Hacker News Reader

Comment by sacred_numbers | Hacker News Reader

Because the death rate is way higher than 0.26%. If the CDC really thinks it's that low for the general population then they are wrong. The general consensus I have seen, based on antibody tests in New York, Italy, Sweden, and Spain point to a death rate between 0.75% and 1.25% for most developed nations. For less developed nations (which, to be fair, is most of the world) the death rate will be lower because there are fewer old people. I agree that we don't have to shut down everything, but everyone should be wearing masks as much as possible in public. Public smoking is highly restricted, at least partially because of the risks of secondhand smoke, and the risks of COVID-19 are way higher than that. People can do what they want at home, but we have to share public spaces. It's not right for someone else to make decisions that endanger me and my family just because they don't care about the risk.

View on HN

Comparing deaths from previous years for the same month(s) (of the pandemic), along with the expected number of deaths (projected) for 2020, is how this sort of calculation is made, on a preliminary basis.

There is an (older) Financial Times article that did an excellent analysis on this (free to read): https://www.ft.com/content/a26fbf7e-48f8-11ea-aeb3-955839e06...

But, this article (also free to read) came out today from the Financial Times and is way better. The UK is doing far worse than the US interestingly: https://www.ft.com/content/6b4c784e-c259-4ca4-9a82-648ffde71...

Countries like France, Spain, and Italy (Mediterranean countries) in particular rank badly (besides having an uncontrollable infection problem) for the calculation of total excess deaths (compared to previous years) because they have some of the highest average lifespans in the world, largest percentages of elderly populations, and have extremely high quality universal healthcare (for example, the third leading cause of death in the US is actually preventable medical errors, so we really do not have phenomenal healthcare, even if you have “great insurance” and “can pay for it”).

The CDC latest best estimate is 0.4% death rate from Covid. This compares to a death rate from the flu of around 0.2%, also based on CDC estimates [1].

Every death is a tragedy, but we didn't do half-lockdowns for the flu before. Why are we doing a full lockdown for something that is 2x the flu.

Antibody testing largely supports a 0.4% death rates conclusion, I've tracked this data in a spreadsheet I made [2]. 0.4% is close to the mean and median of the studies I recorded.

Edit: I know it's gauche to complain about downvotes, but this post is just citing data to make a point. Even if you are worried about Covid, it seems dishonest to downvote data.

[1] https://www.wcnc.com/article/news/health/coronavirus/data-cd...

[2] https://docs.google.com/spreadsheets/d/16onEUBWIV5IqN1RCvTla...

I genuinely mean this with no disrespect but how knowledgeable are you about this & if you aren't why do you think some data collecting from the internet & "back of the envelope" spreadsheet-ing makes you able to reasonably question world experts that dedicated their careers to virology and pandemics?

If you are a programmer how much would you appreciate a medical doctor with no training or experience in software development questioning your technical decisions about what to use when building a software service?

The 0.4% fatality estimate is from the CDC, presumably full of credentialed experts. My spreadsheet is a collection of studies from experts who have presumably dedicated their careers to virology and pandemics. The 0.2% number is also from the CDC. I’m not out there doing science, I’m just looking at the data.

I’m not sure what the basis is if your criticism. If a medical doctor wants to submit a PR I would review it based on merits.

It looks like your spreadsheet divides deaths on a given date by (confirmed and estimated) cases on that same date? This will bias your estimate downward. There is a significant lag between diagnosis and death (and even more so between infection and death -- which of these is relevant depends on how you are estimating unreported cases), but you're implicitly assuming that 100% of unresolved cases will end in recovery.

You are correct, deaths lag a bit, as does reporting of the deaths, which downward biases the IFR. Antibodies also lag before they show up in the serology tests, which upward biases the IFR.

But broadly, the studies I've cited largely back up the CDC's estimate of 0.4% IFR. Some of them have methodology in the papers themselves which give similar numbers, and I encourage you to read the raw data.

But an IFR close to the flu is not at all surprising. Covid is a respiratory infection. It kills the same way the flu kills (eg. pneumonia), so you'd naively expect them to have similar fatality rates and outcomes, and they do.

Probably because it's hard to trust public health experts in a crisis whose very existence has already demonstrated a failure in their authority.

If I wanted to reduce the risk of acquiring or transmitting an airborne infection back in February, I would wear a mask and avoid enclosed spaces without waiting for CDC's permission. Indeed, I would ignore their advice to the contrary, which stood until April (!)

It would be really nice if we could defer to competent technocratic experts, but they earn their trust by having credible track records, not by collecting credentials.

As far as I can see the main screw up was the American and British governments ignoring expert advice. I live in Germany which happens to have a responsible adult in charge & this country fared great in this crisis so far.

No thanks. The Hippocratic oath is meaningless unless the subject decides what constutites a medical procedure.

The train to technocratic ditatorship is paved with... the great glucose poisioning epidemic, loss of free speech, loss of self defense. Todays conventional wisdom is often not looked on favorably in the future.

http://www.youtube.com/watch?v=TRnB9El_V5g

Are we going to wear masks for the rest of our lives (outside FFS)? Pandemics are an authortarian dream. The ability to engineer pandemics will only expand.

A bit of scale: Consider how many children go missing in the US each year.

While I personally also think that the lockdown and movement restrictions have been excessive, it's not especially helpful to compare this new coronavirus to the flu, because people can significantly reduce their risk of catching the flu by getting a vaccination. Lockdown is a strategy put in place not because of the increased mortality rate, but because the "standard" risk mitigation of vaccination is not available for this virus.

Another point of caution here is that CDC adjusts flu mortality rates for under-reporting in an attempt to estimate how many people actually died of the flu, as opposed to how many people have influenza listed on their death certificates.

COVID-19 mortality rates, however, are almost always reported according to narrower inclusion criteria (either positive COVID-19 dx or "presumptive COVID-19"), so there is a risk of an apples-to-oranges comparison here. One ought either count death certificates or estimate the total mortality burden, but not both.

I've looked into this a little bit, actually.

There are two parts of flu/covid IFR: fatalities and infections.

For flu fatalities, the CDC does indeed upward adjust the count. They do this by looking at how many people died of pneumonia, and multiplying that by a number. The number is the estimated percent of pneumonia deaths that are from the flu, which they base on historical data.

For flu infections, it's a bit more obtuse. The best I was able to find was this paper from the CDC [1], which says they use a Monte Carlo model to estimate the number of infections.

For Covid, the deaths are based on diagnoses, not necessarily tests. Diagnoses are from symptoms, so it is not completely accurate. Some flu cases could be misdiagnosed as covid cases if you don't do a test. For infections, the numbers are typically based on the results from antibody tests (not Monte Carlo models).

I agree that it's definitely not apples-to-apples comparison, but it's hard to say if the comparison is in favor of flu being more deadly, or covid being more deadly, because there are confounding factors in both directions.

[1] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4603749/

Thank you for putting that together. Is there a git repo for covid-19.direct or similar?

Yep, the repo is at https://github.com/xunhuang/covid-19 . I didn’t write it, but the maintainer is fairly responsive to pull requests and issues.

Reply on news.ycombinator.com