Comparing Covid-19 Deaths to Flu Deaths Is Like Comparing Apples to Oranges | Hacker News Reader

Comparing Covid-19 Deaths to Flu Deaths Is Like Comparing Apples to Oranges

blogs.scientificamerican.com

94 pointspmorici173 comments

Flu or not, it's possible to compare all-cause mortality this year against previous years. That's about as hard data as it gets. How does Covid compare to "seasonal all-cause-winter-deaths"? That's what people are interested in - how unprecedented is the death rate?

It's possible, and it's being done.

https://www.nytimes.com/interactive/2020/04/27/upshot/corona...

> The recent death count reached six times the normal number of deaths for the city at this time of year, a surge in deaths much larger than what could be attributed to normal seasonal variations from influenza, heart disease or other more common causes.

https://www.nytimes.com/interactive/2020/04/21/world/coronav...

> These numbers undermine the notion that many people who have died from the virus may soon have died anyway. In Paris, more than twice the usual number of people have died each day, far more than the peak of a bad flu season. In New York City, the number has spiked to six times the normal amount.

I'm betting that in the next few years the death rate will drastically fall because the only people left will be really healthy.

If it is killing people a couple of years earlier than they otherwise would have died, it is fair to consider them excess deaths.

Are you counting more than one death per person?

> I'm betting that in the next few years the death rate will drastically fall because the only people left will be really healthy.

If everybody died this year, then next year's death rate will be zero! That's not a good thing.

Except that people who survive severe illness with Covid-19 often seem to suffer severe organ damage as a result. Mortality among Covid survivors is therefore likely to be higher than normal in the coming years, so it’s unclear how the numbers will pan out.

I will take that bet! I'd say that "drastically" means 20% or more. How much would you like to wager?

If you are old or infirm when you contract a bad case of corona, you die.

If you are relatively young and completely healthy, you spend a month in the hospital; get onto a ventilator for weeks requiring heavy sedation; lose your ability to walk, talk, feed yourself and brush your teeth; have to relearn how to swallow; maybe suffer a collapsed lung and upon release if you're lucky are confined to a wheelchair and physical therapy unable to take care of yourself, your loved ones, and return to your job.

Death is not the only negative outcome of this disease. "Recovery" from corona is sometimes the beginning of a long and slow struggle. I'm looking forward to any analysis on the economic cost of this.

If you are relatively young and completely healthy ... you overwhelmingly recover completely fine.

The number of young people that are healthy (no pre-existing conditions like being overweight, diabetes or high blood pressure) that that experience what you describe is extremely rare.

> no pre-existing conditions like being overweight, diabetes or high blood pressure

Well over half the country ticks one of those boxes.

70% of Americans are either obese or overweight.

The coronavirus is going to with us for a long time; we're unlikely to get a miracle cure or effective vaccine anytime soon. On average American adults are about 27lbs (12kg) above their healthy weight. That means if they start losing weight now at a safe, sustainable rate then most people can be down to a healthy weight in a year or so.

Obviously there is a huge amount of individual variation but most people can take steps to improve their individual odds of survival.

> These numbers undermine the notion that many people who have died from the virus may soon have died anyway.

This is known as mortality displacement or the harvesting effect. We won't actually know whether this was a significant factor until we see future all-cause mortality statistics.

I think by "soon" I think he meant a few days but I don't know that anyone is making that claim.

One question I have is how many of these excess deaths is due to coronavirus and how many are due to our response to coronavirus since people are not getting their regular health care. For example, my dad has congestive heart failure and sees his doctor once a month and gets an EKG every time. I've been in appointments and the majority of the discussion centers around changes in his EKGs from the previous months. Right now, he's not getting these regular checkups and appointments. He could have a change in prognosis go undetected and end up with a preventable heart attack.

Is coronavirus the black swan event or is our reaction to coronavirus the black swan event. It's hard to say.

Looking at all cause mortality excess deaths between Sweden and lockdown countries, it's hard to say:

https://www.nytimes.com/interactive/2020/04/21/world/coronav...

One question I'm also curious about is if in the years following the coronavirus, will we see negative excess deaths relative to previous seasonal averages because of the excess deaths right now.

Yeah, its quite amazing how poor our data is. Combine that with how politicized the issue is and I don't think we will ever have reliable numbers.

> One question I have is how many of these excess deaths is due to coronavirus and how many are due to our response to coronavirus since people are not getting their regular health care.

I wouldn't personally know where to start on getting the data, but you can likely tease this out by evaluating populations with already limited access to healthcare - uninsured, low-income folks with $13k deductibles, etc.

Have you considered getting your dad an Apple Watch, so he can do his EKGs at home and send to a doc? Not as effective as a twelve-lead, but maybe it'd help at least somewhat?

I've been looking at the KardiaMobile but the reviews on Amazon look pretty bad. How well does the Apple Watch perform since it's just on a wrist?

I honestly don't have the information to answer that.

I do know the single-lead EKG has some limitations, but they may not be relevant to your dad's particular scenario. Worth a conversation with the doctor, I suspect.

Yeah the real tricky thing is how well they do. We originally thought he had bradycardia because his heart rate as measured by his heart pressure cuff showed it in the high 30s. However once he got an EKG, it became clear that his heart rate was normal but he was afib and all the irregularity wasn't being picked up properly by the BP cuff monitor. I'm guessing that something like the apple watch is accurate enough to inform someone previously healthy that something is now wrong, but wouldn't be accurate enough to tell someone who is unhealthy if their condition is worsening.

The Apple Watch's EKG is definitely good enough for diagnosing bradycardia - you can see very clear QRS complexes showing up on it.

(I would not necessarily trust the 24/7 LED-based rate monitoring.)

Yeah, but brachycardia wasn't the problem and that's easy to diagnose. What I want to know is if it is good enough to know that it wasn't actually brachycardia but was instead afibrillation with weak parts of the cardiac cycle.

In the UK about half the excess deaths don't mention COVID at all on the death certificate.

Of those that do, a very high proportion are extremely old, i.e. they died with the disease rather than of it.

If you take that into account the lockdown kills far more healthy people than COVID.

Germany is in lockdown and seeing no excess deaths (covid or otherwise). The whole of Italy is in lockdown, but areas badly affected by covid are seeing much more more unexplained excess death than other less affected areas.

It seems to me that lock-down is very unlikely to be the thing that is killing people.

Germans haven't been told repeatedly they have a patriotic duty to avoid hospitals to save their national treasure.

The British numbers are public and speak for themselves. ER attendance halved. Cardiac patients halved. People stopped going to hospitals when they needed to - luckily Germany avoided socialised healthcare and hospitals are just seen as normal companies.

As for Italy, there have been similar problems there with elderly homes in particular and hospitals struggling with staffing sure to lockdowns and isolation.

Ultimately this is the same virus everywhere. It doesn't look like it though. So differences in national performances are telling us a lot of important things about how governments have reacted.

Where are you getting your UK stats from please?

People are huddling inside in fear of Covid-19. How many of these excess deaths are caused by fright and the stress thereof?

"Can You Die from Fear? The Baskerville Effect"

https://www.drmirkin.com/histories-and-mysteries/the-baskerv...

And there's an opposite effect. Way fewer people are dying or being hurt by car crashes. Normally one of the leading causes of death in the "prime of life" age range. And even more weird, deaths and injuries for young people are way down as impulsive risk taking and peacocking are reigned in. So weird all the indirect effects.

You've selected anecdotes from two very specific places. But that cuts both ways. There are plenty of places where so far they have been equally or less affected than a flu season.

For instance the UK is doing worse than most other countries in Europe but currently there are quite a few years within my life with similar death numbers by this point in the year: http://inproportion2.talkigy.com/

Or look at the graphs for Europe: https://www.euromomo.eu/graphs-and-maps/

There are some countries where it's higher than the flu seasons in the past 4 years. But there are other countries where it's equal or lower.

Overall weekly mortality in the EuroMOMO reporting countries peaked at approx 2x that of 2016, approx 20k deaths per week more. Given the large populations of these countries this is a tiny change. To put that in perspective 20k people is 0.03% of the British population alone, but you'd also need to include all the other countries to get the real percentage.

>Mortality statistics show 122,000 deaths in excess of normal levels across 14 countries analysed by the FT

https://www.ft.com/content/6bd88b7d-3386-4543-b2e9-0d5c6fac8...

If you count excess deaths, globally where we have data, official Coronavirus deaths are always about half or 2/3rds the excess death.

In other words, it’s likely we are undercounting deaths caused by COVID by a factor of 1.5 to 2 in places which are actually properly tracking COVID deaths.

You also have this effect where this year diagnosed cases of heart attacks and appendicitis are lower than last year. Can't blame people for not showing up at the hospital because that's where the plague is. This kind of excess death where people with serious conditions avoid hospital because of fear of disease isn't coronavirus but clearly caused by coronavirus.

is it caused by the coronavirus or the excessive measures taken to handle the coronavirus?

shutting down half the hospital when there is no need to for example

The shutdown wouldn't happen if the coronavirus wasn't spreading. It's the indirect cause, remove it and the hospitals aren't shut down.

you are confusing the response with the disease - mandating that people cannot go to go to the beach is not caused by the virus, it is caused by a bureaucrat. There is no evidence whatsoever that going to a beach causes infections. If anything makes your body stronger and more able to cope with a disease.

Exposure to the virus is how you get the virus. If there are other people at the beach, some of them might be infected, and you're being exposed by being in proximity. Maybe if you're Far Enough Away you can limit the exposure, but are people really Social Distancing on beaches and in restaurants? Evidence so far suggests the answer is no.

Incidentally there is also a lot of data out there suggesting that the virus can travel through the air a pretty reasonable distance, and it's likely to be breezy on the beach.

In any case, why would a bureaucrat want to go out of their way to close hospitals if there wasn't a good reason? Who benefits from that? Literally no one. It's only a reasonable idea because of the outbreak. The only people you see closing hospitals on purpose are vulture capitalists.

> In any case, why would a bureaucrat want to go out of their way to close hospitals if there wasn't a good reason? Who benefits from that?

Politicians benefit from that. The current situation is no different than "No one ever got fired for hiring IBM". In this case "hiring IBM" is a surrogate for doing whatever your voting base believes in. In blue states, the majority believes in shutting things down aggressively, so politicians do. In red states, the majority thinks we should keep things open, so politicians do.

Politicians by and large in this crisis aren't responding to what experts say, they are responding to their own self interest. This is super easy to spot by seeing what experts are saying or what evidence is raised that they summarily dismiss because it contradicts what their constituency desires.

What evidence would that be? Do you have any reliable data on the number of people infected at beaches? So far the results of real world contact tracing aren't consistent with computerized models of virion distribution in open air.

you just exhibited the fallacy of equating a restaurant to a beach (or being outside). You are also using "exposure" in the most generic sense of "presence" within a distance. That's really not how it works.

Like I said there is absolutely no evidence that you could get infected while in an open environment, moreover plenty of evidence that being outside is greatly beneficial to health and well being.

Now I am asking you the same question you asked me: "why would a bureaucrat want to go out of their way to close a beach if there wasn't a good reason?"

The answer is simple as unpleasant, they don't understand what they are doing.

Why are we talking about closed beaches? We were talking about people with symptoms of appendicitis who choose to stay away from the hospital that is very much open for such cases.

Closing the beaches is frankly idiotic, but postponing elective surgery isn't. It's not wise to go in for a breast enlargement when an epidemic is on the move. With an appendicitis, on the other hand, you don't have a choice.

If a surgery isn't needed to prevent serious permanent injury in the next few days, it's considered elective. Tumor removals, for example, are elective surgery and have generally not been allowed.

Maybe, but the article is doing the exact same thing with the flu, except probably even worse because we're not systematically testing people for the flu in the same way that we are Covid-19.

You don't need to systematically test people in order to assign their death as a flu death.

We have many different ways of counting deaths. We either count the people who had a test; or we count death certificates; or we do surveillance to see what respiratory illnesses are circulating and how many people are searching google for flu-like symptoms and how many people are attending primary care and ED for flu-like symptoms.

(and which ever method we pick, if we pick the same method for covid-19 we see many more deaths)

The whole point of the article is to argue that those flu deaths are "wildly inflated statistical estimates" and not real flu deaths in the same way that Covid-19 deaths - where we are systematically testing increasingly wide-ranging groups of people who are hospitalized or even just in care homes - are real Covid-19 deaths. That the only apples-to-apples comparison is not to statistically compensate for the lack of flu testing at all and only compare tested, confirmed cases.

It's obvious just from comparing the numbers in the article to the ones in the comment I'm replying to that lack of testing has a much, much bigger effect on confirmed flu numbers than confirmed Covid-19 ones - confirmed Covid-19 deaths are at 2/3 to 1/2 of excess deaths, whereas confirmed flu is at 1/4 to 1/8th of estimated flu deaths, and the flu estimate does more work to exclude unrelated deaths than just counting raw excess deaths would. The supposedly apples-to-apples comparisons in the article looks to be even less directly comparable than the apples-to-oranges one it's rejecting.

But note that this is not true at a regional level - many US states, for example, have official coronavirus deaths higher than the excess death.

in the most affected countries in Europe, it is significantly higher (2, 3 times) than the worst flu seasons in recent times.

That being said, we have developed a combination of vaccines and ICU treatments to drastically reduce the mortality of the flu. This is not your '70 flu anymore.

Here's the charts for a number of states in the US. It's pretty dramatic. https://www.nytimes.com/interactive/2020/04/28/us/coronaviru...

Dear god those charts are useless. Different scales all lined up next to each other, no clue as to the value of the bottom of the y axis, no clue as to what the baseline or the peak numbers are, no ability to actually inspect the data... Who are they hiring to do this nonsense at NYT?

Going to be charitable and hope the bottom of y axis is zero.

They're not intended for comparing chart to chart, there's a table with key numbers below, and links at the bottom to country-by-country, state-by-state, and "look up your city" options.

It's a "look at the spikes" illustration in the spot in an article a big photo would typically go.

It's pretty dramatic because they excluded all of the states where it is not dramatic.

Differs on a country by country basis, but generally it's pretty bad in Western Europe anyway. See https://euromomo.eu/graphs-and-maps/ for excess mortality data.

By the way, most countries' local COVID peak is either in April or still to come; it's not really winter deaths.

One caveat; in many countries, death reporting is pretty slow and data from the last couple of weeks shouldn't be considered particularly reliable.

Here's the 2 numbers that matter:

- Covid w/o a shutdown

- Overall deaths from lockdown w/o Covid

Neither number can ever do more than guess at and fight at which statistical model lies less.

One problem with comparing excess deaths is that automotive traffic deaths are down, and automotive traffic deaths are a very large slice of the pie in many countries (including the USA).

There were about 2.8 million total deaths in the United States in 2019 compared to about 40 thousand motor vehicle deaths.

Going further, in 2017 all accidents combined were 6% of deaths in the US (https://www.cdc.gov/nchs/fastats/deaths.htm). Nearly half (1,246,565) of the 2,813,503 deaths in the US in 2017 were from heart disease and cancer. Deaths from heart disease per day are almost as bad as covid-19 right now, so since there is a lot one can do proactively to reduce the chances of heart disease, I'm taking it way more seriously than I did before: don't smoke, hour+ of exercise every day, heart-healthy diet, sleep enough, and less stress.

This was posted here recently: https://www.euromomo.eu/graphs-and-maps

It only covers Europe but effects of Covid-19 are easy to see. For example during the peak, about 85,000 Europeans were dying per week as opposed to the normal range of 50-55K.

I don't know of a source for non-European countries unfortunately.

Until there is an international cooperation to standardize how Covid19 deaths are counted, these type of numbers are all we have for international and historic comparison.

“That's what people are interested in”. I very much disagree. 9/11 didn’t exactly register in all cause mortality. There’s a lot more people are interested in, for good reasons.

The question I have yet to see posed is "How would society be experiencing the common flu if it had never existed before and this is the first time it jumps to humans?"

The first year you see an airborne transmissible disease is likely to be the worst regardless of the disease. The big problem I see is that we're comparing the common flu in its herd immunity steady state to the coronavirus in its first year.

Once coronavirus settles into its steady state, we will be able to compare it to the flu, apples to apples.

That being said, assuming the flu and coronavirus at steady state is approximately comparable, would it justified to react to the flu in its first year the same way we're reacting to the coronavirus in its first year? If we know there will be a modest spike in the first year that mostly claims the infirm and becoming endemic is unavoidable, is it worth doing the damage we're doing to the economy.

That this is going to become endemic and likely reach steady state in a year or two (possibly longer depending on how long we retard reaching that point with lockdowns varying from geography to geography) is at this point a foregone conclusion because not every country has the natural borders of New Zealand and can completely kill this off. It will bounce back and forth between the northern and southern hemisphere indefinitely at this point until herd immunity.

A case of deadly flu with little population resistance would like the Spanish flu pandemic. Yes, I would say that the an extraordinary reaction would be entirely appropriate.

At this point, the disease becoming endemic is all but guaranteed, so we're just delaying the inevitable at great economic cost. The best case scenario is to ride the healthcare system capacity threshold while getting the economy moving again so there aren't excess deaths beyond those that would have been claimed sooner or later.

The people arguing for waiting for a vaccine simply don't understand how long that likely to take. It's an unreasonable expectation to keep the world shut down until a vaccine is available as it will cause far more damage than the virus is liable to cause.

I have a friend who works in a stroke unit and he says almost nobody is coming in for strokes. Instead some people are just suffering or dying at home. Of strokes. Not every single one of them, but some alarming and as yet unknown number.

All that sounds unrelated to Covid-19. But the reason they are not going to the hospital is because they fear going into a medical facility, for fear of Corona virus. Should these deaths be counted as Covid-19 deaths? As shelter in place deaths? Or just as strokes? I'm not sure. It's complicated to sort this stuff out.

And death from other infectious disease will be down as well because of the lockdown. But the demography of the victims from covid is largely skewed against people with a low life expectancy in the first place. And then a severe economic crisis will come with its own health issues.

So the net impact of covid will only really be known at the end of the year at best.

Two important takeaways here.

"In fact, in the fine print, the CDC’s flu numbers also include pneumonia deaths."

"If we compare, for instance, the number of people who died in the United States from COVID-19 in the second full week of April to the number of people who died from influenza during the worst week of the past seven flu seasons (as reported to the CDC), we find that the novel coronavirus killed between 9.5 and 44 times more people than seasonal flu."

For the average person, COVID-19 is better described as "highly contagious and extra deadly pneumonia".

Pneumonia is a common fatal complication of influenza, and I think there's reasonably good statistical evidence that the cases counted towards flu deaths are mostly caused by the flu especially in the bad flu epidemic years which are most comparable to Covid-19. That sentence is a bit like claiming that people don't die from falling out of airplanes because the actual cause of death was hitting the ground.

I think you meant 'flu deaths are mostly caused by [bacterial] pneumonia'. One thought along those lines is that the spike in deaths we see is due to covid19 killing directly, death coming like clockwork X weeks after infection. Compared to the flu, which due to the requirement for a secondary bacterial pneumonia, has a higher variability in the timing of the death: at what time bacterial infection develops, which strain of bacteria, antibiotic interference, etc. Thus, we're going to see higher spikes for covid, and hopefully narrower.

Yeah, and apparently one of the more interesting consequences of this is that people don't necessarily have detectable flu virus anymore once they're hospitalized with pneumonia because it's the bacteria allowed in by the now-gone flu infection that's killing them.

As interesting as this article is... you absolutely can compare Covid-19 deaths to Flu Deaths. It's a far better comparison than, say, the people who note that we are having a 9/11 Terrorist Attack every day. I think the fear is that people will look at the death toll alone -- but if you include timespan as a point of comparison, it becomes very helpful to developing an understanding.

But as long as we're suddenly concerned about understanding the statistics about deaths... the 40,000 people that die annually to "gun violence", over half are suicides.

"There are three kinds of lies: lies, damned lies, and statistics." - made popular by Mark Twain.

Which as you point out, context matters, detail matters. You can make any point you want given enough caveats.

On a somewhat related note - the absolute death toll number isn't super useful either. I was shocked to learn a couple days ago that it seems pretty consistent around the world that 50% of that toll is from people in nursing homes. I did a quick search and found a study that found median/mean stay in a nursing home until death is 5/13 months. I'm not even examining the numbers closely, but that would suggest a huge number of people dying that are in absolutely precarious health. A death toll number does not convey that at all.

But let's compare them anyway.

"we find that the novel coronavirus killed between 9.5 and 44 times more people than seasonal flu."

That was easy wasn't it.

The point was comparing reported COVID deaths to statistically extrapolated flu deaths is apples to oranges. What you quoted was a comparison of reported deaths to reported deaths.

"Comparing reported COVID deaths to statistically extrapolated flu deaths is apples to oranges"

I suppose this would be a better title of the article.

I think comparing just reported deaths is probably even worse, the rate of undercounting of flu deaths and undercounting of covid deaths (in a pandemic with a ton of testing / heightened observation) is not going to be comparable.

Definitely. My takeaway is we should be skeptical of any comparison between COVID and flu statistics.

>That was easy wasn't it.

Was it?

I don’t get it. I found that there so far have been covid 234,133 deaths [0] and season flu can cause up to 646,000 deaths [1].

So I’m not seeing 9x to 44x by any reasonable metric or projection. What am I missing?

If 9 to 44x is possible, there need to be models that would show that. Does anyone have a model that shows we aren’t even remotely near peak deaths?

Even if the article is right and seasonal flu was vastly over reported by the CDC as a “for your own good” tactic, I’m STILL not seeing 9x to 44x.

[0] https://www.statista.com/statistics/1087466/covid19-cases-re...

[1] https://www.medicinenet.com/script/main/art.asp?articlekey=2...

EDIT: Article uses worst week covid compared to worst week flu... talk about Apples to Oranges!?

> What am I missing?

It's right there in your comment:

> so far

(Also, the rest of the paragraph that 9-44x number comes from:)

> To do this, we have to compare counted deaths to counted deaths, not counted deaths to wildly inflated statistical estimates. If we compare, for instance, the number of people who died in the United States from COVID-19 in the second full week of April to the number of people who died from influenza during the worst week of the past seven flu seasons (as reported to the CDC), we find that the novel coronavirus killed between 9.5 and 44 times more people than seasonal flu.

No, I don’t think I am missing that. Which is why I used those words.

I have no doubt covid is worse than flu and have no issue with being more cautious than reckless.

But if you can find me a model that shows we’re no where near peak deaths and are still looking to 9 to 44 times more, I’d like to see that.

Edit: anyone? Sources? Models?

The finding that in like periods in April this year compared to previous years Covid-19 has killed from 9 to 44 times as many people is a historical fact, not a projection. As for peak deaths, the number of people dying each day is still going up. We don’t need a fancy statistical model to tell us this means we haven’t reached peak deaths yet.

Do you think it’s legitimate to take covid single worst week, compare to a week of flu and extrapolate our 9 to 44x worse?

Do you think there will be 10-30 million covid deaths this year?

It's not comparing it to 'a' week of flu, it's comparing it to the single worst week of flu in the last 7 years. I think it's one useful and relevant data point to see how Covid-19 deaths compare to flu, which of course is far from telling the whole story.

I have no idea what the world is going to do over the next 8 months, or what effect it's going to have. Do you?

You're missing how many people have gotten infected and where. If a region has not yet had substantial transmission of the coronavirus, then it won't have many deaths. This would also be like saying in August, "the flu has only killed 100 people, I guess the flu won't be very deadly this year."

Approximately 0.25% of NYC's entire population has died of COVID, which is significantly more than die of the seasonal flu.

What region doesn’t have it? What model shows we aren’t near peak death now?

Edit: seriously asking questions.

I'm not sure if you're being wilfully ignorant at this point. There are 3.3 million confirmed cases world wide, in a population of 7.8 billion. Say they've missed 90% of cases, and we're actually at like 100 million cases, that's still only 1% of the world that has it. Massive regions like India, Africa don't have it.

Are we going to see 9-40 times more deaths this season from those regions?

all of africa has 1663 deaths, and india has 1154. So yes likely.

The origional question was:

>I don’t get it. I found that there so far have been covid 234,133 deaths [0] and season flu can cause up to 646,000 deaths [1].

9-44x is 6-28 million covid deaths this year. Is this the number you think is likely?

> What region doesn’t have it?

Different regions have been infected to different degrees; in the worst hit areas like New York City up to 25% of the population may have caught it by now, but in many (most?) regions less than 1% of the population has been exposed so far.

> What model shows we aren’t near peak death now?

The question is, "What are you modeling?" Are you modeling what happens if the pandemic is contained at this point, and infection rates stay constant, or are you modeling a fast or slow spread to "herd immunity" (currently estimated at 85% of the population catching it)? In some situations, sure, it's almost over, in others, holy shit, it hasn't really started yet, this is the calm before the storm.

> I found that there so far have been covid 234,133 deaths [0] and season flu can cause up to 646,000 deaths [1].

How are you counting covid-19 deaths? You're counting the people who had a test, were confirmed positive for covid-19, who then died.

How are you counting flu deaths? You're looking at people who had a test and were positive for flu, but you're also including people who had flu like symptoms but no test. And you're looking at excess mortality and seeing what strains of flu were circulating. And you plug all this into a formula and a number drops out the other end.

If you only count confirmed deaths you'll see much larger numbers for covid-19 deaths. If you use the same excess mortality method you'd again see much larger figure for covid-19 deaths.

But this way of comparing confirmed covid-19 deaths against statistical modelling for flu gives a misleading impression: that flu and covid are mostly similar. They're really not, many more people are dying from covid-19.

“I’m counting” deaths by the sources I linked.

The article is figuring 9x-44x worse by taking the worst week of covid and comparing to same week of seasonal flu.

I’ll rephrase my original downvoted question, do you think there will be 10-30* million deaths of covid this year? If not - then the 9x-44x is total bullshit.

You likely won't because of the actions people are taking, so there won't be as many people who actually get the disease. There were nearly 200 thousand confirmed deaths from Covid-19 worldwide in April, and that's with much of the world under lockdown. There would be many many more without the lockdown.

Can you find a single source anywhere that modeled or even causally expected 10 to 30 million deaths without lockdown?

Edit: can’t reply directly, but I’d like to see something more than an alarmist article from March 8th. Maybe a recent study? Maybe ANY projection that supports the claim? On March 8th we didn’t know most people that get it will remain asymptomatic.

https://www.news.com.au/technology/public-health-expert-warn...

If you expect 60% of the world to get it and it has a 1% fatality rate that's 46.8 million deaths.

edit:Professor Gabriel Leung, chair of public health medicine at Hong Kong University, is an expert on coronavirus epidemics and played a key role during the SARS outbreak in 2002 and 2003.

It wasn't an alarmist article, it's an article about a projection by Professor Gabriel Leung. I'm going to be honest, at this point I don't think you're going to accept anything, and I don't feel like wasting my time.

> at this point I don't think you're going to accept anything

Oh, no, don’t make that mistake. I am definitely open minded. I can no more predict the future than you.

I just want to confirm you think that there will be 10-30 million Covid19 deaths this year based on a projection a guy made on March 8th and then taking the worst/peak week of covid and extrapolating that out.

That is your position, correct?

It really doesn't matter what my position is. I'm nobody. I've looked at what people who do have actual knowledge on the subject and formed my opinion on the matter, but my opinion is quite literally not relevant to anyone but myself.

I have a simple covid related question, hopefully someone can answer this here.

We have been under lockdown for over 6 weeks here in chicago. But chicago had highest number of deaths all of this week[1]. If we were under lockdown 6 weeks ago how did the people who died this week get infected. Does this mean ppl aren't following the guidelines properly ? I feel like I am missing something obvious here.

1. https://chicago.suntimes.com/coronavirus/2020/4/28/21240216/...

I'd love to know the answer on this question too. The only theory I have is that it's being spread through key workers. So those caring for other like hospitals and care homes.

> Does this mean ppl aren't following the guidelines properly ?

Yes, most likely. Folks still have to go out for groceries. Essential workers spread it to their families. The city had to close down various beaches and trails because people weren't staying far enough apart. etc. etc. etc.

So if make groceries pickup only and hefty fines/jail times for flouting guidelines. We would've had lockdown only for 2 weeks, in theory?

Or civil unrest.

You'd have to do something about the grocery pickup folks, healthcare workers, people maintaining other critical infrastructure, etc. too.

China can manage this (but it took more than two weeks). We probably can't with our current political system.

No. Places with such severe restrictions have still had lockdowns much longer than 2 weeks.

That's what they did in parts of China. Still took some time to get the numbers back down.

Because we don't/can't do perfect lock-downs -- everyone literally stays home for two-to-six weeks and doesn't move from their home -- you can really think of lock-downs as doing two things.

1. It effectively reduces your population.

Not everyone can stay home, because we still have stuff that has to get done, but say 80-90% of your population can stay home, because they're retired or they have a white collar job that allows them to work from home or because they're children and school is closed or because they've been fired because no one is buying what their job is selling.

This means that instead of 85% of your population needing to catch the disease for it to burn itself out, you're only waiting for it to go through 85% of 10% of your population before it burns itself out. (Unfortunately, because that 10% includes the people who take care of the 0.5% of the population that is most vulnerable, this still kind of sucks for that 0.5%, and you're really waiting for it to go through like 10.5% of your population.)

You'll still see increasing deaths during a lock-down until all of the grocery store workers and doctors and firefighters who are still pretty exposed out in the world catch it and some of them die. This also sucks and is kind of unfair, especially since some of the jobs which are pretty essential weren't particularly great before the pandemic.

2. It effectively reduces the R0 of the disease even among the essential personnel.

The R0 of a disease (last estimated at 5.7 for the coronavirus) is not constant, but depends on the conditions it is spreading in. Even among the essential personnel still out there risking exposure, the R0 is lower with everything shut down than it would be without the lock down, because the lock down still reduces the number of people each essential worker comes into contact with on a given day. Dropping off packages at a hundred doors a day puts you in marginal contact with a lot of people, but it's still less than if you did that and then went out to a bar and a restaurant and a movie afterwards.

Suppose the lock down is dropping the R0 among the active population from 5.7 to 2; that's still not great, but that means that instead of 85% of the 10% of the population still out in the world catching the disease, only 50% of the 10% will be expected to catch it. This is more awesome and less unfair, and also protects the vulnerable populations that come into contact with the essential workforce.

So our imperfect lock downs don't stop the disease cold in its tracks, but it does probably reduce the number of people we expect to catch it from 85% of the population to -- if we're lucky -- 5% of the population.

To put it gently, the science behind lockdown limiting the spread of a virus in a urban environment is...limited. But they felt they had to do something. It's kind of like making the medicine taste bad so you'll believe that it is doing something.

> It's kind of like making the medicine taste bad so you'll believe that it is doing something.

No, that's so little kids don't think it's candy.

It's possible to get infected even if you're following the guidelines. For example, essential workers may get infected by a carrier performing essential activities.

Doesn't lend credibility for suggesting lock-downs for future pandemics. Yay.

I don't think that's fair.

Lockdowns have successfully helped flatten the curve - changing spread from exponential to linear. That's worth something.

Why not?

Looking both ways before crossing the street has pretty good credibility. But it's no guarantee.

Has the growth in daily deaths been exponential in the last few weeks, or linear, or sublinear?

Chicago's lockdown may not be very effective. For instance, per https://www.apple.com/covid19/mobility/, there's a way smaller decrease in movement than, say, London or Paris.

> say, London or Paris.

am I looking at this wrong? transit decrease is

London - 85%

Paris - 86%

Chicago - 77%

Look at cars and walking.

The goal of the guidelines is not to prevent 100% of all cases. That goal would be impossible and unreasonable. Even with perfect compliance, the guidelines are not written to prevent all transmission. Again that would be impossible and impractical.

You would have to refrain from opening your doors or windows. Would have to wash all food items, all packaging, and any other items from any store, or anything shipped to your home, with some disinfectant outside of your home before bringing them into your home.

You would not touch anything outside your home without using gloves, ever, and your gloves would be disposed of outside of your home... etc., etc.

And not only you, but every single person, including children, would have to do all these things, and more. And the few examples I wrote don't even account at all for in-home transmission.

The virus is going to spread. That's not going to be prevented. There will be many individual cases. The guidelines are just for slowing it down in aggregate numbers.

If you're on my Facebook feed (which you likely aren't) then you'll see a bunch of people basically saying, "I'm going to live my life and do what I want. Enjoy your quarantine." There's a non-trivial segment of the US population that are ignoring the quarantine and openly flaunting their behavior.

It should be said, there's all kinds of situations where it makes sense to compare apples and oranges. You could compare their nutrition, to see which you should eat. You could compare their price, to see which you should buy. As long as you're aware that you're comparing two things that aren't identical (and you always are), it's often perfectly valid to compare apples and oranges. Just don't claim they're both apples.

I don't think anyone comparing Covid-19 and influenza, is unaware that they are different viruses. But, as when we compare gun violence to drug overdoses to traffic fatalities, it is often a useful and justifiable thing to do, to compare different dangers. Should I take the risk of flying (e.g. after 9/11), or drive 500 miles? Well driving and flying are very different, but it can make sense to compare them.

I have to wonder whether you read the article. Your comment is technically correct but doesn't address the point that it is making: that we are not counting the apples and oranges in comparable ways.

For people getting hung up on the apples to oranges analogy: imagine you're a farmer. You're hearing reports that there's a huge surplus of oranges and everyone is cutting production. You have a lot of orange trees, so you're understandably upset by this and decide to look into it. Some journalists, politicians, and your uncle Jimmy say it's all overblown—there are about as many apples as there are oranges on the market this year. No surplus, no need to cut production.

Then Agricultural American puts out an article showing that data on oranges is being reported by farmers as oranges are picked, whereas data on apples is being estimated based on much fewer reports. The conclusion is that the apples-to-oranges comparison should be taken with a grain of salt.

This is in addition to reports that show orange yields are likely underreported based on overall measures of fruit on the market[1], as well as anecdotal reports of fruit storage being overwhelmed by oranges[2].

Ok I think we can go ahead and report this analogy as another COVID-19-related death.

[1]https://www.nytimes.com/interactive/2020/04/21/world/coronav...

[2]https://news.sky.com/story/coronavirus-body-bags-begin-to-fi...

The surprising thing here is that the often quoted flu death number isn't an actual count it is a statistical estimate based off the must lower recorded count. Not surprising in retrospect but not obvious when people trot that number out.

I think boiling it down to a number is killing understanding. This is far more deadly for elderly than even this estimate makes it sound. The CFR for over sixty year olds in WA being an absurd 28%!! Under sixty? The flu is probably more deadly. Certainly understand 20.

So, just like with different fruits, you can't compete easily when boiled to a single facet. Almost like comparing the color of two fruits, without acknowledging the wide variety in color about the fruits.

I suspect political leadership will avoid bringing up the abject failure of government to protect the elderly. Meanwhile they'll use this as an opportunity to ratchet up emergency power, make precedents, etc.

every now and then I am reminded that there are in fact lots of different metrics that can be used to compare Apples and Oranges.

Yeah, but they are both denominated in the same thing, human lives.

Bit easier for people to get their head into than QALYs

https://en.wikipedia.org/wiki/Quality-adjusted_life_year

One of my takeaways from his article is it sounds like he never actually tests his mortally ill patients to see if they have influenza.

Sounds like he is misleading everyone as bad as the CDC.

I had a bad "flu-like" illness in February and they diagnosed me over the phone and never tested me.

Could have been Covid, could have been the flu, could have been neither.

Even if we're not testing everyone for Covid-19 they're clearly taking it more seriously than flu.

> The CDC should immediately change how it reports flu deaths. While in the past it was justifiable to err on the side of substantially overestimating flu deaths, in order to encourage vaccination and good hygiene

Yeah... except this is exactly what they are doing for COVID deaths, read the below.

https://www.cdc.gov/nchs/data/nvss/coronavirus/Alert-2-New-I...

Why can't we compare apples and oranges? They're fruit. They have nutritions. I don't like this metaphor.

How about oranges vs lemons? That way they are superficially similar but one is a lot more palatable than the other. Regular coronavirus = eating an orange. SARS-CoV-2 = eating a lemon...

We can, but at the same time it would be different than a comparison of say a granny smith and a gala apple.

Oranges share fundamental properties with apples - they are ”fruits” and ”pip” ones at that, and can be compared as such - but there’s a clear distiction and a reason for the grouping of ”apples” and ”oranges”.

> Why can't we compare apples and oranges?

We can! But, we must do so carefully.

"They're both sweet" cannot be used to conclude "they are the same".

The metaphor can apply to some things, but there's no good reason as far as I can tell for us to not compare COVID-19 to influenza. Experts desperately want to be right, so they're going to publish articles like this, which are effectively saying "Hey! Look at my shiny penny!"

so we go from

>3,448 to 15,620

confirmed deaths to

>25,000 to 69,000

estimated deaths? It's no wonder Americans have trust issues with medical experts. That's an absurd number created just to scare people into getting vaccinated. Which appears to be the authors opinion as well.

> While in the past it was justifiable to err on the side of substantially overestimating flu deaths, in order to encourage vaccination and good hygiene

I feel like trust is entirely undervalued in the scientific community. How are lay people expected to follow scientific advice if we can't trust the scientists?

Why the downvotes for this guy? The article makes the specific claim that the flu death numbers are allowed to be inflated through lax CDC reporting rules so that people are encouraged to get vaccinated and practice good hygiene.

Shouldn't that make us skeptical?

After this is all settled we can compare all-cause mortality. It may be complicated by lockdown side-effects though, so we might never know.

There are countries that are under lockdown with relatively small exposure to covid which can be used as control.

Even within a country, lockdown rules are often homogeneous, while the spread of the virus can be quite patchy, allowing for some form of control.

This thing of misleading people on the details so that they make the Correct Decision in the short term seems very dangerous.

I know in tech there's a tendency to do something similar when faced with clueless management, feeding just the information you know they need to hear to get them to make the right choice and think they came up with it themselves. Really don't like this applied on a society level though.

Flu, seasonal pneumonia... what's the diff?

Flu are caused by the viruses, pneumonia is the serious bacterial infection of the lungs potentially caused by having influenza.

You would count seasonal pneumonia if you wanted to see how serious that seasonal flu was.

Technically no, in addition to bacterial pneumonia (which is often caused by opportunistic infection after a virual invection), there is straight viral pneumonia without any bacterial infection. Even fungi by themselves can cause pneumonia.

The difference is that flu does not always cause pneumonia.

> The difference is that flu does not always cause pneumonia

Did not think I implied it did. But no, you’re right, it’s why the flu is more dangerous old people.

My understanding of fungus introduced to the lungs is that it’s also carrying bacteria. But yes, there are lots of things you can introduce the lungs cause light to severe pneumonia.

If you want to pro tip, stirring your protein powder well, if you don’t and inhale as powder it on your first sip you can have a bad couple weeks.

From the article:

Doctors, scientists, and the CDC, basically everyone we've been told to always trust when it comes to these matters, are and have been 'substantially' inflating flu death numbers to manipulate public behavior (e.g. to get more people to be vaccinated).

As well intentioned as that may be it is exactly why some people are skeptical about the flu vaccine. They sense manipulation and run from it. This patronizing behavior from the scientific and political class is destructive.

I agree. But within that same batch of people, faced with undeniable first-hand evidence of the severity of something like covid-19 will inexplicably choose to blame 5G antennas. Nutso people are nutso is my only conclusion.

> have been 'substantially' inflating flu death numbers to manipulate public behavior (e.g. to get more people to take vaccinated).

This is the article writer opinion, it is not the CDC official stand on the matter. Similar calculations are currently done for Covid-19. Many articles show the discrepancy between total reported deaths, Covid-19 reported deaths and medium values in previous years. It makes sense.

That the numbers are not perfectly correct does not mean that they are not good enough to use for policies and to inform the public.

> This patronizing behavior from the scientific and political class is destructive.

I like the transparency of publishing the data and the methods to calculate it. That is the contrary of patronizing. I hope that transparency continues to be the norm, even when makes the number easier to criticize for people with little information.

Agree that lying to people (even by selective omission) is a great way to undermine confidence and trust.

Does anyone feel like this happens with some approaching hurricane coverage where the scientists may overestimate things like storm surge to try and prod people to evacuate? Now when the next hurricane approaches they remember what happened last time and might be less likely to leave.

Article critiques data set and gives zero details on the CDC calculation methodology, and provides zero relevance to any new or interesting data analysis, but it surely mentioned Trump a few times.

A simple Google search will provide you with actual data sets, nationally, state wide, and even globally, by very competent scientists, some of whom have studied this specifically for decades.

There are various places you can research for example: FluView, FluMomo, Virologic Surveillance, Journal of Infectious Diseases, etc.

I recently read, Influenza-attributable deaths, Canada 1990–1999 Published online by Cambridge University Press , so this MD's Scientific American blog post comes across as pure garbage.

so, how are his conclusions different than the research you're referring to?

Inaccurate comparisons like this one are often made in bad faith by someone who knows full well that it's misleading but it serves a political purpose.

Assuming good faith intentions by the current US political leadership is comically naive at this point.

> The former [COVID] are actual numbers; the latter [Flu] are inflated statistical estimates

With doctors being instructed by the CDC to count every death that could have potentially been COVID as a COVID death, in the face of near total lack of testing, saying that the numbers are not inflated is a bold faced lie.

https://www.mprnews.org/story/2020/04/07/covid19-death-certi...

https://www.cdc.gov/nchs/data/nvss/vsrg/vsrg03-508.pdf

Fortunately you can look at mortality rates and you can see that in Europe it's even worse than the official statistics. Excess mortality is twice higher than recorded COVID deaths in the Netherlands for example.[1]

Even though there is a harvesting effect and you can expect mortality to be lower than average in the next few months (the famous "people that would have died anyway"), the data is not exactly substantiating conspiracy theories such as yours.

[1] https://www.rivm.nl/en/news/excess-mortality-caused-by-novel...

Excess death in Italy is estimated about twice the reported death count. In the most affected regions it is 3-4x.

The author may have a point but his numeric example is also comparing apples and oranges:

> If we compare, for instance, the number of people who died in the United States from COVID-19 in the second full week of April to the number of people who died from influenza during the worst week of the past seven flu seasons (as reported to the CDC), we find that the novel coronavirus killed between 9.5 and 44 times more people than seasonal flu

He is comparing the number of deaths from a disease to which we are largely immune (the flu, through flu shot or previous strains) to a new disease to which we have no immunity. Of course the number of cases will widely differ. But I think most of the educated people who compare covid 19 to the flu mean in term of IFR, i.e. the ratio of death to the number of people infected. Of course if that ratio is applied to a much greater number of infections you will get more absolute deaths.

Also is the way we account for deaths between flu and covid that different? I understand people accounted as covid death don't really die from the virus but from the effects of the virus (pneumonia, heart attack, etc). Isn't the flu killing the same way, and do we do flu tests for each death from pneumonia (or whatever the flu ends up inducing)?

he's just pointing out that nobody writes the flu as the cause of death on a death certificate. The CDC simply releases estimates, while we've been tracking the severity of COVID-19 deaths by asking clinicians to write the cause of death as COVID-19 on the death certificate if they tested positive.

Patient has flu and dies of a stroke? Flu can increase the chance of stroke by 50% but the cause of death was the stroke, not the flu. Same thing happens to covid-19? They died of covid-19.

The point is that the data collection methods are so different there's no possible way you can really look at the numbers and tell if one kills more than the other.

Your explaination seems sensible but I don't think this was the point that the author made. The way I read it was he meant that the CDC overestimates the number of flu deaths and therefore anyone thinking covid's lethality is comparable to flu is wrong as the flu kills a lot less.

Reply on news.ycombinator.com