It is missing the point, to be charitable.
posted by They sucked his brains out! at 8:48 PM on May 29, 2020


Ummm - what ManInSuit said.
posted by ElGuapo at 11:18 PM on May 29, 2020


how to die, with statistics
posted by chavenet at 12:03 AM on May 30, 2020 [2 favorites]


I don’t think this chart actually helped put the deaths into greater context at all. I don’t know if I was supposed to be gobsmacked or whatever, but comparing it against poisonings and hepatitis really isn’t helpful for context when the big killers of today tend to be cancer and heart disease, you know? Actually what I’d like to see it compared against are the other viruses it keeps getting likened to, such as the Spanish Flu. Of course that killed millions which isn’t the point they want to make by this cherry picking of data, as Umami Dearest points out. (And yes, I realise the Spanish flu has come and gone so wouldn’t be able to appear on this timeline.)
posted by Jubey at 12:54 AM on May 30, 2020 [2 favorites]


Mod note: Checking out this post since it has just been flagged, and seeing the comments here, it does seem that this is maybe more an exercise in playing with the software to create a shiny thing rather than something of careful statistical / informational significance or value. However, the thread itself may be useful as a point of reference if the visualization has been shared widely? We'll be monitoring flags in case folks feel it's best to delete.
posted by taz (staff) at 1:12 AM on May 30, 2020 [4 favorites]


There’s a certain person on my neighborhood NextDoor who keeps posting a table of total US deaths along with percentage of deaths from COVID-19, broken down by age ranges. She thinks this table proves that the “news” (always in quotes!) has been lying to us, and that COVID-19 is not dangerous for people under 65. Because it’s a low percentage of the deaths for those age ranges.

And every time she does this, a handful of people try to nicely explain that her table doesn’t show what she thinks it shows, you can’t compute case fatality ratio from it, etc. I feel like it’s a microcosm for the willful innumeracy that’s behind many of the specious arguments we keep seeing for reopening. With online platforms like that, you can keep reposting the same stupid argument over and over, as long as you’re not breaking site rules. It’s exhausting.

The methodology complaints about this posts chart remind me of that. Smart MeFites will debunk it today, but that doesn’t matter because it’s going to keep getting reposted to Reddit with new audiences. Social media brings us endless echoes of anything shiny enough to reap some karma.
posted by qxntpqbbbqxl at 2:02 AM on May 30, 2020 [7 favorites]


how much COVID deaths are outpacing influenza deaths is enough to make it worthwhile, and useful for rubbing it in the face of people who say "bbbut the flu!1!1!1"

Unfortunately even that is a bit suspect. The WHO estimates that the flu kills 290,000 to 650,000 people per year, so I don't know where the chartmaker got his figure of 82,000 for halfway through the year - it doesn't really add up. (And the chart ManInSuit links to lists Lower Respiratory Infections as 2.38 million.) So while Covid deaths are certainly on track for being very bad, currently they're at a level of around 60% that of a bad flu season, globally speaking.
posted by Umami Dearest at 2:17 AM on May 30, 2020 [2 favorites]


Any assessment of the deadliness of COVID that ignores the fact that much of the world went into an unprecedented lockdown for months in an attempt to minimize transmission is suspect. Hopefully we never get proof of exactly how deadly it is.
posted by snofoam at 2:50 AM on May 30, 2020 [6 favorites]


I guess it would be interesting to see causes of death for three-year-old, third-year-olds and eight-year-old as the pandemic continues. But there are still so many unknowns in our figures, the eventual outcome, and the effect of the lockdown: I can't think that we can draw many conclusions yet.
posted by alasdair at 3:57 AM on May 30, 2020


The problem with this visualisation is that although it shows the rate of change, it does not give even a glimpse of what the future trajectory might look like, in the way that a simple line chart would.
I can imagine a lot of people looking at this and thinking: Oh well covid is only 50% worse than malaria deaths, I don't know anyone who has died of malaria, it's all a rounding error close to zero - lets forget about it and throw a pool party!

The FT have an interesting set of charts which show that Covid deaths in the US are bad in absolute terms, but measuring 'excess deaths' adjusted against similar historical time periods, (so removing differences due to different testing regimes) - shows the US is doing about average.
posted by Lanark at 5:49 AM on May 30, 2020


Any assessment of the deadliness of COVID that ignores the fact that much of the world went into an unprecedented lockdown for months in an attempt to minimize transmission is suspect.

It makes this view of the data even more remarkable, in that the rate is as high as it is, even with such action taken by different countries at timepoints along the way.
posted by They sucked his brains out! at 6:19 AM on May 30, 2020 [1 favorite]


Not impressed with this chart as good for context. The chart claims it only shows 7% of deaths, and doesn't mention how that 7% was selected, though "things less deadly than Covid" seems to be a likely candidate. That is not a full context.
posted by mrgoldenbrown at 7:01 AM on May 30, 2020 [2 favorites]


I think the non-COVID numbers are extrapolations based on monthly/annual data, not based on daily counts. I'd expect them to be more-or-less constant compared to COVID, but not that rock-steady.
posted by RobotVoodooPower at 7:20 AM on May 30, 2020


Please don't share this. It's exactly the sort of thing Edward Tufte was always warning against, for good reason.
posted by aspersioncast at 7:26 AM on May 30, 2020 [11 favorites]


I can't find it right this moment but I've always felt the most relevant chart was normalized for quality adjusted life years.

If a heart attack killed a 90 year old, it would make less of an impact on that chart because that person was at the end of their life anyway, so they might have lost a couple of year of life, and they were confined to a nursing home anyway before the heart attack.

Malaria killing a 2 year old though? Massive impact, because that person could have had an entire lifetime ahead of them.

A disease that cripples someone for life but doesn't actually kill them? Also massive impact on that chart when it affects a younger person.

It showed the top two things that are really killing us in developed countries are car accidents and mental health (self harm, suicide, depression).
posted by xdvesper at 12:56 AM on May 31, 2020


xdvesper, I understand the impetus to start comparing deaths like that, but I also find it really problematic. For one thing, there are so many different ways you can measure impact. A two-year-old's death is absolutely devastating for the child's parents, and will affect them for the rest of their lives. A 35-year-old who is statistically likely to have young children and a romantic partner, will probably leave an even greater number of bereft immediate family members, however. The impact on their children's lives will be huge and probably continue to have impact on them for the next 70-80 years. They are also likely to have dozens or even hundreds of friends who are impacted by their death. This will not be the case for the two year old.

Another problem with that kind of comparison is the risk of taking it a step further and starting to make judgements about 'quality' of life, as in fact you point to with your assumption that the quality of life of someone who is living in a nursing home means it is less of a loss than someone who isn't. It's only a short step to considering disability to make someone's life worth less.

Since we aren't actuaries and aren't required to figure out the calculus of what exactly a life is worth, I really think we shouldn't do so, but rather treat all deaths as equally tragic.
posted by lollusc at 1:19 AM on May 31, 2020 [4 favorites]


Interview with Betsey Stevenson:

GONZALEZ: Valuing people differently was kind of tried once in 2003. The EPA, under the George W. Bush administration, was considering a new clean air standard. How clean the air should be. And the EPA was like, well, people over age 70 have fewer years to breathe clean air. They should count less. And a lower value on some lives would essentially mean looser clean air standards.

MALONE: The EPA suggested that, perhaps, older people should be worth 37% less money.

STEVENSON: It became known as the senior discount.

GONZALEZ: The senior death discount, actually. People protested. They held up signs that said granny for sale. It did not go over well.

STEVENSON: The backlash was so intense they, very immediately, walked away from that, and you haven't seen any kind of similar proposal since.

posted by Comrade_robot at 1:53 AM on May 31, 2020 [2 favorites]


Fair points, I'm thinking from inside the medical system rather than looking in from the outside. I realise a lot of the visceral reaction against universal health care is rooted in this very idea, where a "death panel" will use QALYs to justify not treating grandma despite her paying into the system with her tax dollars. A lot of it stems from overestimating how magical medical interventions are, e.g. only 10% of doctors said they would want CPR done on them to save their life if their heart stopped. In the general population 75% of people said they would want the CPR. The "customer" isn't always right.

These whole "cause of death" charts are misleading because 100% of people die, of something or another. The goal of healthcare is to see through that noise and see where we can make the most difference to people's lives, the QALYs are one way to do that. With limited resources, we are charged with making the best use of what we have. And making sure the cure isn't worse than the disease.

It's staggering the amount of effort and disruption we went through to combat Covid and I only wish we put a similar priority to mental health. I actually find it hard to imagine why we tolerate so many motor vehicle related deaths: imagine a hypothetical future where all cars on the road were required to be at current safety standards (crumple zones, airbags, seatbelts) and then installed with mandatory speed limiters to exactly 60kmph / 40mph. You'll cut deaths to virtually zero and significantly reduce emissions. The cost would be just taking a few minutes longer to get where you want to go. We value saving a few minutes more than we value not dying, I guess.
posted by xdvesper at 7:57 AM on May 31, 2020 [3 favorites]


The cost would be just taking a few minutes longer to get where you want to go. 

That shouldn't be trivialized. A life is only 41 million minutes.
posted by justsomebodythatyouusedtoknow at 9:59 AM on May 31, 2020


xdvesper I love that idea, but I think it will actually need to be 20mph. Most people will still die if they're hit at 40 mph.
posted by aniola at 1:13 PM on May 31, 2020


That shouldn't be trivialized. A life is only 41 million minutes.

I think what will happen is we will spend the same time commuting, we would just choose closer destinations. It's similar to induced demand - why building larger and larger highways don't solve congestion. Travel time falls, people are "induced" into wanting to travel because it's convenient, then you're back at square one as traffic increases.

xdvesper I love that idea, but I think it will actually need to be 20mph. Most people will still die if they're hit at 40 mph.

Ha, and here I was talking about the vehicle occupants being 100% safe!

But yes fair point about pedestrians. Europe pretty much mandates AEB (autonomous emergency braking) so that cars detect humans, cyclists and animals using radar and cameras and will automatically stop to prevent the collision. Crucially, it's limited to operation at slower speeds (40mph and below). Nothing is 100% certain but it would definitely save a lot of pedestrians from drivers who missed seeing them. The current safety standards are starting to force design changes such as pop up hoods and energy absorbing bumper beams / geometry that scoop pedestrians up to the hood which pops up to prevent them hitting the hard components in the engine bay. In contrast the US has zero pedestrian protection standards (afaik)...
posted by xdvesper at 6:58 PM on May 31, 2020


Look at this chart of global causes of death. The visualization in the link leaves out all the biggest causes of death, which still outpace covid by a lot. If the visualization showed covid leaping up to a spot below cancer, heart disease, diabetes, dementia, and diarrhea, it might seem less alarming.

Ok, yes, but covid just killed almost 400k in the last four months. If it behaves and just does that (and good luck with that) for the next 8, that’s 1.2M/yr or up in ‘car accident’ territory, def in the top 10. Out of nowhere. That’s alarming enough.
And (like I have had to reiterate til I’m blue in the face) the covid numbers are where they are because everybody’s been locked up at home for 3 months and standing 6ft apart and denuding the earth of alcohol, hand sanitizer, and n95 masks.
Hmm...has anybody tried to figure out where we’d be now if we hadn’t done all that. Would heart disease be in the rearview or just cancer?
posted by sexyrobot at 8:10 PM on May 31, 2020 [1 favorite]


(Or what snofoam said)
posted by sexyrobot at 8:33 PM on May 31, 2020


All that having been said, can we just take a second to consider how many people apparently die from drowning?

I find that figure extremely hard to believe. If drowning really killed at double the rate of fires, wouldn't we expect a lot more resources to be put into lifeguards?
posted by Buck Alec at 8:02 AM on June 1, 2020 [1 favorite]


In Canada it is an order of magnitude more dangerous to your children to have a backyard pool than to have a firearm in your house.

The World except for the US Health Organization has a good summary on drowning deaths:
The global burden and death from drowning is found in all economies and regions, however:
  • low- and middle-income countries account for over 90% of unintentional drowning deaths;
  • over half of the world's drowning occurs in the WHO Western Pacific Region and WHO South-East Asia Region;
  • drowning death rates are highest in the WHO African Region, and are 15-20 times higher than those seen in Germany or the United Kingdom, respectively.
So a US/Canadian/European assessment of the risks doesn't extrapolate well to the whole world. Flood deaths are significant, and daily commuting or journeys made by migrants or asylum seekers over water in high drowning rate areas are often on overcrowded equipment with minimal/no safety equipment and are significant contributors to global deaths by drowning.
posted by Mitheral at 8:41 AM on June 1, 2020 [2 favorites]


imagine a hypothetical future where all cars on the road were required to be at current safety standards (crumple zones, airbags, seatbelts) and then installed with mandatory speed limiters to exactly 60kmph / 40mph. You'll cut deaths to virtually zero

We would not get to "virtually zero" car-crash fatalities by having speed limitations built into cars by any stretch of the imagination. While speed is the top killer on the road (you'll often see drunk driving listed as #1 but the majority of deaths in drunk-driving crashes happen because the drunk driver was going too fast), it's far from the only killer. A significant number of motorists die each year simply because they don't follow the right of way rules, for example, or they tailgate, or they drive too close to the left side of their lane.

Hell, falling asleep at the wheel accounts for ~1,000 deaths a year on average, and texting while driving is racing up the cause-of-fatality list with similar acceleration to COVID racing up this graphic. And I can rant forever about seat belts. It doesn't matter if we have seat belts if people won't wear them. The number of annual crash fatalities simply because the person wasn't wearing a seat belt makes my brain spin. Wearing a seatblet cuts your chances of dying in a fatality-possible crash in HALF, people! 50%! WEAR YOUR FUCKING SEAT BELT ARRGHTHGHTH

</Driver's Ed fatality statistics lesson tangent>
posted by tzikeh at 12:53 PM on June 1, 2020 [1 favorite]


Also on QALYs: done wrong, they're incredibly regressive. Poor people die younger, so a naive implementation says it's not worth spending as much on them.

It would, of course, be extremely cynical of anyone to suggest that such a thing might be viewed, by some, as a desirable feature of such a system.
posted by regularfry at 2:33 PM on June 1, 2020


All that having been said, can we just take a second to consider how many people apparently die from drowning?

I was just saying this last night in chat...like wtf. But yeah, remember 90% of accidents happen in the home, soo...bathtub? And all the things mitheral says, a big chunk of which sounds like “monsoon season”
posted by sexyrobot at 5:39 PM on June 1, 2020


We would not get to "virtually zero" car-crash fatalities by having speed limitations built into cars by any stretch of the imagination. While speed is the top killer on the road (you'll often see drunk driving listed as #1 but the majority of deaths in drunk-driving crashes happen because the drunk driver was going too fast), it's far from the only killer.
It depends on what the limit is. If all cars were limited to 1 mph, I'm pretty sure we would see zero fatalities from crashes. So somewhere between that and the current speed limits is the threshold.
posted by floomp at 7:02 PM on June 1, 2020


It depends on what the limit is. If all cars were limited to 1 mph, I'm pretty sure we would see zero fatalities from crashes.

Well, sure, but only because no one would use cars anymore. I think it's fair to say that as long as people use cars, they will usually want to travel faster than 20 mph, but even around this speed, upwards of 2% of traffic collisions with pedestrians in the US are fatal. Assuming that a maximum speed is the only policy change made, there does not appear to be any threshold at which people would drive cars with zero risk of killing pedestrians.

So, as much as I enjoy applying the Intermediate Value Theorem to the real world, I don't think it works in this case :)
posted by J.K. Seazer at 7:32 PM on June 1, 2020


« Older 40 years of feeding the hungry   |   How long I have waited… Newer »


This thread has been archived and is closed to new comments