Deconfinement: The Price of Human Life | Hacker News Reader

Deconfinement: The Price of Human Life

mondaynote.com

58 pointsevo_959 comments

Some good points, although I'd toss another thought into the mix. The confinements and shutdowns and shelter in place orders also cost lives.

We know that recessions cost lives. One study (published in The Lancet, so worth taking a second look at) suggested 500k excess cancer deaths from the global financial crisis[1], and another study (published in the BMJ) suggested that excess suicides in 2009 alone across the US and 53 other countries totalled some 5k[2]. I don't know of a single comprehensive effort, but we know recessions reduce the resources available at the margins for food, medicine and shelter, and increase the suicide rate. The net effect for something like the GFC seems likely to be on the order of tens of thousands of deaths (and perhaps significantly higher), and the recession caused in part by Covid-19 and in part by our reaction to Covid-19 seems to be within an order of magnitude of the GFC.

Thus, it seems fairly obvious that at least some responses to Covid-19 would cost more lives than they saved. If you're evaluating a measure that, let's say, would save X lives from Covid-19, but would cause the resulting recession to go from, Y% as bad as the GFC to Z% as bad as the GFC, there are absolutely values for X, Y and Z which would make the measure net negative on lives.

I'd be happier if that tradeoff got acknowledged a bit more often, even in a perfunctory "well, we're clearly MILES away from the break even point, so let's ignore it" sort of way. If you pretend that tradeoffs don't exist, you're unlikely to reach an optimal decision.

[1]: https://www.telegraph.co.uk/news/2016/05/25/financial-crisis...

[2]: https://www.bmj.com/content/347/bmj.f5239

If the movement restrictions cost lives, why are total fatalities down? (even including SARS deaths)

Because total = increase - decrease.

If the decrease is bigger than the increase, you will see the total drop.

Also, movement restrictions cause deaths in longer terms than the virus.

How?

Movement restrictions causes financial difficulties for companies, which cause job losses, which can cause depression. All this can take time.

These deaths aren't as acute. Deaths from economic depression will take longer to manifest than a sudden illness.

That sounds like a call for a coherent policy rather than the do nothing that's being proposed here.

The primary risk factors of economic recessions take years. Most cancers take years to kill people. The effect of unemployment on suicide peaks at around 2.5 years of unemployment [0]. I think that we're going to have an economic downturn with repercussions that will last long after a SARS-CoV-2 vaccine exists. I also think the ends currently justify the means, but that's just a gut feeling I've got.

[0] Milner, Allison et al. “Long-term unemployment and suicide: a systematic review and meta-analysis.” PloS one vol. 8,1 (2013): e51333. doi:10.1371/journal.pone.0051333

They undoubtedly do cost lives but:

1) You're looking at net lives, and there's no particular reason to think we're net negative yet. Let's imagine a brief lockdown to avoid overwhelming the emergency rooms with people needing ventilators is very net positive. And let's imagine crippling the economy by fully locking down everything until a vaccine passes clinical trials in 18 months is very net negative. Even if you accept these as true, we might have weeks to go until the crossover point arrives...or we might have already passed it.

2) Which brings us to the next point - even if we're already in net negative territory, deaths from Covid-19 are (hopefully) front loaded; they're happening now. Deaths three years from now because the long term economic growth rate is 0.2% lower than it otherwise would have been are still in the future. I linked to some studies suggesting the GFC cost a fair chunk of lives in the end, but you wouldn't have been able to detect any of that the week after Lehman failed; the fallout had yet to land.

And finally:

3) Coming back to the point in the linked article, there's more to life than strict maximising quality-adjusted life years. There's a lot of ways you could, in principle, save lives via a stroke of a pen. Boxing, white water rafting, surfing, diving, and mountain climbing are all shockingly dangerous sports. We could ban them! It'd straight up save lives. Should we? I, for one, would strongly argue no. But that cuts both ways. Even if we had inarguably correct data showing continuing the lockdowns was starting to cost lives, that's the starting point for the debate, not the conclusion. We don't (and shouldn't) just do the things that would save the most lives.

Regarding point 3, I think your examples, being individual sports that people choose to participate in, are kind of straw men. For one, relatively few people die doing them. For another, as I mentioned, they're activities that people participate in, knowing that there are some serious risks.

A better example would be "let's end all farm subsidies for sugar and tobacco." You could arguably include corn here, as well. Go a little further, and you can start talking about national sugar and tobacco taxes. All these measures simply increase the price of sugar and tobacco, which has been proven to reduce consumption, and would certainly have the effect of reducing obesity and cancer.

I find it a lot harder to argue against those things than to say we should ban surfing, boxing, or rafting, because a few people get injured or die every year doing them.

My point isn't that they're equivalent (they're not), but that the analysis never stops at "this will save lives, so we should do it". There's always more layers, such as here your very valid analysis of whether people are choosing to participate.

No, but you'd get the point across better by using a less ridiculous example -- that's my point. Nobody's ever going to support a ban on surfing because it's just ludicrous on its face. Ending tobacco and sugar subsidies is something that could get real consideration today.

> but we know recessions reduce the resources available at the margins for food, medicine and shelter, and increase the suicide rate

OK, agreed, because of hardship.

But a lot of that hardship is due to enormous wealth inequality (I'm assuming you're a US person). How about you don't have titanic inequalities instead? Would that work?

It would. The fact that a science-based public health response to a public health crisis is causing extreme economic hardship for so many people right now indicates that our economic system is broken.

The science behind reducing transmission of the virus is sound. What's going to get people killed in the near term is twofold: people undermining or protesting the quarantine orders; and people who need to go to work in order to support their families and not get evicted, who end up contracting the virus and either dying or spreading it to others. The former is either due to some kind of fucked up value system that interprets a temporary public health quarantine as some kind of intolerable infringement of rights. The latter is because the American form of capitalism leaves so many people so far behind that they can never get ahead.

The confinements and shutdowns and shelter in place orders also cost lives.

You have to remember the economic slowdowns started before the lockdowns were in place. Even without lockdown we were in for a lot of economic pain.

> We know that recessions cost lives.

Interesting, I read a different view: https://www.theguardian.com/commentisfree/2020/apr/28/a-rece...

I am not sure how to reconcile this.

Haven't they also cost lives from hindered access to medical care for stuff unrelated to Covid-19?

>The Hong Kong Flu caused more deaths than are projected for this current pandemic and yet, at the time, no one thought of bringing entire economies to a halt.

The other way to look at it, even with the quarentine this pandemic will kill a similar number of people.

If the reports were true that "half the workers were in bed," that would mean there were millions of cases, yet only ~50k deaths. Compare that to the current pandemic, where France has 25k deaths in ~150k cases.

There have been many, many more cases than the official confirmed (=medically tested/diagnosed positive) cases. For example data in Spain points to 5% prevalence (~2M infected) with ~25k deaths (official number, may be higher due to unverified and side cases).

Do you know if the serological data in Spain has also per-city estimates? I doubt the 5% is evenly distributed across the population, I'm expecting higher percentages in Madrid, for example.

If I remember correctly the data from Spain is also available itemized by region (autonomous community), and indeed the most population-dense regions have a higher prevalence. I believe the data isn't broken down by city, but it should be obvious that they're going to have a higher prevalence as well.

I can't follow the argument. The Hong Kong Flu caused more deaths without the economy shutting down than this epidemic with the economy shutting down, therefore what?

The shutdown looks like it will lead to both epidemics having roughly the same amount of deaths. Unless you think the shutdown had minimal effect, without it the death rate would be far worse than the Hong Kong Flu.

> In the end, these determinations all place an unspoken price on human life

We are already doing this by letting people drive with personal vehicles every year even they cost dozens of thousands of lives. There is no zero risk, but we decide (we, as in, the society at large, not individuals) what level of harm is "acceptable".

But here the decision making is localized to individuals, and their decisions can formally be translated into pricing human life. I'd hate to be the one having to make these calls.

I mean, various insurances price human lives already, so it's not exactly novel, but here the decisions directly cause these tradeoffs to play out (whereas e.g. a life insurance does not cause a death for monetary gain).

The line between deciding that a particular level of harm is acceptable, and accepting it by default, is blurry. And yet allowing, through inaction, a human being to come to harm is not functionally or morally different than deliberately harming them. And so we wash our hands of responsibility by telling ourselves we didn't do anything. But who are we fooling?

If that definition were taken, and people were moral, you could infinitely blackmail moral people by threatening harm.

We have incredible amount of regulations on what personal vehicles can do and how they can be made. We did not just accepted deaths from automobiles, we wrote driving rules, mandate seat bells, automobile makers have to prove their cars pass safety tests.

Yet we still have dozens of thousands of people dying from them every year and nobody stops sleeping at night at the same time.

Unites States have 35,000 to 40,000 car crashes deaths per year. It is what we get. Cumulative Covid deaths so far were 90,694 and it is not done yet.

> nobody stops sleeping at night at the same time

When driving deaths go up, people do react. There are driving safety campaigns, dont drive drunk campaigns, wear seatbelts campaigns. The roads are changed, signs are put up, speed limits are placed.

> When driving deaths go up, people do react. There are driving safety campaigns, dont drive drunk campaigns, wear seatbelts campaigns. The roads are changed, signs are put up, speed limits are placed.

Yes, but these show marginal improvements. I don't think you can make the case that any of these measures, even combined, would be able to reduce the death toll by even 50% over time. It means that, all in all, we are pretty much willing to sacrifice at least X amount of people per year for a number of reasons.

Something that seems not to be considered by many observers is that if a disease is not eliminated, it will continuously either claim lives or reduce efficiency (which claims lives).

Is there reasonable hope for a virus like this to be completely eliminated? I figure if it's in that sweet spot of contagiousness, incubation time and mortality, it's hard to eradicate completely. And unless it's eradicated completely, we'll absolutely have to live with reduced efficiency in hyper-alert testing of anyone with symptoms, unless we find that it ceases to be a serious problem once humans have had contact with it.

It probably depends on how long acquired immunity lasts, which we don't know yet. People remain immune to some diseases, like measles, for a lifetime after surviving the disease or being vaccinated. Immunity to other diseases, like influenza, disappears quickly (even immunity to the same strain lasts less than a year).[1]

If COVID-19 is in the latter category, it will be difficult to control and we'll probably have to learn to live with it, as we have learned to live with the flu causing hundreds of thousands of deaths each year worldwide.

1: https://www.sciencemag.org/news/2019/04/how-long-do-vaccines...

That likely depends on how effective we are at eliminating it in the coming months. I doubt it will become endemic in Korea or Singapore.

I wonder if even influenza could be erraticated in the coming years with mass testing and contact tracing. The lockdowns ended flu season early this year.

We'd have to eliminate it worldwide, or it would travel from country to country like the flu does.

In particular, from Mexico. There's too much travel between the US and Mexico to prevent it from crossing the border, and there are 50,000 confirmed cases in Mexico already, with roughly 2000 more being found every day.

Given the relative number of cases, Mexico should be stopping travel from the US, not the other way around.

Reducing travel from Mexico has been a popular position in the Us for decades.

Let's not discount pharmacological options. HIV doesn't have a vaccine, yet it can be fought with a proper (admittedly very complex) drug regime.

The same can be probably done for SARS-CoV-2, although we'll have to wait for specific drugs to reach better efficiency.

And I add, we don't need to make this virus disappear (it would be better, but even with a very efficient vaccine it is unlikely). If we can prevent hospitalizations (and possibly deaths), we can likely live with it like we did with many other pathogens.

Bubonic plague, tuberculosis and polio still exist outside the West, but they don't cause much inefficiency here. I doubt doing some ongoing testing for SARS now that it's been mostly eliminated in Asia will permanently cripple Asia's economies. The inefficiencies of continued outbreaks on less well organized nations are harder to predict but possibly much larger.

I'm not qualified to answer that;

but as a layman I share your concern. From lay reading it seems that the large organisations were hoping that situations like this could be arrested at source or in the early stages.

To double down on my earlier comment; the inefficiencies the disease cause are not strictly the testing etc.

Look at the regular flu [1] [2] it has been "measured" as costing economies some pretty significant sums every year ($87bn USA, £1.35bn UK, different standards used); of course this could be said to pale against the size of these leading economies [3] landing at less than 1%. An annual flu tax is certianly not ideal.

Additional diseases will have different "tax rates", and I have no idea to the way in which they will interact.

You have to bear in mind that a tax rate is a form of inhibitive cost, and the burden of disease could be holding other economies back.

[1] https://www.cdcfoundation.org/pr/flu-costs-United-States-87-...

[2] https://www.hrmagazine.co.uk/article-details/flu-costs-the-b...

[3] https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nomi...

I'm not sure the person you're responding to was using it in this sense, but the word "eliminate," in epidemiological terms, means to get the virus under control such that outbreaks are reduced in scope to local clusters that can be dealt with using methods like contact tracing. In that sense, yes, we absolutely can, and should, eliminate this virus.

You're right - I was using it in a totality sense. While I know enough to know it won't get there the point was to draw the extreme and highlight that deaths aren't the only cost on the scale from here to there.

Unlikely, as it can be passed to animals, then mutate and return to humans. The flu vaccine is designed and distributed annually based on the most likely strains. My guess is a similar pattern for strains of coronavirus.

We are in better position then before tho. We have cheaper tests now and we can manufacture them more quickly. The way people get infected is also clearer now and we can have partial limitations on economy more effectively. Pretty soon, we can be in position to test, say, 50% of workers in factory every week and having asymptotic sick at home. People got used to wearing masks in public indoor spaces and to sanitizing their hands upon entering stores.

Overall, we are in much better position to make selective limitations, opening and closing depending on what numbers are. Countries that locked down soon are opening now with small numbers of infections - they are able to do tracing and test all contacts of any new infection.

We are getting better at controlling the pandemic.

Yes, people are really underestimating the effects of the virus that aren't fatal. While I know nobody that's died from it I know several who've been off sick for weeks.

To add, it's a bit concerning seeing that the military has put a ban on enlisting those who have tested positive, were hospitalized, and recovered without the ability to be waivered. (I mentioned this in another post about a week ago when it was reported).

Then just a few days ago, it was reported that ~13 sailors who were positive, recovered, isolated, then tested negative two times in a row, retested positive and some were showing symptoms again.

It could be due to inaccuracies in testing, but given that some showed symptoms again and tested positive, it seems possible that they either didn't develop an immunity and caught it again, or this could be a chronic illness.

The reason for banning enlistees being a vague "as a precaution", is a little suspicious. Who knows though. Regardless, it's a something to keep an eye on as it develops.

> just a few days ago, it was reported that ~13 sailors who were positive, recovered, isolated, then tested negative two times in a row, retested positive and some were showing symptoms again.

It's unfortunate that the guy who highlighted the problems on the Roosevelt was also fired.

The literature has a lot of case reports, but nothing substantial last time I checked. This makes a non-sensationalistic look at those issues (and more importantly, how much they occur) difficult.

Regarding sending people to the moon, I don't believe we've changed how we calculate the value of human life in this scenario at all. Remember, astronauts volunteer to strap themselves to gigantic tanks of explosives, all for the chance to experience the universe personally in a way very few humans ever have.

Rather, I think the issues here are:

* America literally watched 14 people die on national TV.

* At the time of the Columbia disaster in 2003, the Space Shuttle was already a 30 year old vehicle (Columbia itself was 22 years old when it was lost), based on 40 year old designs.

* We realized we just don't need humans on the moon. Compared with robots, we're terribly fragile, inefficient, and largely superfluous on the surface of the Moon in a lunar exploration mission.

> We realized we just don't need humans on the moon. Compared with robots, we're terribly fragile, inefficient, and largely superfluous on the surface of the Moon in a lunar exploration mission.

Robots are inflexible. There's nothing like human eyes and hands.

Even if I agreed with you in this context (I don't -- hands inside space suits are clumsy, and eyes limited in what they can perceive compared to instruments), we can keep the people here on Earth operating the robots with ~1 second latency, and it will be fine. We even operate rovers on Mars with ~20 minute latency, and they've worked out well.

They've worked out well considering the extremely limited scope of what they can do. They can't, for example, brush the dust off their solar panels. Or even the simplest repair. Or dig a hole with a pickaxe.

I am under the impression the author is missing the point of the lockdown. The main point is not about "saving lives" / "avoiding deaths".

Lockdown is about (1) keeping hospitals under capacity and (2) hopefully slowing the virus to the point it's not an epidemic anymore.

If you don't keep hospitals under capacity, people who can be cured have to be either moved somewhere else at great cost (this happened, there have been people moved from France to Germany) or left to die because they cannot be taken care of (I am pretty sure this happened too). Please note people who have another issue than the Coronavirus are also affected. That's what everyone REALLY wants to avoid. Needless to say, letting the virus run its course and overwhelm hospitals would have an order of magnitude more deaths.

The hard issue is about how to re-open without letting the virus rampage again. It's a lot of difficult decisions, and no one have conclusive data on what works and what does not. It's going to be a lot of trial and error, as shown by the fact that no 2 countries have the same rules to reopen.

C2 are really fun to drive if you are at the wheel, but the back seat is like a big beach chair. The ones with a piece of cloth suspended on a frame. Do you think in a world of autonomous driving cars, we can make the cars lighter again? I suspect that in the future, around 2100, there will probably be a lucrative industry around turning old-timers into self driving vehicle. Once a technology is commoditised it needs to be commercialised in another way. Maybe then driving will be fun again, like a big theme park.

A friend of mine had one which went up in flames. There was a lot of paper ducting around the engine that eventually got soaked in oil with inevitable results.

It was interesting to drive, mostly I remember it would lean a long way over going round bends.

I find the complaint that it is harder to do something today, for safety reasons, a bit odd. Isn't the attempt to eliminate more and more human suffering a sign of moral progress?

It seems to me that inevitably, if you attempt to save marginally more lives, the technology to do so will become more complex.

But conversely, why have all the associated complexity of technology, if the human suffering is not reduced? What is the goal then?

The "Get and keep America open" plan is really a "keep dying" plan, but this is something few want to admit. Someone else pointed out how deceptive the "flattening the curve" charts were. Most showed a symmetrical rise and fall. In reality, that plan just keeps the hospitalization rate and death rate down to what the medical system can handle. Both stay high for many months until either saturation ("herd immunity") is approached, or there's a vaccine. The "flat" part is probably a year long, somewhere around the current death rate, which is in the 1000-2000 deaths per day range. (That may be underestimated. See this Financial Times article, which is looking at "excess mortality", or the change in deaths from all causes.)[1]

"Herd immunity" is a long way off.[2] That only happens when 70% or so of the population has had this virus. While there's not much data yet about how many people have cleared the virus and acquired antibodies, most figures are around 5% of the population, except a few cities, primarily New York, where it may be somewhere above 20%. So the dying is maybe 6-10% done, nationally.

The CDC is now avoiding death forecasts beyond mid-June.[3] Previous longer-term projections have been so far off as to be useless.[4]

Don't bunch up, and wear a real mask that protects you, not others.

[1] https://www.ft.com/content/6bd88b7d-3386-4543-b2e9-0d5c6fac8... [2] https://www.msn.com/en-us/health/medical/no-place-on-earth-i... [3] https://www.cdc.gov/coronavirus/2019-ncov/covid-data/forecas... [4] https://www.foxnews.com/politics/coronavirus-model-estimates...

Reply on news.ycombinator.com