Hacker News Reader

Comment by mikekij | Hacker News Reader

This entire plan is based on models that do not align with what we are seeing. The graph on slide two[1] of the official CA govt plan shows hospitalizations With Intervention rising exponentially, exceeding our hospital capacity in early June. The actual new infection data shows that the rate of new infections has been decreasing since at least 4/4 [2], meaning that the graph of active hospitalizations should be a bell curve with a maximum value of ~10,000 on ~4/20.

The mismatch between their models and the observed data does not engender confidence.

[1] https://www.gov.ca.gov/wp-content/uploads/2020/04/California...

[2] https://covid-19.direct/state/CA

View on HN

Something weird happened in California, where 20% of their tests were coming back positive, then all the sudden 04/04/2020 they started seeing <10% positive tests:

https://raw.githubusercontent.com/lettergram/covid19-analysi...

They also barely tested at all prior to 04/04/2020:

https://raw.githubusercontent.com/lettergram/covid19-analysi...

In generally though, the rate of positive tests is increasing nation wide:

https://raw.githubusercontent.com/lettergram/covid19-analysi...

Implying we are still on a solid growth curve, as our testing capacity has maintained relatively constant levels:

https://raw.githubusercontent.com/lettergram/covid19-analysi...

I can't speak to the hospitalization levels, but I suspect people are avoiding the hospital. My roommate decided not to get tested and stay home, even though he was having difficulty breathing. He seems past it after a couple weeks.

>Something weird happened in California, where 20% of their tests were coming back positive, then all the sudden 04/04/2020 they started seeing <10% positive tests:

This is not surprising. When tests were scarce, tests would be rationed to those most suspected of infection.

Tests are still scarce.

Maybe but it’s all relative. Even a little less scarcity can change those numbers in significant ways.

I can't figure out what you mean, can I go someplace to get tested and put my mind at ease that I'm not about to die or kill my familymembers? That's the only "relative" that matters here.

Are they scarce everywhere? I was under the impression that LA county, at least, could test everyone that was symptomatic.

No, they have been available to only “seniors and high risk individuals”.

Literally just today Mayor Garcetti announced testing for everyone is available.

That's great for LA county but there are 320 million of us who don't live there. That is: should they be scarce anywhere?

Remember, this is the USA.

This whole thread is about California's roadmap. LA County is something like 25% of the state. Why stop at the USA? With people moving between places, outbreaks anywhere will matter until there's a reasonable vaccine/treatment.

Maybe or may not be true.

I just got an email from local urgent care that they have the fast test equipment that can offer a result within 15 minutes if you are negative.

The situation is changing quickly

My friend in SF was told he wasn't severe enough to get tested a few weeks ago, but he's convinced he had it.

Assuming that's the Abbott 15 minute test, there's only going to be 50,000 of those a day for the near future (and it seems less than that in the immediate future).

That's a big chunk on top of current testing, but probably not enough to make it easy to move forward with a testing based strategy.

And the stories I've read say their accuracy is questionable.

The cases reported per country, or state are only a narrow window into what is actually happening. The most advanced medical testing company in the world, Roche, is Swiss-German so Germany's testing capability relative to most countries is off the charts. Combined with a highly distributed, yet also highly co-ordinated health care system they have been ahead of the curve in testing on every conceivable metric. As a result they have reported relatively high case totals, because they have been doing mass testing for a long time, but very low death rates because this enabled them to effectively isolate cases quickly.

As countries increase their testing capacity it is quite possible some countries (or states) will end up eventually reporting accelerating numbers of diagnoses even if the actual new case rate in the population is falling.

One way to mitigate that is to report only cases requiring hospital admission. This isn't ideal by any stretch, but is a reasonable proxy for the spread of the virus and at least gives a good indication of the impact on the health care system.

But hospital admission cases depends on perceived hospital resources too... if people think going to the hospital is worse than staying at home, or impossible, that will distort conclusions also.

That's true, but is mainly a problem when comparing one country or region with another. Within a region, assuming a roughly stable general perception of hospital capability, it should till give a reasonable indication of the relative rise and fall in numbers of severe cases. I don't know, it's a really hard problem.

> One way to mitigate that is to report only cases requiring hospital admission. This isn't ideal by any stretch,

It doesn't count all the frail people living in care or nursing homes who are put onto a palliative pathway and allowed to die, for example.

This could be about half of all Covid-19 deaths.

https://ltccovid.org/2020/04/12/mortality-associated-with-co...

> Data from 3 epidemiological studies in the United States show that as many as half of people with COVID-19 infections in care homes were asymptomatic (or pre-symptomatic) at the time of testing

> Data from 5 European countries suggest that care home residents have so far accounted for between 42% and 57% of all deaths related to COVID-19.

Those charts look very different from the Worldometer data, which shows the April 4th high water mark in new cases still holding 10 days later.

Their graph is linear growth in new cases per day until 4/4, followed by a gradual but spiky decline, even as tests/day keep increasing.

That’s because worldometer data is community reported. This data comes from the states (mostly):

https://covidtracking.com/

When testing is scarce it is used as a clinical tool to help make decisions about treatment. For example: patient has pneumonia. If its bacterial pneumonia a course of the right kind of antibiotics will be very helpful. If its COVID, those antibiotics won't do anything, and the patient may not get the therapies they really need.

Once testing becomes more available, doctors use it to rule out COVID even on patients who aren't very sick, to help make sure they those patients aren't spreading it, instead of just making treatment decisions.

Ideally once testing capacity gets high enough we could start testing every contact of an infected person, or just randomly sample people.

Additionally, notice that the orange data points (actual hospitalizations) are significantly under the "With Intervention" plot. So why even include the slide? Clearly the assumptions behind the "With Intervention" plot do not match reality TODAY, let alone in a few weeks.

Not a good sign if people are making actual decisions based on this.

My former physics professors would have not approved of a graph like this. D+ on the lab at best.

https://i.imgur.com/dJ7kHch.png

It's pretty sloppy to show surge capacity on the same graph as total hospitalizations, not active active hospitalizations.

> Clearly the assumptions behind the "With Intervention" plot do not match reality TODAY, let alone in a few weeks.

I feel like I’m taking crazy pills with the way people are talking about this whole thing still being out of control, when data shows the situation is even better than the most optimistic models were predicting. And by that I mean, pretend I was on HN 3 weeks ago predicting 25K deaths in the USA by today. I’d have been downvoted into oblivion and told I was delusional. But here we are.

Clearly NYC was hit far harder than the rest of the country, but it seems like every state and municipality is simply basing their models on NYC case numbers scaled by population. I don’t get it.

> I feel like I’m taking crazy pills with the way people are talking about this whole thing still being out of control

A week ago, I remember reading how the Bay Area was bracing for a bracing for a surge and it was the calm before the storm. The Santa Clara county covid-19 dashboard told a different story: day-over-day, new cases were roughly the same for the past two weeks. 2+ weeks into the lockdown, there was obviously no surge coming.

https://i.imgur.com/LrFZj3S.png

Looking at that chart on the April 7th, nothing pointed to a surge.

Isn't this just California being in the "3.6 roentgen, not great, not terrible" phase? I.e. number of new cases reported is limited by the number of tests that are able to be made?

Every region (and almost every country) is generally believed to be severely undercounting cases because of limited testing. There's not much reason to believe the problem is worse in California than elsewhere, and a severe surge would be evident in hospitalization and death rates no matter how bad testing is.

In mid-March we were told that we are "10 days behind Italy" and shelter-in-place is not to stop the spread (already too late for that), but to give hospitals a bit of a breather before the surge of critical cases and the Death Wave that follows. That was the justification for such extraordinary measures as a state-wide lockdown.

It's mid-April now. Where is the Death Wave? Either this virus is much less infectuous than it was touted, or much less deadly. Or both.

Sweden never locked down. Denmark is opening up after a lockdown. Italy that was touted as the prime example of hell on earth, had ~21,000 people died by official count. Even if it was undercounted by 50%, that's roughly 40,000 people, mostly elders on their last breath, out of country of 60 mln. And that's the worst case so far, in the world of 7 billion people.

Disregard all that, we're still being told daily that it's "war zone", the zombie apocalypse is just around the corner, etc. Only Big Brother like surveillance is going to save us. The lockdown is not an extraordinary measure anymore, it was an absolutely necessary first step, and now we have to take it further. Much further.

I wonder what it felt like for people in 1932 Germany.

Sweden didnt lock down and as a result has around TEN TIMES more deaths than its neighbours, and the death toll is raising by hundreds per day, when its hovering around 10 with Finland and Norway. I don't think you want to give Sweden as a good example.

Wikipedia says 1,200 deaths so far in Sweden. Sure that might be ten times more than in Finland and Norway but given the dire predictions of two months ago, and even more dire predictions of a month ago, you'd expect this number to be ten times higher than it is now, or a hundred times. What about exponential growth and stuff? In a country with no lockdowns this disease should go rampant, right?

1,200 people dead of 10.3 mln is 0.012% of the populace. Not good but not scary either, for the Worst Pandemic Of Our Lifetime.

I don't understand what you're suggesting, that California should ease up faster? It seems to me that reality being better than models had predicted is more the result of everyone flying blind due to lack of testing... California has probably been lucky that it turned out its infection was lower than New York's turned out to be. It is just hard to make a model starting from extremely shitty and noisy base data about facts on the ground, but someone still has to make very difficult public health decisions based on the best models available.

I think there was genuine concern that the quarantine as implemented would not be enough. You can see lots of stories locally wherever you are about how people aren't obeying the order. Italy was still in a lot of trouble despite having had some quarantine measures in place. Hindsight shows the quarantine was effective enough to get numbers down. The US, despite reacting too late, did lockdown fully relatively quickly.

I would agree that modeling after NYC seems like a bad idea given the lack of density and public transportation. On the other hand, Michigan looked pretty exponential, and as the lagging cases came in from other states, they also began to look fairly exponential as well. Despite cases staying elevated for now, the exponential growth has died down.

The US missed the worst of hospitals being overrun and it running away in multiple cities.

>The US missed the worst of hospitals being overrun and it running away in multiple cities.

I want to add ", so far" to the end of that sentence. I do not hold hope that this will be true as we fight among ourselves to reopen various portions of the country/economy.

> I do not hold hope that this will be true as we fight among ourselves to reopen various portions of the country/economy.

But things won’t be the same after reopening as they were before. People will be wearing masks (I see a lot of that already), keeping distance, etc. Couple that with common sense restrictions like “keep the elderly and sick at home”, “no big public events”, “no sporting events”, and “lots of restaurants have gone out of business”, it’s clear the opportunity for infection spread will be lower.

The line is the average of high and low estimates. The issue we are dealing with exponential growth so 1/10th x to 10x makes it look like the estimate is wildly off even when it’s surprisingly accurate.

Further, these estimates are much older than the data so showing their accuracy is useful.

I think the point of the plot is to suggest that intervention efforts are being quite successful relative to expectations at the time the "With Intervention" model was generated.

The brighter dashed line is there to show what happens if intervention is disabled completely.

I really like that the milestones are placed as questions, rather than dates. It is the right way to message the unknown.

This is ridiculous criticism. There are so many unknowns with this virus. We don’t know how many people have it, how fast it is spreading or what the mortality rate. Everything is a fairly rough estimate.

It turns out that social distancing is working better than expected, or the virus isn’t as contagious as expected, or slightly less severe than expected. That doesn’t mean the lockdown was the wrong thing to do, or that it isn’t still necessary. What it does mean is that we’ll be able to start easing restrictions as the hospitalizations continue to drop — and perhaps we can stop it completely with testing and contact tracing.

Whenever I see the IHME charts[1] (which are great) I think “I wish I could see the what the model predicted before today’s date so I can assess it.”

This shows laudable transparency, in my view.

(Also the model was directionally correct, it’s not like it was entirely the wrong shape.)

[1] https://covid19.healthdata.org/united-states-of-america/cali...

The IHME model is proving to be wildly inaccurate on the downslope due to bad data coming out China. Their projections will continue to creep up in terms of deaths over the next few weeks because the predicted drop offs won’t be happening, sad to say.

If you want to see examples, check out their predictions on Italy. Their figures have been underestimating by over 50%, and the gaps will widen each day over the upcoming week. Heck, the ranges for the last few days have been outside their confidence interval.

A significant portion of this probably is compliance, but I'm with you completely on "the numbers out of China, particularly at the tail end are farcical". There was a drop off a cliff in cases.

Could you share how you would assess the model?

Do you have a background in disease modeling or anything?

I'm curious.

Why should infections follow a bell curve? Shouldn't it be the derivative of a the logistic curve, which should be approximately exponential on both ends. I don't see how a quadratic should show up in the exponent; either mathematically or epidemiologically.

You're pointing out a great ambiguity in the state's graph. Is the hospitalization line a cumulative count of all hospitalizations? Or hospitalizations at any given time? Since the state's hospital capacity is related to hospitalizations at a given time, and not cumulative, I'm assuming the fit of that graph would be a bell curve with a max, followed by a decrease. People either die or leave the hospital.

Does anyone have the ability to walk through these explanations in a little more detail. I'm having trouble understanding why infections would be a bell curve, or derivative of a log curve.

Cumulative infections would not be a bell curve. They'd be a sigmoid curve, approaching some upper limit (e.g. 100,000).

Current Hospitalizations, which is the number we should be most concerned about, and is reflected by the orange dots on the state's graph, should start low, max out at a number, and eventually return to 0. That should follow something like a bell curve.

I think the confusion is between the "total infections" curve and the "current hospitalizations" curve. The difference between them is huge, and really important. The "Flatten the Curve" idea refers to keeping the number of hospitalizations at any one moment below some upper limit, ideally the state's number of available ICU beds. That's a bell curve.

> think the confusion is between the "total infections" curve and the "current hospitalizations" curve.

I think the confusion is that if the CDF is the logistic function, then the PDF is the logistic distribution. This is what archgoon is saying.

Bell curve refers to a Gaussian distribution, which is different but looks similar.

https://en.m.wikipedia.org/wiki/Logistic_distribution

http://visionlab.harvard.edu/Members/Anne/Math/Logistic_vs_G...

So the basic assumption of an exponential growth curve is: "rate of change is proportional to a population."

However, in the case of a virus, you eventually run out of population to consume; so you tweak the model to be "Let the rate of infection be the probability that a infected person encounters a non-infected person" (nicely, for a large population, this will give you the same results initially). This is, approximately, proportional to the product of the number of people who are infected times the number of people who are not infected (think back to chemical reagents and reaction rates). That is, (1-infected(t)) * infected(t). A function who's rate of change is that is the logistic function (and you can verify by taking the derivative of 1/(1+e^-t))

This is a simple model, ignores geography, ignores population change, and ignores changing behavior. It basically pretends everyone is an ideal gas molecule in a volume. But if you want to model something that "grows exponentially with a limit" it works alright.

I don't know under what assumptions the rate of change would yield a bell (normal) curve. (e^-x^2).

Here's a neat video by 3Blue1Brown that shows some simple simulations of an epidemic, and give a good visualization of the bell curve.

That's a good question. If control measures are undertaken to "flatten the curve" down to the capacity of the medical system, we'd expect to see new infections per day and deaths per day level out, maybe decline a little, but not decline a lot. Which is what the actuals show for the areas worst off.[1]

At some point, there's saturation, and "herd immunity". For this virus, that's somewhere in the 60%-80% range of the population. US population is around 333,000,000, so saturation is somewhere above 200 million.

We're nowhere near that. Current known US total cases are nearing 2 million. Current estimates of undetected cases are in the 80% range, so current total cases is probably around 10 million. So about 20x more people have to get sick before herd immunity. That's well over a year away.

[1] https://www.ft.com/coronavirus-latest

It seems like both the perception of and response to this pandemic are not reflective of reality. What am I getting wrong here?

- Around 200 million Americans will eventually get it, barring extraordinary events (like a vaccine arriving in 6 months). We cannot change this.

- Around 0.5% will die, overwhelmingly those who had a low remaining life expectancy to begin with. This is a much larger body count than most people seem able to accept, but also much less dangerous to the average American than many believe.

- All we can do is alter the timetable to keep the hospitals functioning, which has been done successfully in many places throughout the world albeit at great cost.

There's both good news and terrible news in those facts but none of it really seems to be in the consciousness of the public or of leaders, whereas "oh no you can maybe get it 8 hours after a guy sneezed in a lobby" is everywhere. Along with "everyone who stays at home is a hero" as we set ourselves up for a second Depression.

It’s not 0.5% it’s closer to 2%.

You don’t have a normal economy if people don’t want to go out in public or the supply chain is interrupted as workers get sick. A pandemic will harm the economy no matter what we do, these efforts, it is hoped, will reduce the impact while also saving lives and the hospital system.

And that also assumes everyone can get medical care, if all of the cases are happening at the same time, the mortality rate would increase as resource constraints force medical providers to pick who gets care.

What's your source for 2%? A recent study which was featured on Hacker News estimated the infection fatality rate at 0.66%. [1]

I have seen a number of studies putting the case fatality rate at 2-3%, but this number cannot be generalized to the entire population because current testing is skewed toward the most serious cases. [2]

Long-term mortality in the total population may look something like 0.66% * 0.7 assuming there are no advances in treatment (possible but unlikely, given the unprecedented efforts that are underway, and how naive we were on day one).

In general more recent studies estimating IFR should be more accurate, early ones came with a lot of caveats. If there is a better estimate of IFR than 0.66%, I would like to read it.

[1] https://www.thelancet.com/journals/laninf/article/PIIS1473-3...

[2] https://en.wikipedia.org/wiki/Case_fatality_rate

[3] https://news.ycombinator.com/item?id=22828691

The CFR in Italy is nearly 13%. Even if you make incredibly generous assumptions that there are no unreported deaths and they're only catching 10% of cases that still leaves the infection fatality rate at twice the number you propose

With an older population, a problematic medical system, and being caught totally unprepared, isn't Italy the worst case scenario? I don't think it would be representative of what's going to happen in other countries from here on out.

I don't know where this meme came from that Italy has a "problematic medical system" but that's certainly not true in Lombardy.

The US has a population riddled with comorbitities (obesity, heart disease, diabetes, etc.), a problematic medical system, and was caught totally unprepared. Isn't the US the worst case scenario, with the most deaths and most infections in the world?

>Isn't the US the worst case scenario, with the most deaths and most infections in the world?

Adjust your numbers for population. Per capita numbers for the US are far from the worse case scenario.

You pretty much just described the US.

That's almost certainly untrue. Recent studies have show that as many as 38% of the Italian population has been infected at this point [1]. That gives a fatality rate 0.03%. Let's say the official death count is 10x undercounting, you're still left with 0.3%.

[1] https://www.reddit.com/r/CoronavirusUK/comments/fxgk8q/early...

Testing is severely undercounting cases here. People who just experience fever are told to self-isolate and aren't tested. Not all of these (luckily) end up hospitalized, so those are never, ever counted.

This published scientific article shows you why the CFR in Italy is so high -- and why it doesn't apply to the general population: http://jamanetwork.com/journals/jama/fullarticle/2763667

Takeaways: Old population, Pre-existing conditions, Skewed testing

I use the 0.66% number myself when estimating things, but it's worth noting that that is that papers best-guess estimated infection fatality rate for China, based on its demographics.

The paper has an overall estimate of the infection fatality rate for all ages of 0.2-1.6%.

No, 2% is the case fatality rate. The true overall mortality is far lower. Germany did a randomized test recently of a town and calculated that 2% of the population is actively infected, 14% possess antibodies indicative of a current/prior infection, and overall 15% have been infected at some point. Once this broader testing was completed, the overal mortality was found to only be 0.37%. For comparison, a typical flu season is 0.1%.

https://www.tagesschau.de/regional/nordrheinwestfalen/corona...

For comparison, a typical flu season is 0.1%.

That is incorrect. The worst flu wave to hit Germany in the last 20 years, 2017/2018, killed appoximately 25000 people (excess mortality). That's 0.03%, and it's not typical. Often, the number is much lower.

Note also that the results from the Gangelt study you're referring to are preliminary; there's a press release but no paper yet, not even in preprint.

https://edoc.rki.de/handle/176904/6253

> That's 0.03%, and it's not typical. Often, the number is much lower.

0.03% or 0.1% or 0.66%, the number is low enough to be perceived as not too dangerous by general populace. Contrast that with 13% mortality rate for Italy that was widely spun a month ago. Now _that_ was scary!

> Around 200 million Americans will eventually get it, barring extraordinary events (like a vaccine arriving in 6 months). We cannot change this.

"Significantly less than 200 million Americans catching it" is still on the table.

"Herd immunity" is just the point at which the reproduction rate of the disease falls below 1 because the average infected person does not encounter anyone else to infect (because everyone else has already been infected) before they recover. It's not a binary switch, though -- as the percent of the population that has recovered grows, the reproduction rate shrinks steadily.

This means, for instance, that if we had X infected individuals before the lock downs, and then -- because we can't stay completely locked down forever -- we at some point have X infected individuals again after the lock downs end, we're actually in a much better place, because the reproduction rate of the disease is lower and that X will become 2X much slower than it would have before the lock downs.

This also synergizes with stay-at-home orders, which also reduce the reproduction rate of the disease. Right now we're relying entirely on stay-at-home to reduce the reproduction rate to as close to (or below) 1 as possible, but in a month or two, once a portion of the population has recovered, we'll be able to rely on a mix of "herd immunity" and stay-at-home to achieve the same thing; it's entirely reasonable that we can drag it out for years, and that less than 50 or 100 million Americans will catch it before a vaccine is developed.

It isn't hopeless.

> Around 0.5% will die, overwhelmingly those who had a low remaining life expectancy to begin with. This is a much larger body count than most people seem able to accept, but also much less dangerous to the average American than many believe.

We're currently losing 12-17 years of life per death, based on the statistics out of New York. It's older folks, sure, but it's hardly just taking people off their death beds.

This is a great comment, thanks. So isn't there an argument in favor of infecting more people sooner in order to increase the immunity rate of the population, particularly if those people are at lower risk of complications for the disease?

For example why not pursue a policy such as: kids can go to school unless they live with someone who is at risk (since we know that children are at very low risk for complications). Similar to how we managed chickenpox before there was a vaccine.

I think we need to explore alternatives to lockdowns because I believe the economics of lockdowns will make them impossible to continue for very long. At the moment people see this as a "be heroic and do the right thing" issues, but most people don't yet understand the severity of the economic impact and haven't yet been personally affected by it. Lockdowns will probably become politically untenable by May or June whether anyone today likes that fact or not.

>For example why not pursue a policy such as: kids can go to school unless they live with someone who is at risk (since we know that children are at very low risk for complications). Similar to how we managed chickenpox before there was a vaccine.

Because kids are the most likely to ignore even simple ways of preventing infection. Then you get exponential growth via kids and the health system implodes.

How does this square with Denmark, which is opening schools tomorrow? It's possible they're wrong, but they must at least have some reason to think what you're saying won't happen.

Eh, it's kind of like the decision on how bags of groceries to carry from the car to the house at once -- it's faster the more you take, but the more you carry the more likely you are to drop them and the more catastrophic the accident becomes when it happens.

It's tricky to take large bites here because of the short doubling time compared to the relatively longer recovery time -- in the ~3 weeks it takes your currently-infected cohort to recover, your infected population could increase 100-fold if you aren't careful. Mistiming things -- misjudging how many people are currently infected -- by a couple of days can be difference between exposing 10% of your population to the risk and exposing 15%. There's a reason pandemics are often compared to forest fires.

This will get easier as time goes on, as the doubling period gets longer and longer.

Based on current estimates of how many people will die in the US, you can infer that we currently expect 5-10% of the population to be exposed in this first wave; I believe what widespread testing there has been in Italy leans towards about 10% of that population having been exposed there as well.

Chicken pox was a steady state. It sucked that kids got it, but their parents and most adults were likely immune, so when they brought it home from school, it didn’t spread to parents.

But if you expose kids to COVID now, they will in turn expose their parents, who will expose their coworkers, etc, etc

I could see that being effective strategy if we didn’t have a vaccine in 30 years, but for now, most people are still susceptible.

""isn't there an argument in favor of infecting more people sooner in order to increase the immunity rate of the population, particularly if those people are at lower risk of complications for the disease?""

Only if you would like to see the hospitals overwhelmed with cases that don't result in fatalities but kill lots of other people and also kill people with underlying conditions that would otherwise live full and productive lives.

> Around 200 million Americans will eventually get it, barring extraordinary events (like a vaccine arriving in 6 months). We cannot change this.

We don't know that's inevitable. Korea, Japan and Taiwan suggest another outcome is possible, and while a vaccine may not arrive in 6 months, lots of other things might: better treatments, better masks, better testing, better tracing, etc.

Death isn't the only metric. A lot more than those dying will be the amount of people that are hospitalized in very serious shape. Some of them will have long-term damage, and some of them will die if the hospitals are overwhelmed.

The death rate is 0.5% IF there's really good medical treatment. If there isn't then not only does the death rate from the covid go up but so do death rates for every other condition requiring medical treatment. And not only during the pandemic but for a some time afterwards as the health system recovers. That's not counting those who survive but have some form of permanent damage.

The way to avoid that is to at least spread out the impact over time. Telling people it'll be okay leads to no one listening and the health system collapsing. That's human nature. Telling people it's the apocalypse means most listen and the health system survives. Welcome to humanity.

I'm just very skeptical of this kind of noble lie. If authorities are known to tell people what's convenient for them to believe rather than what's true, is anyone going to end up listening to them in the end?

It is not a noble lie to tell people it won't be ok if it really won't be ok.

Really, if every jurisdiction in the world overwhelms their hospital capacity it won't be ok.

Agreed, it'd be wrong to tell people it's okay. It's not okay, a lot of people are going to die, and even more people would have died if we didn't take costly measures to stop it. This has been and will be the worst experience of a lot of people's lives, and it'd just be a lie in another direction to pretend that's not so.

What seems to be a noble lie is the apocalyptic mindset, where the coronavirus is literally the only thing that matters and we must never ask if a particular mitigation is worth the cost. Many authority figures are promoting this idea, even though they clearly don't believe it themselves and couldn't formulate effective policy if they did.

> even more people would have died if we didn't take costly measures to stop it.

The example of Sweden seems to disagree with this assumption. They never locked down, and the current mortality is (1,203 / 10,330,000) * 100 = 0.0116%

First case on Jan 31st, no lockdowns, death rate has already flattened. Where is that crazy scary exponential growth?

Sweden as the great success story when it is suffering the same economic devastation as its neighbors but a higher death rate is an interesting argument.

The thing about the lockdowns is that evidence so far indicates that stopping 80% of non-essential economic activities voluntarily is about as bad for the economy as stopping 90+% on a mandatory basis, but it seems that the health outcomes are much better in the latter situation.

> The death rate is 0.5%

You can't know that without testing the whole population, or at least getting some controls.

They did random anti-body testing in a small town if I remember correctly and found it to be 0.5% or so.

Based on Italian serological and all cause mortality data I have seen, the death toll would more likely be about 1.4%. 0.5 IMO is quite optimistic. 0.1% of New York's population has already died.

This also discounts the unknown possibility of disability, reinfection, and limited time immunity.

I was under the impression that NYC had a confirmed infection rate of ~1%. Have they really had a tenth of confirmed cases die already?

No. Currently >200k confirmed, <11k deaths in New York State.

NY population is 18.8M. 0.05% of the population has died.

New York City's population is around 8.5 M, and more then 10,000 people there have died from from the virus so far (https://www.nytimes.com/2020/04/14/nyregion/new-york-coronav...). That puts the toll over 0.1 %.

I cited the correct, current numbers for the entire state.

You’re citing numbers released by the city where they haven’t confirmed cases, so it doesn’t answer the OP’s question.

Can you cite them, rather than saying that you cited them?

Just in support of your argument, new data is showing that the fatality rate is lower then previously estimated, due to a large number of undetected cases.

Just one example, tests of pregnant women coming into labor in New York show that 1/6 of them are positive for nCov, but 85% are asymptomatic or presymptomatic. (https://twitter.com/CT_Bergstrom/status/1249836446569181186)

Another example, Denmark is finding potentially as high as 30-80x undercounting of cases, which would bring the fatality rate down to 0.1-0.2% (like a bad flu season) [https://www.reddit.com/r/COVID19/comments/fxk917/covid19_in_...]

I've seen a few more of these data points coming out, and nothing is 100% conclusive, all the new data points are showing that it's less bad than mainstream thinking suggests.

> Around 200 million Americans will eventually get it, barring extraordinary events (like a vaccine arriving in 6 months). We cannot change this.

This doesn't seem right to me. With social distancing now, and testing and tracing once we have enough tests available, it seems to me that we should be able to keep the total number of infected people well below 200M, even if a vaccine takes the expected 18-24 months. If there's a solid argument that this is not the case, I'd like to hear it.

1) SARS-2 is highly contagious.

2) Testing and tracing is working fairly well in S. Korea because they're treating it like a war, and they've been orgainzed since fairly early on. The US is not organized.

What most people, including HN, miss is that passing a test today means nothing about tomorrow.

I discount any comments that argue IFR as the sole basis for how quickly we can let the disease spread to save jobs on a short term basis.

Hospitalization is a more important factor as it the death count goes up dramatically without intensive care.

These comments always come with a healthy dose of "people aren't seeing reality!?!?", which is a bonus given the level speculation surrounding IFR and the variance therein.

The "curve" in "flatten the curve" is an epidemic curve, not a Bell curve.

https://en.wikipedia.org/wiki/Epidemic_curve

Fair. But the curves on the document referenced above are not epidemic curves. They’re unbounded exponential growth curves. My criticism of the report stands.

Yeah it was a great post. Just adding that note.

The exponential curves are somewhat useful early in an outbreak. It's confusing right now because some places are still "early" and some places are way past the point where those curves are descriptive. So they won't describe NYC at all but they're still somewhat working nationally.

And "bell" in "bell curve" shouldn't be capitalized -- it's not someone's name, it's just that the curve has (somewhat) the shape of a bell.

> Current known US total cases are nearing 2 million.

It's 2 million worldwide. The US is now above 600,000:

https://coronavirus.jhu.edu/map.html

It won't follow a 'bell curve' but the curve will 'look like a bell curve'.

You can approximate one with the other; it used to be common when normals were harder to work with early computers.

Eg, https://www.johndcook.com/blog/2010/05/18/normal-approximati...

the dotted line is what we would expect if the intervention is removed as of today. the red dots are actual hospitalizations as of right now. the blue line was the predicted line for hospitalizations... she mentions that we are doing better than the prediction as of right now. it was not altogether that clear but in the context of the press briefing it's more clear.

This is poorly designed graph because it can give the wrong impression when it isn't accompanied with the explanation that was given in the press conference. I am not usually one to make a "this should be higher" comment on HN, but this is the proper interpretation and lots of people are being misled by not seeing it.

> The graph on slide two[1] of the official CA govt plan shows hospitalizations With Intervention rising exponentially

It does not. The exponential curve, per the graph legend, is what would be expected without intervention.

The light blue curve, labeled "With Intervention", is an upward-facing concave curve. Even if it were linear, it does not align with the observed data referenced in the link labeled [2].

OK, looking at the new cases curve on your covid-19.direct page. As presented, you can't really see what the green curve is doing. If you remove the black, total curve, the green new case curve is visible but it seems to be going up and down fairly inconclusively. Until 4/1, it looks like exponential growth. After that, it goes down and then up. Sure, the optimistic view is it's plateaued or going downward and I hope that's true but counting on just twelve days of data seems foolish.

"If you remove the black, total curve, the green new case curve is visible but it seems to be going up and down fairly inconclusively."

The best fit line of new cases after 4/4 is a linear function with a definitively negative slope. It is conclusively not a function that contributes to the upward, concave blue line on the state's graph.

Who cares what the best fit is. There is simply not enough data. You're talking twelves data points with considerable variation. Using statistic doesn't turn muddy and ambiguous data into a clear signal magically.

My point is that there is no extrapolation of the data we’ve seen to date that yields the blue line on their graph.

Sure, but I think they are tracing the worst case scenario based on the underlying assumption that the virus increases exponentially.

What going to happen is indeed inherently uncertain so we pretty much have to prepare for the worst plausible series of events. Once it becomes clear that a better outcome is fairly likely, then the plans can shift.

Can you trust that data to be perfectly comparable? Testing capacity ramped up at a different rate than cases. Testing guidance has been spotty too. The days right after increases in testing access could be very noisy.

I would look to the deaths curve much before the new cases curve, even though it's less responsive to changes. I assume fewer deaths have slipped through the cracks (though I'm sure some have).

What bothers me the most is that the comparison is always against a hypothetical "do nothing" strategy, ignoring the fact that different countries have employed different strategies, some of them have been highly successful in containing the disease without any lock down (Korea, Taiwan, and Australia) come to mind.

Aren't the techniques that Taiwan and Korea have employed (testing, isolation of known cases, contact tracing) irrelevant at this stage of contagion? I doubt of the efficacy of contact tracing when you have enough of a percentage of population infected that a single visit to Costco or any other supermarket would mean a very high chance of being there at the same time as an infectious person...

I think that the answer is that we don't know what's the effect of each measure. The fact that Japan & Australia did not employ contact tracing, but also managed to contain the spread shows that there are other factors at play. It might be face masks use, or temperature (for Australia), but it does show that there are potential alternatives that could be viable.

"some of them have been highly successful in containing the disease without any lock down (Korea, Taiwan, and Australia) come to mind."

How are you defining lock down here? I live in Australia and most people I know have been working from home for weeks and have had kids out of school during that time. All restaurants have been closed for a while now. There are stops/checks at border entry points between states, even at points 1-2 hours in from the border, enforced quarantine for anyone arriving interstate. Weddings restricted to 5 people, funerals to 10. People are getting fined $1k+ for non-essential travel (group of kids this week copped big fines for getting up and congregating to see the sunrise).

At almost every "essential" business I've visited, there have been spacing markers on the floor. Supermarkets have plastic shields between the checkout operator and shopper. Gloves, hand-washing, etc at any food pick-up place I've seen.

I don't favour our current leaders at either state or national level, but since Morrison's Hillsong weekend, he's barely put a foot wrong. Senior politicians have been serious and escalated restrictions as appropriate (ignoring Ruby Princess). I'm sure there are more serious levels of lock down, but it's been part of the response in Australia.

I wish this was the situation here. The limitations in most of Europe and united states are much much harsher, including no schools and mandatory closure of all non-essential business.

Most of Europe hasn't closed all non-essential businesses. Has anyone, aside from Italy and Spain? Non-essential public venues (gyms, museums, some stores, etc) are closed, but that's true in Australia, too.

Schools have closed in one Australian state at least. More than half of children are at home in my home state, from what I can tell, where schools are not closed. They've polled for the coming term and suggest that two-thirds will be at home.

One serious loss in productivity comes in overseeing homeschooling while trying to work from home. Can't imagine many non-essential businesses are running anywhere near 100%.

That makes you think it should be a bell curve? Are you arguing it will be a bell curve with such a peak, even if restrictions are weakened or removed? That is unlikely.

I am not. I am arguing that the current trajectory data we're observing [2] does not align with the projections of our current trajectory provided by the government of California (the light blue line labeled 'With Intervention' [1])

[1] https://www.gov.ca.gov/wp-content/uploads/2020/04/California.... [2] https://covid-19.direct/state/CA

But you wrote "meaning that the graph of active hospitalizations should be a bell curve with a maximum value of ~10,000 on ~4/20."? You didn't mean that?

I meant that "the graph of active hospitalizations should be a bell curve with a maximum value of ~10,000 on ~4/20" with the current mitigations in place. That is not what the government report is projecting. They're projecting unbounded, exponential growth with current mitigations in place, even though that is not in line with the data we've seen over the past two weeks.

Also the hospital bed line that many draw are all based upon having medical staff for them. So does somewhat depend upon no staff shortages due to illness impacting capacity of hospital bed numbers.

This plan also ignores the question of the majority of people who have likely been infected with no symptoms that could be identified with widespread Serology/Antibody tests. I find it alarming that this plan ignores that, is the governor priming is for population wide vaccination that is likely to be unnecessary? If we confirm widespread recovery and immunity then larger numbers of people can get back to life as usual quite soon.

The plan ignores that because people working on this are consistently finding infection rates of less than a few percent. Stanford tested 2888 samples that were negative for flu and other respiratory viruses and didn't find any positive for SAR-COV2 until Feb 21st and Feb23.

A source for what you are saying is the linked SF article [1] which indicates that of ~3k evidently sick people with flu symptoms, 0.1-1% were confirmed to have COVID-19. Though note this is not random population sampling, on an interesting scale, so ignores asymptomatic question. Worldwide analysis of death rates [2], indicates #infected-so-far = 700-1600 x #dead so far. And data from Iceland [3], the only country with widespread sampling of asymptomatic population, is aligned with these numbers. Based on the analysis and support from Iceland, Covid-19 testing to see if patients actively have the disease are likely inaccurate and also a policy of only permitting testing of severe symptom patients is depressing reported numbers of once infected.

[1] https://www.sfgate.com/coronavirus/article/Think-you-got-the...

[2] Analysis of #infected today from #dead today, using following assumptions:

1. death statistics are reliable, since cause of death tests can adequately collect fluid samples (a major source of inaccuracy in COVID-19 testing), source = talking to medical staff at Bay Area sites.

2. E[days between first infection and death] = 14 (lost the source at the moment), Therefore, number of dead today D => number of infected 14 days ago = 100 D

3. rate of death increase consistently trends around 25%/day or 15% with full lockdown, source CDC situation reports worldwide excluding regions with restricted access to journalists

4. therefore, #infected today = #infected 14 days ago growing at 25% daily = 22X, or 7X if your region was locked down for last 2 weeks

5. so #infected=2200D or 700D

[3] Last week, CDC reported 2 deaths today for Iceland, and admittedly-non-random sampling by DeCode indicated upto 1% of population have/had virus, upto 3600. Iceland has done strict quarantining of confirmed and suspected cases, and extensive testing since February, so despite no lockdown, I would put them at the 700X end of the estimate. They report ~1000 confirmed cases, so they would actually estimate 3600-1000 infected so far. My estimate would have indicated 2200-700 X D (= today's deaths, 2) = 4400 to 1400 cases.

San Miguel county in Colorado has found 0.5% of their population has confirmed antibodies after doing 1600+ tests (out of 8,000 people in the county).

As another data point, the theory of this thing being widespread in the population seems far fetched to me.

Well, Gangelt showed 15% had antibodies (subject to, potentially fair criticisms re: methodology) so I think it's fair to say it varies by locale.

While I understand that many folks from Wuhan travel all over the United States, the non-stop options from Wuhan to NYC and, if I recall correctly, SFO and LAX meant that more opportunities existed for early spread than the middle of Colorado. In fact, I can't actually pinpoint a city in San Miguel County, Colorado -- can you? (https://goo.gl/maps/F7CJ3yfDKaf1gQjDA) I guess there's a post office in Egnar... Telluride maybe?

Certainly the sparse population created natural social distancing precluding its spread.

San Miguel County is home to Telluride ski resort, an international destination with about half a million visits per year [1]

[1] https://www.telluridenews.com/news/article_1547199f-46bc-547...

True, population 2484.

The strain that is widespread in New York is the European strain: https://www.nytimes.com/2020/04/08/science/new-york-coronavi...

Very cool! I wonder if that's true in the Bay Area too.

I like how a factual point that goes against HN wishful thinking gets downvoted.

There could be some limitations in sampling. Multiplex respiratory panels at Stanford that come up negative for everything... what does that patient look like?

14% of women giving birth in NYC test positive; 86% asymptomatic. https://www.nejm.org/doi/full/10.1056/NEJMc2009316

That's what a 15% infection rate looks like. New York City right now.

And you guys want to take it to 85%.

No one is priming for vaccination because a vaccine doesn't exist yet and we don't even have a reliable schedule estimate for one.

Reply on news.ycombinator.com