A few thoughts:<p>1) Extrapolating from the infection rates of very specific groups (for example, evacuees) to the entire population without taking into account transmission dynamics and the time between infection and detection does not make very much sense. The authors naively multiply the infection rate among evacuees by the population of Wuhan to conclude that Wuhan must have had 178,000 infections at the end of January. By comparison, epidemiological models have estimated there were around ~20,000 infections at that time [1][2][3]. What conclusion should we draw here? If you use sloppy, back of the napkin math to over-inflate the infection count by 10x then you can correspondingly deflate the mortality rate?<p>2) Speculating about the mortality rate of Covid-19 based on several gigantic assumptions ("If our surmise of six million cases is accurate, that’s a mortality rate of 0.01%") seems borderline irresponsible. Numerous researchers have been modeling this virus and have generally arrived at numbers in the range of 0.5% to 1.6% [4][5][6][7][8][9]. The authors don't present any compelling reason why we should doubt those numbers.<p>3) Ultimately the mortality rate is not as important a number as the hospitalization rate. The authors would have you believe this virus is no worse than the flu, but this is not congruent with the number of reports coming out of places like Italy and New York saying they're about to run out of ICU beds, or China rushing to build temporary hospitals to house all of the patients that need critical care. What the mortality rate might be under ideal circumstances where every patient receives adequate medical care might be significantly different compared to a scenario where you've run out of ICU beds and have to start rationing ventilators.<p>[1] <a href="https://www.mdpi.com/2077-0383/9/2/419/htm" rel="nofollow">https://www.mdpi.com/2077-0383/9/2/419/htm</a><p>[2] <a href="https://www.medrxiv.org/content/10.1101/2020.01.23.20018549v2.full.pdf" rel="nofollow">https://www.medrxiv.org/content/10.1101/2020.01.23.20018549v...</a><p>[3] <a href="https://www.mdpi.com/2077-0383/9/2/523/htm" rel="nofollow">https://www.mdpi.com/2077-0383/9/2/523/htm</a><p>[4] <a href="https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-severity-10-02-2020.pdf" rel="nofollow">https://www.imperial.ac.uk/media/imperial-college/medicine/s...</a><p>[5] <a href="https://institutefordiseasemodeling.github.io/nCoV-public/analyses/first_adjusted_mortality_estimates_and_risk_assessment/2019-nCoV-preliminary_age_and_time_adjusted_mortality_rates_and_pandemic_risk_assessment.html" rel="nofollow">https://institutefordiseasemodeling.github.io/nCoV-public/an...</a><p>[6] <a href="https://cmmid.github.io/topics/covid19/severity/diamond_cruise_cfr_estimates.html" rel="nofollow">https://cmmid.github.io/topics/covid19/severity/diamond_crui...</a><p>[7] <a href="https://www.medrxiv.org/content/10.1101/2020.03.04.20031104v1.full.pdf" rel="nofollow">https://www.medrxiv.org/content/10.1101/2020.03.04.20031104v...</a><p>[8] <a href="https://www.medrxiv.org/content/10.1101/2020.03.09.20033357v1.full.pdf" rel="nofollow">https://www.medrxiv.org/content/10.1101/2020.03.09.20033357v...</a><p>[9] <a href="https://www.nature.com/articles/s41591-020-0822-7" rel="nofollow">https://www.nature.com/articles/s41591-020-0822-7</a>
They have a good point about missing data, but they're glossing over the fact the fatality rate goes up drastically if the medical system fails.<p>I 100% agree that we should be running antibody tests to see how many people have caught COVID-19 and recovered. All of our testing until recently was based on RNA tests, which may only be positive during a relatively short window. (I saw a case study claiming that even some hospitalized patients are testing negative inside 7 days.) We <i>need</i> to know how many people catch this and beat it quickly.<p>That said, even in populations that have been RNA-tested early and extensively (such as the Diamond Princess and South Korea), the number of completely asymptomatic cases is less than 50%. Using the most optimistic data, I personally have less than a 0.5% chance of dying <i>assuming I get all the medical care I need.</i><p>And that's the problem. This virus hospitalizes about 20% of identified cases. They require some supplemental oxygen, and maybe an IV. With good care, probably less than 1% die.<p>So there are really <i>two</i> key fatality rates:<p>1. How many people die if they get all the care they need, and<p>2. How many people die if 30+% of the population catches this at the same time?<p>Even if we're overestimating (1) by a factor of 10, that's still enough to make (2) catastrophic. What happened in Wuhan and Lombardy can happen here, and there's absolutely no reason that it couldn't get 10x worse. Even if we're overestimating the disease.<p>So let's start testing aggressively for antibodies. Until we get that number, I'm all for extreme caution.
> So if 100 million Americans ultimately get the disease, two million to four million could die.<p>Seems to miss any mention that people who get sick - but <i>don't</i> die - seem to be having pretty severe (sometimes permanent looking) damage.<p>Aka, they're only counting "deaths", when they should also be including other very serious negative consequences too. :(
> the real fatality rate could in fact be closer to 0.06%<p>If you take a look at Diamond Princess.<p>~ 4000 on board<p>* 712 cases<p>* 10 deaths<p>Gives a case fatality rate of 1.4%.<p>The average age onboard was ~60 years (don't know if it includes staff) which is definitely higher than average.
However, the two numbers (1.4% and 0.06%) sounds very far apart.
FTA "But a 20,000- or 40,000-death epidemic is a far less severe problem than one that kills two million."<p>Meanwhile, <a href="https://www.worldometers.info/coronavirus/" rel="nofollow">https://www.worldometers.info/coronavirus/</a> shows Deaths: 19,603<p>I'm not an epidemiologist but how can we barely at the mid-point of the spread and say things like the above, while the hard facts already show we're at 20k deaths?<p>While I agree the actual rates are hard to know until we have robust antibody assay in larger populations, it seems a bit hand-wavy to say "oh it's just not that bad, here look at these random samples we looked at it and it'll be ok."
Related pieces by John PA Ioannidis:<p>- Coronavirus disease 2019: the harms of exaggerated information and non‐evidence‐based measures[1]<p>- A fiasco in the making? As the coronavirus pandemic takes hold, we are making decisions without reliable data[2]<p>There are two ways of getting the answer to the question: get relevant data to make reliable predictions (e.g. test for prevalence in the population) or run the experiment (which some governments seem prepared to do)<p>1- <a href="https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13222" rel="nofollow">https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13222</a><p>2- <a href="https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/" rel="nofollow">https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-a...</a>
Everything feels "deadly" in the hyper connected world.<p>The hyperconnected world is like a baby's brain which is more connected than an adult. As learning happens connections are culled [1]. To the baby every new piece of info is mesmerizing or frightening. The brain hasn't yet understood how to process things, what to filter out, what to focus on etc<p>With this new networked world/hive mind, reactions to every new event are much like the reactions of a Baby as it blunders about discovering a new world.<p><a href="https://www.edge.org/conversation/alison_gopnik-a-separate-kind-of-intelligence" rel="nofollow">https://www.edge.org/conversation/alison_gopnik-a-separate-k...</a>
Knowing the fatality rate requires knowing the number of people infected: we all know the figures we currently have are mostly a low estimate for a number of reasons already discussed at length.<p>We must also know the number of fatalities, and this will only be known <i>after</i> the epidemic once we calculate the over-mortality compared to a normal year. So the current numbers are a low estimate too.<p>In the end we can tell nothing about the fatality rate which is the <i>ratio</i> of those values.<p>Hence governments make a guess and act accordingly. Only those who test large scale and isolate only positive people act rationally.
> <i>Elective procedures will need to be rescheduled. Hospital resources will need to be reallocated to care for critically ill patients. Triage will need to improve.</i><p>If we'd let this thing run its course, wouldn't almost all corona patients have to be refused in hospitals and instead die at home? Normal "triage" doesn't cut it. You'd basically paralyze normal health care for months, which would come with an additional death toll. This is already happening even with the lockdowns.
Something I'm wondering about all this, if you look at the current and unprecedented almost worldwide policy of self isolation and distancing through the prism of unintended consequences, cobra effect, natural selection, and Spanish flu with deadly second wave (and corresponding deadlier strain), I'm wondering: aren't we creating a perfect ground to selecting the most durable and transmissible strain of the sars-cov2?
Spain just recorded 738 deaths in a single day. We will certain know within a week of monitoring Spain and Italy as to who was right and who was wrong. It isn't something I would gamble on personally. I also don't see the act of a complete shut down as being more economically-damaging then a drawn out partial shutdown.<p>Also, the other side of the equation is said to be perfect, but as was shown the other day, a lot of elderly people are left for dead in their homes or retirement villages and not reported. Also, I don't know if all deaths in hospitals are checked for coronavirus.
Both the authors of this piece are not infectious disease epidemiologists. One (<a href="https://profiles.stanford.edu/eran-bendavid" rel="nofollow">https://profiles.stanford.edu/eran-bendavid</a>) is an infectious disease doctor who focuses on the effects of economics, politics, and the environment on health.<p>The second (<a href="https://healthpolicy.fsi.stanford.edu/people/jay_bhattacharya" rel="nofollow">https://healthpolicy.fsi.stanford.edu/people/jay_bhattachary...</a>) focuses on "the constraints that vulnerable populations face in making decisions that affect their health status, as well as the effects of government policies and programs designed to benefit vulnerable populations. "<p>This is not the time for folks to hit the pages of the WSJ to start developing their viral infectious disease modeling muscles. Imagine if you asked a compiler writer to start developing a commercial OS.<p>These folks fundamentally do not understand viral growth models. Their article demonstrates their lack of understanding and they keep falling back to the normal models (pun intended) of epidemiology based on the usual statistical machinery. This is a common problem in this analysis. I made a more detailed comment further down that demonstrates the pitfalls of this thinking.<p>(edits: grammar)
Hello, just an info, in my hometown, Bucharest, population approx 3mil,they just announced they will test everyone. Some good data will come out if this.
The conclusion of the article isn’t controversial but the logic is incomplete.<p>Of course policy-makers should continuously evaluate whether or not shutting down a whole economy is worth it. Definitely we need better testing to understand the true number of infections.<p>But you can’t completely disregard current human responses when a) the condition can be fatal (game over, there’s no retry) and b) the virus can leave behind long-lasting damage.<p>In this case a purely quantitative argument based on fatality statistics feels myopic.<p>For sure agreed we need to find ways of gathering more and accurate information.
Re-posting the most informative visualization I have found on this topic so far: <a href="https://www.politico.com/interactives/2020/coronavirus-testing-by-state-chart-of-new-cases/" rel="nofollow">https://www.politico.com/interactives/2020/coronavirus-testi...</a><p>See dramatic ramp-up in testing, and where the number of tests administered is significant, a reasonable approximation of the infection rate can be established. So the recent blow-out in NY was actually a good thing and it does not reflect the true daily infection rate: they just didn't know how many cases they had. They still don't in fact, asymptomatic cases don't get tests, and those who had the virus and now have immunity (of which Cuomo's advisors suspect there's at least 100K) can't be tested without the new serological test currently under development.<p>The only reliable metric of severity remains the number of deaths, and until that gets into tens of thousands (i.e. exceeds that of flu), any panic is premature. We're not Italy. We're not Spain. The current level of response to this is unprecedented.
Somethng else I wonder about all this: Let's look at Germany and say we have about 10000 new cases per day with the current measures, including unknown infections and some handwaving. That's a lot of days until you reach a significant portion of the polulation (~80 million).
There's insufficient data to say whether the death rate is actually low or not. But this will be interesting to see play out politically, if the Republicans agree to open up and there is mass deaths as a result this could sink the party for decades.
Case fatality rate (CFR) is meaningless for the general public. A case is basically any person that went to the doctor and received the permission to get tested and eventually became a patient. Knowing the CFR for a disease is therefore very valuable for doctors because they know that x% of their patients may die. However, it is very possible that the vast majority of Covid-19 cases never become patients.<p>This doesn't mean that the quarantine is pointless. Even if covid19 is as "harmless" as influenza, having two influenza type diseases is still worse than only having one.
Q: how accurate are the covid19 reported deaths? My guess is that they measure immune cells from blood (cytokine storm), but they can't know for sure if covid19 was they cause of death. It means we have an upper bound of deaths.
> We should undertake immediate steps to evaluate the empirical basis of the current lockdowns.<p>I think there's not enough data available to evaluate if these lockdowns are justified. Especially the cost of the lockdowns are hard to assess. We won't know until it's over.<p>But even if the mortality rate is vastly over-estimated, it will be hard for a leader to justify that their country has twice the mortality rate as the neighbor country (even this rate is low, and even it a lockdown was avoided).<p>Besides, most countries are imposing a lockdown anyway, so "do like everybody else" is a safe bet for our leaders. Plus, the population asks for it.
There are people who build and grok models (A) and people who build epidemiological models (B) and then people who built epidemiological models of infectious diseases (C).<p>It is a common error for folks in set A to believe that they can understand models built by folks in set C. It is a subtle but serious error for folks in set B to believe they can understand models by set C e.g. John Katz of Yale who is a doctor focusing on diets and wrote a woefully misguided article in the NY Times.<p>Roughly, I find the confounding factors about models about the current epidemic seem to the following:<p>1- 1% of the population severely affected etc. (see: Diamond Princess). What they fail to appreciate is that the Diamond Princess was an enclosed environment and the outbreak was controlled and <i>limited</i> to the population on board. In the general population, a virus with an R0 of 3 will infect 50,000+ people in 10 infection steps.<p>2- What most people miss is the collision of the 1% severity with the capacity of the medical system. This second-order effect is something hard to understand. 1% of the US is 3.6 million. Even if 10% of those cases turn up at the ER in the same year and occupy the beds for 2 weeks each it will be a disaster.<p>3- #1 and #2 interacts with life as usual demands on the health care system – accidents, heart attacks, strokes, etc. to create third order effects – more deaths as there are no beds and no personnel to deal with them.<p>4- PPE running short and causing infections amongst medical personnel leading to their quarantine, hospitalization or death (Wuhan, Italy, Spain) decimates their ranks and accentuates the stress on the health care system……. 20% of lost capacity translates to some fraction more deaths and more stress on the rest of the medical population.<p>I am sure there are factors I have missed.<p>We have seen many contrarian viewpoints. None of these contrarian thinkers make any concrete suggestions (Ionnadis, Katz, Friedman, Gillespie, Hanson) except making ominous predictions to how wrong we are and we need more data. Meanwhile ER doctors say this is the worst they have ever seen and bodies keep piling up. These contrarian thinkers provide no simulations of how saving the economy will lead to lesser deaths. Just hand waving and finger jabbing.<p>The contrarians seem to think that the playbook to deal with the epidemic has been improvised. Or that this is something modelers are thinking up on the fly. These playbooks have existed for decades with very good understanding of the dynamics and were used during SARs, Ebola and heck even back during the Spanish Flu. What seems to be lacking was parameters specific to Covid-19. Its R0, CFR, co-morbidities, etc.<p>The reason Taiwan, Korea and Singapore were able to act so decisively and fast was that they just dusted of the SARs playbook and knew almost exactly what to do. Taiwan enforced the first measures in early Jan.<p>I will take the word of the viral epidemiologists over the contrarian's armchair speculations and continue to overreact. I recommend you should too.
for the last time it’s not the death toll it’s the <i>massive</i> numbers of immunologically naive people of all ages who need advanced medical care all at once.
One of the biggest reasons the rate may be too high is the method of attribution. For example, in Italy, post-mortem testing of people who were origially presumed dead of other illnesses accounted for a large portion of testing. When it turned out they had COVID-19 at the time of death, the <i>cause</i> was attributed to COVID-19 regardless of comorbidities: <a href="https://www.bloomberg.com/news/articles/2020-03-18/99-of-those-who-died-from-virus-had-other-illness-italy-says" rel="nofollow">https://www.bloomberg.com/news/articles/2020-03-18/99-of-tho...</a><p>In other words, many of those who died today <i>would have died</i> within the year. Many were already in and out of hospitals for serious diseases.<p>So, not only is the IFR wildly unknown without serology tests. The CFR itself is also severely biased.<p>Currently in the U.S. you cannot get tested unless you have severe, life-threatening illnesses.<p>The fact that the U.S. has such a low CFR speaks very well of healthcare in the U.S.... considering the majority of those tested have severe symptoms.
It has been horribly fascinating watching the (coordinated?) campaign to downplay either the fatality rate or just write off a percentage off the population dying as the "cost of doing business".
at this point who cares? we know that older people are at risk and do we really want to risk our parents/grandparents/neighbours health? it's not worth it. isolate yourself, clean your hands and face, and help fellow humans.
TLDR; the article argues that the actual fatality rate is possibly 0.06% as opposed to 1-6% touted in media.<p>The reasoning is based on estimating actual cases given confirmed cases. For example, in Italian town Vò, the entire population was tested to find a prevalence rate of 2.7%. Apply this to the whole province to estimate actual cases and then divide that by confirmed deaths. So assuming that unconfirmed cases mostly recovered without an event, the actual fatality rate goes down to 0.06%.<p>Arguably, the author doesn't have a lot of other strong data to back this up. Also, this would imply that a large part of the infected population simply recovered without needing to possibly sick treatment.<p>If this is true, however, it would mean we just had a 2 trillion dollar party :).
tl;dr The case fatality rate is usually much higher than the real mortality rate.<p>People and health professionals have pointed that out from the start and it should be hard to find anyone who doesn't know that by now.<p>> <i>A universal quarantine may not be worth the costs it imposes on the economy, community and individual mental and physical health.</i><p>Sure, you can use expected utility theory to calculate that. What result you get depends a lot on the monetary value you attribute to a human live, or, if you think that's better, on the value of quality-adjusted life years (QUALYs).<p>If you're particularly sociopathic you can also reverse-engineer the simple models used in order to find the value judgments that will give you the decision result you want. Just tweak the values for QUALYs, lower the estimates here and there, and the result is that it's better to save the economy (or vice versa, depending on what you want). Call me cynic, but I'm sure plenty of people around the world are doing that right now.
Is there any explanation as to why this thread is "flagged"? It seems like the original article and the discussion here are of good quality even if some of the ultimate conclusions maybe flawed.
Let's prove it. Go, go, go! Let's beat Iceland! We can do it. <a href="https://www.cloverpop.com/blog/we-need-a-covid-19-testing-moonshot" rel="nofollow">https://www.cloverpop.com/blog/we-need-a-covid-19-testing-mo...</a>
The WSJ has a history of climate change denial too. They are 100% not a publication that follows science at all.<p>You may also not be surprised to know they're owned by News Corp.
if its really so low then why are they scrambling to pass a 2T bailout package. just let it rip through the population because 'its just a flu bro' and it will be over in a month and a half.<p>/s<p>as I see it this is BS propaganda from WSJ to normalize whats coming next, push to open businesses.
There are some credible voices skeptical of the current doomsday predictions from Dr. Fauci et al, vis-a-vis the national quarantine.<p>Unfortunately, this too has become a partisan political issue with dems for shutdown/quarantine, and republicans for opening up.